Re: ext3_inode_info i_datasync_tid field
On Fri, Dec 10, 2010 at 10:27 AM, Onkar Mahajan kern.de...@gmail.com wrote: What is the use of i_datasync_tid in the ext3_inode_info struct ? Hi Onkar, This seems to be in conjunction with the field i_sync_tid in the inode structure. They are used to store the transaction id of the journal. Depending on what system call you use fsync() or fdatasync() to flush the contents of your file, it will use i_sync_tid or i_datasync_tid and will commit the appropriate transaction to disk. Hope that helps... -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: VFS layer filesytem tracking
On Wed, Dec 8, 2010 at 8:00 AM, sanoj k.u.sa...@gmail.com wrote:
Hello all,
I want to track the read/write operations for a particular inode at the vfs
layer.
I added the following structure as a field to struct inode.
struct meta_info
{
unsigned int read_count;
unsigned int write_count;
};
Looking at the fields it seems you want this only incore...and I guess
you want to count how many reads or writes came to a particular file.
This caused a problem in mounting the root device , what could be the cause
?
Can you tell us what was the error/problem ? Stack trace would be better.
Am i allowed modify the struct inode by adding a new field ?
Sure :-) ... you are allowed to modify incore ondisk anything as long
as you make code understand those structures (or if you are adding
something in spare bytes which might get ignored). If you change
something ondisk structure you will need to fix other userspace
utilities which may parse this (in the worst case throw away your old
filesystem and create new one if it changes the size of ondisk inode).
Also, i would like to know if there are any existing mechanisms that can
help me track the workload .
at various levels
Thanks,
HTH
sanoj
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: hi
On Wed, Nov 17, 2010 at 10:54 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On Thu, Nov 18, 2010 at 13:12, navatha reddy navat...@gmail.com wrote: Hi, plz send the information about file system.its not supporting the usb file system without ext2 and ext3. guys, have we ever heard usb filesystem? Or is this Reddy is talking about something that we need deep meditation first? Mulyadi, Take a deep breath..press Delete -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: about EXT3_SUPER_MAGIC
On Sat, Nov 13, 2010 at 8:29 PM, Onkar Mahajan kern.de...@gmail.com wrote: What is EXT3_SUPER_MAGIC ? How is it arrived at ? There is no formula... it is just an identifier. Most of the times it is the name of filesystem in hex. I am writing a file system for which I am referring to the ext3 code... I am not able to arrive at a magic number for my filesystem please help Have you put a name for your filesystem... else you can try 0x6d796673 :-) #define EXT3_SUPER_MAGIC 0xEF53 How this is arrived at ? Regards, Onkar -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Beginner books/articles/papers for Filesystems
On Wed, Nov 3, 2010 at 9:14 AM, Nobin Mathew nobin.mat...@gmail.com wrote: See this http://www.amazon.com/UNIX-Filesystems-Evolution-Design-Implementation/dp/0471164836 On Wed, Nov 3, 2010 at 8:50 PM, Anand Arumugam anand.aru...@gmail.com wrote: Try Maurice J Bach's book. On Wed, Nov 3, 2010 at 9:17 AM, Sankar P sankar.curios...@gmail.com wrote: Hi, Are there any articles or books or whitepapers that are useful for understanding the basics of filesystems (like what are extents, backing-device-infos, caching etc.) ? I dont want a generic OS book, but something that explains about linux filesystems. Any recommendations ? Shankar, As I see, there are two aspects of learning filesystems. a) Understanding the ondisk layout. Once you are comfortable with it, you will be able to understand how it works, and how various ops commands manipulate the ondisk layout. This is file system specific. So if you are looking for this I would suggest some simple filesystem first like ext23 (not ext4 or some other btree based filesystems). b) Upper supporting layers for the filesystem ops. VFS, caching etc. They are generally common to all filesystems and will tell you how to reach to your filesystem, get data from it, and other things perform IO etc. -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: BUG: sleeping function called from invalid context at mm/slab.c
On Wed, Oct 13, 2010 at 11:54 PM, Amit Nagal helloin.a...@gmail.com wrote: Hi , i am getting this kernel log . BUG: sleeping function called from invalid context at mm/slab.c:3055 in_atomic(): 1, irqs_disabled(): 128, pid: 192, name: Front Display Backtrace: [c002cf94] (dump_backtrace+0x0/0x110) from [c02fb048] (dump_stack+0x18/0x1c) r6:c4c0 r5:00d0 r4:c41f1160 r3:4093 [c02fb030] (dump_stack+0x0/0x1c) from [c004cf08] (__might_sleep+0x108/0x128) [c004ce00] (__might_sleep+0x0/0x128) from [c00ccfd0] (kmem_cache_alloc+0x30/0x148) [c00ccfa0] (kmem_cache_alloc+0x0/0x148) from [bf0edf80] (spi_write+0x29c/0x4d4 [fpi]) I guess the interesting part is spi_write+0x0/0x4d4 . Looks like it is trying to do some memory allocation from the place where it is not supposed to do ie... where the code isn't expected to sleep. r8:8013 r7:83bd8d08 r6:a013 r5:0004 r4:c41f1160 [bf0edce4] (spi_write+0x0/0x4d4 [fpi]) from [c00d1498] (vfs_write+0xb4/0x128) [c00d13e4] (vfs_write+0x0/0x128) from [c00d15d4] (sys_write+0x44/0x70) r6:0004 r5:83bd8d08 r4:c1000a60 r3:c77adf70 [c00d1590] (sys_write+0x0/0x70) from [c0028980] (ret_fast_syscall+0x0/0x2c) r9:c77ac000 r8:c0028b28 r6:02b90784 r5:0004 r4:83bd8d08 can anybody guide me what is meant by this bug : BUG: sleeping function called from invalid context at mm/slab.c:3055 and what kernel functions it points to where problem lies . Thanx Regards Amit -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: link list implementation in linux kernel easy question syntax not clear
On Fri, Oct 8, 2010 at 11:38 PM, Bond jamesbond.2...@gmail.com wrote:
Hi,in
http://lxr.linux.no/#linux+v2.6.18/include/linux/list.h
following section of code
there is a structure which is defined as
struct list_head {
struct list_head *next, *prev;
};
It is used in another file as
#define LIST_HEAD_INIT(name) { (name), (name) }
#define LIST_HEAD(name) \
struct list_head name = LIST_HEAD_INIT(name)
static inline void INIT_LIST_HEAD(struct list_head *list)
{
list-next = list;
list-prev = list;
}
I came across a book where the code is given as follows in an example
include/linux/list.h
struct list_head {
struct list_head *next,*prev;
};
#define LIST_HEAD_INIT(name) {(name),(name)}
#define LIST_HEAD(name) struct list_head name = LIST_HEAD_INIT(name)
#define INIT_LIST_HEAD(prt) do {\
(ptr)-next = (ptr);(ptr)-prev= (ptr);\
}while(0)
I was not able to understand above code segment.
I am aware of what a #define is but still I could not understand above thing.
Can some one help in understanding with some example.
Not related to Linux Kernel a normal example where I
can make a link list with above defined way.
http://kernelnewbies.org/FAQ/LinkedLists
--
http://vger.kernel.org/vger-lists.html
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Digest of Questions Solutions
On Tue, Oct 5, 2010 at 12:21 PM, Linux Kernel Explorer mylinux...@gmail.com wrote: Hi James Bond John Mahoney, Thanks a ton for replying. I guess, I did my share of search experiments before posting this one out.I am good on Hygiene factors. John - appreciate you posting that link for me but I guess, I already have it. Infact, if you read my mail once again you will realize that I asked for a compilation of queries posted solutions offered over the years in the pdf format (if not all the years then probably for all months of this year). However, really appreciate your effort. I haven't tried so I am not sure... but with little scripting and all those free pdf convertors it should be easy. And then you can post your script for others too use too :-) Regards, Linux Kernel Explorer On Tue, Oct 5, 2010 at 4:24 PM, Bond jamesbond.2...@gmail.com wrote: On Sun, Oct 3, 2010 at 2:25 PM, Linux Explorer mylinux...@gmail.com wrote: Hello Everyone, Hi explorer things do not work that way as you asked. Do some experiment and then mail here may be some of us can point you to right thing. -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Debugging a kernel oops message
On Fri, Oct 1, 2010 at 3:52 AM, Tushar Dadlani tush...@gmail.com wrote: Hi, What is the best way to debug a kernel oops message? I see some call trace , and some memory addresses , but I am not able to figure out the main reason for my module to fail. Most useful data that I have found is the code data that is sprayed. Have a look at below thread. It might help http://kerneltrap.org/mailarchive/linux-kernel-newbies/2009/9/2/6373273/thread Regards, -- Tushar Dadlani Manipal Institute of Technology +919986292434 -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: open file related system calls
On Thu, Sep 30, 2010 at 5:50 AM, mohit verma mohit89m...@gmail.com wrote: hi all, if my system is shared among several root users . and one of my (say root user S5) process opens a file . is there any system call or any kernel function that can be used by root user S2 to trace my opened file? Question is not clear (atleast to me)... what are root users ? (uid == 0 ?). When you say shared do you mean root logged in from different sessions ? What does tracing an opened file mean ? please reply, guys -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: my kmalloc implementation
Bond, You might want to look at the example of malloc at the end of KnR. On Mon, Sep 27, 2010 at 11:05 PM, Bond jamesbond.2...@gmail.com wrote: I have to write my own kmalloc. I am not given any sort of Kernel API to assign or delete memory. Suppose I have 4GB of memory on Ram. Some of which is filled and some of which is not filled. My question is what data structure do I need to maintain in order to be able to assign memory to any userspace program when the program requests some bytes of memory which can be 1 or more. My logic for this implementation was to maintain a hashtable. For example 1-- points to all the memory addresses which are 1 byte and free 2-- points to all the memory addresses which are 2 byte and free 3-- points to all the memory addresses which are 3 byte and free 4-- points to all the memory addresses which are 4 byte and free . . . . . . . n-- points to all the memory addresses which are n byte and free How can I improve the above schema because to know the location where 1byte memory is free I will maintain a pointer which can be u64 or u32 which itself is costlier than the free memory itself. So what should I be doing to be able to do above. -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: printk and \n
On Sat, Sep 25, 2010 at 9:15 AM, mohit verma mohit89m...@gmail.com wrote: hello guys, are there any gains ( in terms of security or anything else) to flush the buffer in kernel printk() as soon as \n is encountred ? i am asking because this does not happen in the user space as all of us know. This happens in userspace too. -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: why mknod() ??
On Tue, Sep 21, 2010 at 8:06 AM, mohit verma mohit89m...@gmail.com wrote: hey guys, i have a question silly question : as we can create the simple files with open,creat and for directories there is mkdir system calls. so what is the need of mknod() to create special files and as well as ordinary files ? why can't we wrap the special file building code in one of open,creat or mkdir system calls? Because they are special files and some of the commands/operations (eg.. truncate) may not be available for them. If you look at the implementation if is kind of a wrapper over regular create. -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: difference between read permission and executing permission
On Sun, Sep 12, 2010 at 9:07 PM, Parmenides mobile.parmeni...@gmail.com wrote: 2010/9/12 Mulyadi Santosa mulyadi.sant...@gmail.com: On Sun, Sep 12, 2010 at 00:28, Parmenides mobile.parmeni...@gmail.com wrote: Hi, For a specified directory, we can go through it when the kernel parsing path, though we can not read it. Actually, parsing path also need read the directory file. So, how does the kernel distinguish between these two permission? Sorry for replying so late due to some trifles. For a directory, if it is allowed to read, that means we can list all items in its directory file, namely that we can read it completely. May be I understood the question wrong but you can't ls in a directory if it doesn't have execute permission. Or rather even if a directory has read/exectue permission you can't ls on it till all the directories in its path have execute permission. Is there something else that you are asking ? /home/mkatiyar/codetree/main cd /tmp /tmp mkdir -p a/b/c/d /tmp ls -lR a a: total 4 drwxr-xr-x 3 mkatiyar mkatiyar 4096 2010-09-12 22:26 b a/b: total 4 drwxr-xr-x 3 mkatiyar mkatiyar 4096 2010-09-12 22:26 c a/b/c: total 4 drwxr-xr-x 2 mkatiyar mkatiyar 4096 2010-09-12 22:26 d a/b/c/d: total 0 /tmp chmod 000 a/b /tmp ls -lR a a: total 4 d- 3 mkatiyar mkatiyar 4096 2010-09-12 22:26 b ls: cannot open directory a/b: Permission denied /tmp cd a/b ksh: cd: a/b: [Permission denied] /tmp ls a/b ls: cannot open directory a/b: Permission denied If it is allowed to execute, that means we can go through it in a process of path parsing. Path parsing will open every directory in a specified path. For example, when parsing a path like /home/parmenides/src/hello.c, the kernel will open a sequence of directory files, namely /, /home, /home/parmenides and /home/parmenides/src, and read its subdirectory for each. So, both of read permission and executing permission mean that the kernel need to read the corresponding directory file. As such, the common 'readable' has two meanings for a directory. One corresponds to read the directory file when listing it, the other corresponds to read the directory file when paring path. On the other hand, system call opendir(const char *pathname) has no way to distinguish these two meanings whereas both listing and path parsing need call opendir(). My quesition is exactly that how these two permissions take their effect. -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: question about linked list implementation in kernel.h
On Sat, Sep 11, 2010 at 5:33 AM, Bond jamesbond.2...@gmail.com wrote:
I read
I was going through include/linux/kernel.h
encountered following code
#define container_of(ptr, type, member) ({ \
const typeof( ((type *)0)-member ) *__mptr = (ptr); \
in above code I am not clear with (type *)0
If you want some more details about this macro.
http://www.spinics.net/lists/linux-usb-devel/msg11766.html
how is it working any link?
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Query related to mount point identification.
On Tue, Sep 7, 2010 at 4:27 AM, Tayade, Nilesh nilesh.tay...@netscout.com wrote: Hi, I am currently working on parsing the packets for NFS protocol. I want to know, in case of a file operation, how can we identify the mount point where the file is mounted? E.g. scenario is- Client has mounted the node already. You apply the NFS parser and start reading/writing to the file. Here, parser did not know if the node is mounted or not (as we enabled the parser after mounting the node. And I want to identify that at the arrival of first packet). So is there any way to identify that based on the file name (or inode may be)? I am looking at the fstat() system call, but unable to find anything useful. Looks like given the filename you want to find the mountpoint. May be this will help. http://www.mail-archive.com/bug-coreut...@gnu.org/msg20630.html -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: changing kernel variables at run time
On Tue, Sep 7, 2010 at 10:29 AM, Tapas Mishra mightydre...@gmail.com wrote: How can one change the value of Kernel Variables at run time. Depending on what you what to change, you may want to look at /proc /sys -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: multiple alternate superblocks
On 7/7/10, Onkar Mahajan kern.de...@gmail.com wrote: Is it possible to supply multiple alternate superblocks while doing a mount in Linux ? Why would you want to do that ? What is the expected behavior in such cases ? -Regards, Onkar -- Your work is going to fill a large part of your life, and the only way to be truly satisfied is to do what you believe is great work. And the only way to do great work is to love what you do. If you haven’t found it yet, keep looking, and don’t settle. As with all matters of the heart, you’ll know when you find it. - Steve Jobs -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Understanding NULL pointer dereference
On Fri, Jun 18, 2010 at 10:11 AM, Prasad Joshi prasadjoshi...@gmail.com wrote: Hi All, I am trying to understand a a kernel oops report. Here are some of the fields from the report BUG: unable to handle kernel NULL pointer dereference at 0034 Oops: [#1] SMP Pid: 6478, comm: cp Tainted: P 2.6.31.5-127.fc12.x86_64 #1 Inspiron 1525 RIP: 0010:[810faac1] [810faac1] do_sys_open+0x7a/0x10f CR2: 0034 As I know, when a page fault occurs, the address the program attempted to access is stored in the CR2 register. So probably the pointer is pointing to address 0034 and is being access. The BUG string is bit confusing, it says NULL pointer dereference at 0034, I know the address 00034 is not valid but why is it interpreted as NULL pointer? The NULL pointer as I know should point to address 0. As ppl have already mentioned, this is some offset into a structure. Next thing you can try is to do a disas do_sys_open and find out which member/structure is being accessed which has this offset. You can also try to get the assembly instructions from the code dumped. if you already have the source code for your running kernel you can also try l *do_sys_open+0x7a. Thanks and Regards, Prasad -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Understanding NULL pointer dereference
On Fri, Jun 18, 2010 at 5:09 PM, Prasad Joshi prasadjoshi...@gmail.com wrote:
On Fri, Jun 18, 2010 at 1:19 PM, Manish Katiyar mkati...@gmail.com wrote:
On Fri, Jun 18, 2010 at 10:11 AM, Prasad Joshi prasadjoshi...@gmail.com
wrote:
Hi All,
I am trying to understand a a kernel oops report. Here are some of the
fields from the report
BUG: unable to handle kernel NULL pointer dereference at
0034
Oops: [#1] SMP
Pid: 6478, comm: cp Tainted: P 2.6.31.5-127.fc12.x86_64 #1 Inspiron 1525
RIP: 0010:[810faac1] [810faac1]
do_sys_open+0x7a/0x10f
CR2: 0034
As I know, when a page fault occurs, the address the program attempted
to
access is stored in the CR2 register. So probably the pointer is
pointing to
address 0034 and is being access.
The BUG string is bit confusing, it says NULL pointer dereference at
0034, I
know the address 00034 is not valid but why is it interpreted as NULL
pointer? The NULL pointer as I know should point to address 0.
As ppl have already mentioned, this is some offset into a structure.
Next thing you can try is to do a disas do_sys_open and find out which
member/structure is being accessed which has this offset. You can also
try to get the assembly instructions from the code dumped. if you
already have the source code for your running kernel you can also try
l *do_sys_open+0x7a.
Thanks Manish for your reply. The problem was with the return value in FS
create() call. The create() call was returning the +ve error code. I looked
into some of the other file systems, they use PTR_ERR to return the error
code. Basically the -ve error code is assigned to unsigned long causing,
thus converting it in a big number.
struct file *f = do_filp_open(dfd, tmp, flags, mode, 0);
if (IS_ERR(f)) {
put_unused_fd(fd);
fd = PTR_ERR(f);
} else {
fsnotify_open(f-f_path.dentry); === OOPs
fd_install(fd, f);
}
So IS_ERR() was returning false and causing panic in else path.
Cool so looks like you crashed while dereferencing d_parent of a
negative dentry.
(gdb) p (int)(((struct dentry *)0)-d_parent)
$5 = 52
(gdb) p/x (int)(((struct dentry *)0)-d_parent)
$6 = 0x34
Thanks -
Manish
Thanks a lot for your help.
Regards,
Prasad
Thanks and Regards,
Prasad
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Question about do_sync_read()
[adding krenel newbies in cc, as there might be others who have better answers]
On Sat, Jun 19, 2010 at 12:49 AM, Anand Pai anand_...@yahoo.com wrote:
Manish:
You are absolutely correct ! I agree with you that once I decide to use the
filp, then the inode operations will be used.
But ...your response has caused me to ask a more fundamental question.
This is the scenario I am trying to initiate a read of LKM read of a file
into a kernel buffer.
So .. I do a kernel_read.
This ... does a vfs_read.
vfs_read uses file-f_op and file-f_op-read to decide ... WHICH read it
will do:
do_sync_read OR file-f_op-read
Now ... I dont see where file-f_op-read or file-f_op are being initialized.
kernel_read needs a file * as the first argument, What are you passing
there ? How did you get file* there.
Thanks -
Manish
I have shown all the greps etc. Thias has been awhile. Please help me if you
can.
Also, another question is ... when IS do_synch_read SUPPOSED to be used (over
file-f_op-read) anyway ? What is the philisophy if you will ?
Thanks so much Manish. I do look forward to your response. Regards
-Anand
--- On Wed, 6/16/10, Manish Katiyar mkati...@gmail.com wrote:
From: Manish Katiyar mkati...@gmail.com
Subject: Re: Question about do_sync_read()
To: Anand Pai anand_...@yahoo.com
Date: Wednesday, June 16, 2010, 8:32 AM
If you happened to trace the code enough, you will realise that from the
pathname we try to get a file * (filp)... tracing somemore you will find that
ultimately __dentry_open(...) will be called which has the below line :-
static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt,
struct file *f,
int (*open)(struct inode *, struct file *),
const struct cred *cred)
{
.
...
f-f_mapping = inode-i_mapping;
f-f_path.dentry = dentry;
f-f_path.mnt = mnt;
f-f_pos = 0;
f-f_op = fops_get(inode-i_fop);
..
Hope that clears your doubt.
On Wed, Jun 16, 2010 at 3:48 AM, Anand Pai anand_...@yahoo.com wrote:
Hi Manish:
Thanks so much for your response. As you can see from the attached screen
shots,
ext3_file_operations is getting attached to inode--i_fop.
kernel_read (and vfs_read used by it) uses file-f_op, and I cant see where
the appropriate
plave where any function/ method gets attached to file-f_op.
This is noted in the third jpg (File Operations 3.jpg). I just cant find
which method of file
read is used by vfs_read. Could you please advise ?
Thanks, Regards
-Anand
--- On Sun, 6/13/10, Manish Katiyar mkati...@gmail.com wrote:
From: Manish Katiyar mkati...@gmail.com
Subject: Re: Question about do_sync_read()
To: Anand Pai anand_...@yahoo.com
Date: Sunday, June 13, 2010, 7:38 PM
For ext3 filesystem, see the structure ext3_file_operations{} in
ext3/file.c. Similar thing is done for other filesystems. and trace the
function generic_file_aio_read().
On Mon, Jun 14, 2010 at 1:58 AM, Anand Pai anand_...@yahoo.com wrote:
Hi Manish:
Cane across your question on do_sych_read I was wondering if you could
help point me to exactly where the buf is getting written on the read from
file
I am trying to do a kernel_read of a file from within an LKM, but the buf
is not
getting updated after completion of the kernel_read.
I would appreciate any help at all. Thanks so much
-Anand
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: why is Size of sigset_t 128 bytes ?
On Tue, Jun 15, 2010 at 6:30 PM, Siddu siddu.s...@gmail.com wrote:
Hello all,
I came across an article (http://www.linuxjournal.com/article/6483) about
signals . Its a very nice article
It says and also i verified that the size of sigset_t showing 128 bytes ?
while code from /* /usr/include/bits/sigset.h */
Maybe this is not the right place to ask this Sorry, Hoping someone
would be knowing the answer
# define _SIGSET_NWORDS (1024 / (8 * sizeof (unsigned long int)))
typedef struct
{
unsigned long int __val[_SIGSET_NWORDS];
} __sigset_t;
typedef __sigset_t sigset_t;
which is definitely not 128 bytes .
Why not ???
Assuming 32 bit machines unsigned long int is 4 bytes, and you are
declaring and array of 32 elements of 4 bytes each.
Thanks for your inputs !
--
Regards,
~Sid~
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - -
Gazing through the window at the world outside,
Wondering will mother earth survive ?
Hoping that the mankind will stop abusing her
Sometime !
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - -
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Stack Tracing using Ftrace
On Mon, May 3, 2010 at 7:02 PM, Neependra Khare neependra.kh...@gmail.com wrote: Hi, I wrote a module and somehow it is corrupting the thread_info of the process. After some debugging I concluded that its corrupting the kernel stack. Then I came across Ftrace to get the stack trace, which confirmed high stack usage. http://lwn.net/Articles/366796/ I see around 7K in the depth section of first entry in stack_trace. At the same time I run a stap script to get the current stack usage using stack_used function. It comes around 4k. Given that you are eating so much of stack space, you must have lots of local variables or some *big* ones. It might be easy to just open your .ko check the sizes of the variables and compute from there. You can also enable below variables in your .config. CONFIG_DEBUG_STACKOVERFLOW=y CONFIG_DEBUG_STACK_USAGE=y Thanks - Manish To my understanding the depth section of first entry in stack_trace and value retuned by stack_used should match. Is my understanding correct? Thanks Neependra -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Dynamic Kernel Module Section Addresses
On Wed, Apr 28, 2010 at 11:33 AM, Venkatram Tummala venkatram...@gmail.com wrote: On Tue, Apr 27, 2010 at 8:01 PM, Manish Katiyar mkati...@gmail.com wrote: On Wed, Apr 28, 2010 at 6:38 AM, Venkatram Tummala venkatram...@gmail.com wrote: On Tue, Apr 27, 2010 at 5:04 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: Hi Venkatram... On Wed, Apr 28, 2010 at 05:34, Venkatram Tummala venkatram...@gmail.com wrote: 1) Is there any way to tell the kernel not to load these module sections at different addresses every time the module is inserted removed . That is, can we have the module loaded at exactly the same address everytime? None that I know. AFAIK the only way is to recheck the /sys entry and reinsert the symbol file (your kernel module file, which has been compiled with -g before) to proper position With this approach, i can set breakpoints in every function except for the function that initializes the module because i have to insert the module to get the kernel module section addresses from /sys as a result of which, the initializition routine finishes executing which means i can not set any breakpoints in the module initialize function. This is not a problem for the the module exit function or any other function though. There should be a way to tell the kernel that Look, I dont care where you put the segments of this module but put them at the same place everytime i re-insert the module again (atleast until the next reboot). Conceptually, this should not be hard to implement (if its not there already) because it is the same kernel the same unchanged KO module that i am asking it to insert. 2) Is there any way to overcome this problem so that i can debug the module using vmware even though the module is loaded at different addresses? Probably i didn't understand the question. Can't you find it by looking at p modules from gdb ? Hey Manish, to find these addresses by p modules, the module has to be already loaded.right? So, the initialization function of the module must have been already executedright? So, the question is how can we step through the initialization function of the module..? Yes, you can't step through initialization function using this. Thanks - Manish p modules won't help us..probably looking at /sys is a faster way to find module segment addresses, or am I missing something here in p modules?? I will try out Mulyadi's trick get back ! Regards, Venkatram Tummala Thanks - Manish AFAIK, again none that I know. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: question about get_user
On Wed, Apr 28, 2010 at 12:24 PM, Yang Fangkai wolfgang.y...@gmail.com wrote:
Hi, Nidhi,
Thanks! Now I know what is the problem.
But in my project, I am developing a virtual char device driver
such that user can read any process's memory contents by specifying
its pid and virtual address. Therefore, a device /dev/gh is
registered, and the user try to do something like:
echo -n targetpid/dev/gh echo -n addr/dev/gh
Therefore, the target process can't be modified. The device driver
will retrieve the value by pid and virtual address and return it to
user at terminal.
Isn't it something similar to ptrace with PTRACE_PEEKTEXT, PTRACE_PEEKDATA
Thanks -
Manish
I don't know how to solve this problem with get_user
under my scenario. Thank you!
Fangkai
On Wed, Apr 28, 2010 at 1:42 AM, nidhi mittal hada
nidhimitta...@gmail.com wrote:
you should somehow call the kernel module from user space program ...
may be using ioctl ..
so that you are in same process context --- as the one you are sending
address from .
write char dev driver with ioctl method defined
then call ioctl from user space
ioctl(fd,IOCTL_CMD,addrees you want to send)
Now in kernel module
get_user(kernel data variable, user space address);
Nidhi
On Wed, Apr 28, 2010 at 11:55 AM, Yang Fangkai wolfgang.y...@gmail.com
wrote:
Hi, Nidhi,
Thanks for your reply! Yes, you are right. I pass the address
to the module from bash command echo, therefore when the address is
referred, the current pid is bash's pid, instead of the simple program
I wrote.
But how can I fix this problem?
Thank you!
Fangkai
On Wed, Apr 28, 2010 at 1:01 AM, nidhi mittal hada
nidhimitta...@gmail.com wrote:
in your kernel module try to print current-pid
is it same as the user space process id ?
i think when in kernel module you are not in the same process context
whihc
you want to refer ...
Nidhi
On Wed, Apr 28, 2010 at 10:38 AM, Yang Fangkai wolfgang.y...@gmail.com
wrote:
Hi, all,
I have a problem with get_user() macro. What I did is as follows:
I run the following program
int main()
{
int a = 20;
printf(address of a: %p, a);
sleep(200);
return 0;
}
When the program runs, it outputs the address of a, say, 0xbff91914.
Then I pass this address to a module running in Kernel Mode that
retrieves the contents at this address (at the time when I did this, I
also made sure the process didn't terminate, because I put it to sleep
for 200 seconds... ):
The address is firstly sent as a string, and I cast them into pointer
type.
int * ptr = (int*)simple_strtol(buffer, NULL,16);
printk(address: %p,ptr); // I use this line to make sure the cast is
correct. When running, it does output bff91914
int val = 0;
int res;
res= get_user(val, (int*) ptr);
However, res is always not 0, meaning that get_user returns error. I
am wondering what is the problem
Thank you!!
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks Regards
Nidhi Mittal Hada
Scientific officer D
Computer Division
Bhabha Atomic Research Center
Mumbai
--
Thanks Regards
Nidhi Mittal Hada
Scientific officer D
Computer Division
Bhabha Atomic Research Center
Mumbai
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Dynamic Kernel Module Section Addresses
On Wed, Apr 28, 2010 at 6:38 AM, Venkatram Tummala venkatram...@gmail.com wrote: On Tue, Apr 27, 2010 at 5:04 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: Hi Venkatram... On Wed, Apr 28, 2010 at 05:34, Venkatram Tummala venkatram...@gmail.com wrote: 1) Is there any way to tell the kernel not to load these module sections at different addresses every time the module is inserted removed . That is, can we have the module loaded at exactly the same address everytime? None that I know. AFAIK the only way is to recheck the /sys entry and reinsert the symbol file (your kernel module file, which has been compiled with -g before) to proper position With this approach, i can set breakpoints in every function except for the function that initializes the module because i have to insert the module to get the kernel module section addresses from /sys as a result of which, the initializition routine finishes executing which means i can not set any breakpoints in the module initialize function. This is not a problem for the the module exit function or any other function though. There should be a way to tell the kernel that Look, I dont care where you put the segments of this module but put them at the same place everytime i re-insert the module again (atleast until the next reboot). Conceptually, this should not be hard to implement (if its not there already) because it is the same kernel the same unchanged KO module that i am asking it to insert. 2) Is there any way to overcome this problem so that i can debug the module using vmware even though the module is loaded at different addresses? Probably i didn't understand the question. Can't you find it by looking at p modules from gdb ? Thanks - Manish AFAIK, again none that I know. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: delete from hlist ???
On Thu, Apr 8, 2010 at 11:28 AM, Onkar Mahajan kern.de...@gmail.com wrote:
I am not able to understand how the marked like skips the node n :
static inline void __hlist_del(struct hlist_node *n)
{
struct hlist_node *next = n-next;
struct hlist_node **pprev = n-pprev;
*pprev = next; ---
if (next)
next-pprev = pprev;
}
in this case
static inline void __list_del(struct list_head * prev, struct list_head *
next)
{
next-prev = prev;
prev-next = next;
}
It is clear ...
Why is not clear in the first case. It is doing exactly the same
thing. Except the fact the in list_head both the fields prev and next
are pointers while in hlist next is pointer but prev is pointer to
pointer.
NO ?
Thanks -
Manish
Please help.
Regards,
Onkar
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: list query
2010/4/8 Cédric Augonnet cedric.augon...@gmail.com: 2010/4/8 Onkar Mahajan kern.de...@gmail.com: What is LIST_POISON1 and LIST_POISON2 ?? #define LIST_POISON1 ((void *) 0x00100100 + POISON_POINTER_DELTA) #define LIST_POISON2 ((void *) 0x00200200 + POISON_POINTER_DELTA) and why they are used ? Regards, Onkar Hi Onkar, A poison value is a convenient value that you can detect when debugging. For instance if you find out that a pointer is 0xdeadbeef, you can easily imagine this is some uninitialized value. If you had used a NULL value, it would be harder to determine whether it's something uninitialized ... yes... specially if it is some special value (like LIST_POISON1) in the core and you can immediately go and look at the places where all they get set and you can start from there. Setting it to NULL or something generic like 0xdeadbeef will make this hard. thanks - Manish Hope this helps, Cédric -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: need to get blocks per group in user space
On Wed, Apr 7, 2010 at 1:06 PM, nidhi mittal hada nidhimitta...@gmail.com wrote: hello All I want to get some of filesystem specific information like blocksize , s_blocks_per_group etc in userspace c program How do i get it ? Use libext2fs or see the code of dumpe2fs i have obtained block size by stat() function but not more than that :( -- Thanks Regards Nidhi Mittal Hada Scientific officer D Computer Division Bhabha Atomic Research Center Mumbai -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Allocating memory with spinlock held
Hello all,
Is the following code valid ???
foo() {
...
spin_lock(lock);
ptr = kzalloc(size, GFP_KERNEL);
spin_unlock(lock);
.
}
I guess the answer is no, but would like to confirm. Detailed/all
possible reasons for not doing so are welcome. IIRC, I was getting a
kernel crash with similar code with error message as spinlock bug :
unlocked on wrong CPU which looked like it was getting locked on one
cpu and due to memory allocation was getting unlocked from another
cpu.
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Allocating memory with spinlock held
Hello all,
Is the following code valid ???
foo() {
...
spin_lock(lock);
ptr = kzalloc(size, GFP_KERNEL);
spin_unlock(lock);
.
}
I guess the answer is no, but would like to confirm. Detailed/all
possible reasons for not doing so are welcome. IIRC, I was getting a
kernel crash with similar code with error message as spinlock bug :
unlocked on wrong CPU which looked like it was getting locked on one
cpu and due to memory allocation was getting unlocked from another
cpu.
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Allocating memory with spinlock held
On Thu, Mar 4, 2010 at 2:16 AM, Frederic Weisbecker fweis...@gmail.com wrote:
On Tue, Mar 02, 2010 at 09:32:16AM +0530, Manish Katiyar wrote:
Hello all,
Is the following code valid ???
foo() {
...
spin_lock(lock);
ptr = kzalloc(size, GFP_KERNEL);
spin_unlock(lock);
.
}
You can't because GFP_KERNEL involves the fact
kzalloc might sleep while looking for memory
somewhere (swapping out pages to get some free
rooms).
And sleeping while holding a spinlock may
lead to a deadlock:
Task A holds spinlock in cpu0. Task A sleeps.
Task B is scheduled in cpu0, tries to take the
spinlock and then deadlock for ever.
You can use GFP_ATOMIC instead, but it's usually
not recommended as the GFP_ATOMIC pool is a pretty
limited resource.
Hi,
Thanks for the clarification. That is what I had suspected too and
wanted to clarify.
The best is to allocate before you take the spinlock.
Yes, that is what I ended coding up.
Thanks again -
Manish
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: [ANNOUNCE] Online Hierarchical Storage Manager (OHSM v1.2)
On Tue, Feb 23, 2010 at 7:57 PM, Jan Kara j...@ucw.cz wrote: Hello Manish, We are pleased to announce the first official functional release of Online Hierarchical Storage Manager (OHSM v1.2). This is a RFC release and not yet aimed at mainline inclusion. OHSM is a tool to manage and move data across various class of storage. It can help users to selectively place and move data across tiers such as SSD, Raid 10, Raid 6 based on the attributes of the data. OHSM supports background movement of data without any visible change in a files namespace to users and user applications. OHSM is built as an external module with recompilation required for ext4. The current version of OHSM is based upon kernel 2.6.32.2 This release includes 3 core kernel patches: 1) An initial ext4 patch to adjust ext4's block allocation to use a preferred block range per Ted Tso's Dec 2008 write-up (http://markmail.org/message/qp7zjhhdzxum7rfn). This patch is not ohsm specific. 2) A ohsm specific ext4 patch to provide callouts to the ohsm module. 3) The ohsm module itself The source code for OHSM v1.2 is freely distributable under GPL. The latest stable OHSM v1.2 is available at : http://sourceforge.net/projects/ohsm/files/OHSMv1.2.tar.gz/download Sources are available as git repository at : git://ohsm.git.sourceforge.net/gitroot/ohsm/ohsm Looking at the code, the code deciding where to put a given file isn't really tied to ext4, is it? Hi Jan, ext4 was our default choice to get started, because of the ioctls provided by Akira and since e4defrag already exists it was much easier for us to refer. Though we have tried to keep the OHSM code as independant as possible so that it would be possible to support other filesystems in future too. Using pathnames to decide where to put files is ... natural but it has its downsides. For example your code in build_path() is racy because the directory hiearchy can change while you climb the tree - not mentioning the fact that the static buffer of 4096 for pathname is ugly and implementing that simple walk recursively isn't good either. Kernel already has a d_path function for constructing path in at least somewhat reliable way. Thanks a lot for your feedback. We have fixed some bugs recently to get rid of static allocations so they will go away. I will have a look at d_path(). Anyway I suggest you have a look at Tomoyo which implements pathname based security module so that you can avoid mistakes they did... Sure, will do. Definitely it will help us learn more and make the code base better. Thanks - Manish Honza -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Absolute Path from linux_dirent64
On Sat, Feb 13, 2010 at 11:14 PM, Sam Carter sam.w.car...@gmail.com wrote: Hi guys, Can I get the string of the full path to a linux_dirent64 file somehow? If I understand your question correctly, that is not possible. a linux_dirent64 will have information about your current filename (ie.. name, record len etc.). Just by using this it is not possible to get a unique parent for the filename. For example if you have a filename foo you may have different parents like /home/mkatiyar/A/foo /home/mkatiyar/B/foo In such cases what would you expect to get ? the parent should be 'A' or 'B' Thanks - Manish I've looked at stat() but it wasnt very helpful. Also d_path(), but that takes a dentry struct, not a dirent like I have. Thanks, Sam -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
[ANNOUNCE] Online Hierarchical Storage Manager (OHSM v1.2)
Hello all, We are pleased to announce the first official functional release of Online Hierarchical Storage Manager (OHSM v1.2). This is a RFC release and not yet aimed at mainline inclusion. OHSM is a tool to manage and move data across various class of storage. It can help users to selectively place and move data across tiers such as SSD, Raid 10, Raid 6 based on the attributes of the data. OHSM supports background movement of data without any visible change in a files namespace to users and user applications. OHSM is built as an external module with recompilation required for ext4. The current version of OHSM is based upon kernel 2.6.32.2 This release includes 3 core kernel patches: 1) An initial ext4 patch to adjust ext4's block allocation to use a preferred block range per Ted Tso's Dec 2008 write-up (http://markmail.org/message/qp7zjhhdzxum7rfn). This patch is not ohsm specific. 2) A ohsm specific ext4 patch to provide callouts to the ohsm module. 3) The ohsm module itself The source code for OHSM v1.2 is freely distributable under GPL. The latest stable OHSM v1.2 is available at : http://sourceforge.net/projects/ohsm/files/OHSMv1.2.tar.gz/download Sources are available as git repository at : git://ohsm.git.sourceforge.net/gitroot/ohsm/ohsm The following list describes the features of OHSM * Support for Ext4. * File System callouts * Policy based File Migration * Per inode information maintainer in user xattr area * XML based policy files * Support for RAID * DTD based XML validation - Automatic Relocation engine * Selective Relocation of mountpoint/Subdirectories * Selection Relocation based on specific rule * File Placement based on: + UID + GID + File Type + Directory Name * File Migration based on + UID + GID + FILE SIZE + ACCESS AGE + MODIFICATION AGE - File Access Temperatue - File I/O Temperature + (*=current features, -= on our todo list, +=support already present) Here are some simple examples :- $ ohsm enable sample/ohsm.xml OHSM enabled successfully on '/tmp/mnt' $ ohsm relocate -m /tmp/mnt Relocation performed successfully. $ ohsm status /tmp/mnt OHSM is enabled on '/tmp/mnt' $ ohsm disable /tmp/mnt OHSM disabled successfully on '/tmp/mnt' $ ohsm Usage: ohsm subcommand [optstring ..] [parameters ..] Subcommands : enable - Enable OHSM disable - Disable OHSM status - Show status of OHSM relocate -Trigger relocation on filesystem update - Update OHSM instance parameters help -Show usage *Note* that the code for OHSM is still in early stages Feedback, bug reports and patches, as always, are welcomed. (ohsm-de...@lists.sourceforge.net) -- Thanks - OHSM team -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: ieee papers
On Fri, Feb 19, 2010 at 12:47 AM, Krushnaal Pai krisonea...@gmail.com wrote: OK... SORRY FOR USING UR PORTAL FOR IEEE PAPERS WONT DO IT AGAIN And typing in caps is again *something to avoid* while posting in mailing lists. Thanks - Manish On Thu, Feb 18, 2010 at 11:29 PM, Denis Kirjanov kirja...@gmail.com wrote: Recently there are too many messages not related to the Linux kernel topic. And this is a bad sign :/ On Thu, Feb 18, 2010 at 7:04 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On Thu, Feb 18, 2010 at 4:24 PM, Krushnaal Pai krisonea...@gmail.com wrote: Anybody got any full sample ieee papers/abstracts?... Plz mail me guyz.. Since when kernelnewbies becomes the new place to ask for IEEE paper? PS: Rik, could these things happen because they don't read the kernelnewbies' posting rule? -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Regards, Denis -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: latest kernel compile-- system panic
On Thu, Jan 28, 2010 at 5:05 PM, waqar afridi afridi.wa...@gmail.com wrote: I used to get these errors when I started with kernel compilation and by this I mean a few weeks back. Where do u put the kernel source when you compile it, I used to download it and extract it to my home Directory and then the list of commands above, but it never worked. The Solution to this problem is to Extract the Kernel source to /usr/src/ and then do the above process (make etc), may be this will help... Where you put the kernel sources to build is not at all related to these panics (Otherwise you don't need a build symlink in your /lib/modules/version). Did you build the same config ? Thanks - Manish On Thu, Jan 28, 2010 at 3:25 PM, nidhi mittal hada nidhimitta...@gmail.com wrote: Hello This time also . i just copied the .config of my running kernel to the downloaded new kernel's directory and did make and further steps. and it executed successfully. no kernel panic. is it the right way .? But i helplessly do it as last resort everytime when after my all efforts new kernel panics on boot everytime. Nidhi On Thu, Jan 28, 2010 at 2:53 PM, nidhi mittal hada nidhimitta...@gmail.com wrote: hi all i compiled latest kernel linux2.6.32.6 after 4 steps of make menuconfig make make modules make modules_install make install when i rebooted these error messages came Unable to access resume device (/dev/sda3) mount: could not find filesystem /dev/root' setuproot : moving /dev failed : No such file or directory setuproot : error mounting /proc : No such file or directory setuproot : error mounting /Sys : No such file or directory switchroot: mount failed No such file or directory earlier also whenever i compiled new kernel .. i used to copy the .config of old running kernel into new liinux folder as everytime if i create .config file with make menuconfig .. some unknown mistake gives kernel panic i need to know how to set proper .config so that it caters to all requirement of my present machine hardware i got to know make cloneconfig but it gives error as make[1]: *** No rule to make target `cloneconfig'. Stop. make: *** [cloneconfig] Error 2 will make oldconfig work ?? what to do Nidhi -- Thanks Regards Nidhi Mittal Hada -- Thanks Regards Nidhi Mittal Hada -- Waqar Afridi Android Developer SERG IM|Sciences -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: latest kernel compile-- system panic
On Thu, Jan 28, 2010 at 6:02 PM, nidhi mittal hada nidhimitta...@gmail.com wrote: yes when i get unhinting errors -- i guessed something , changed a few params in .config and did make make modules so on but again same errors . finally i copied earlier running kernel's .config into this ...didnot do make menuconfig at all and in .config changed my required parameter of ext4 to be build as module. that's all now make , make modules , make modules_install , make install and it worked well now my ques is ...is it right what i am doing ... if not how to know correct parameter of .config isnt there some make someconfig instead of menuconfig which prepares correct .config file seeing hardware requirement on my pc or working kernel's .config file ... make oldconfig make help Please CMIIW anywhere. Nidhi On Thu, Jan 28, 2010 at 5:13 PM, Manish Katiyar mkati...@gmail.com wrote: On Thu, Jan 28, 2010 at 5:05 PM, waqar afridi afridi.wa...@gmail.com wrote: I used to get these errors when I started with kernel compilation and by this I mean a few weeks back. Where do u put the kernel source when you compile it, I used to download it and extract it to my home Directory and then the list of commands above, but it never worked. The Solution to this problem is to Extract the Kernel source to /usr/src/ and then do the above process (make etc), may be this will help... Where you put the kernel sources to build is not at all related to these panics (Otherwise you don't need a build symlink in your /lib/modules/version). Did you build the same config ? Thanks - Manish On Thu, Jan 28, 2010 at 3:25 PM, nidhi mittal hada nidhimitta...@gmail.com wrote: Hello This time also . i just copied the .config of my running kernel to the downloaded new kernel's directory and did make and further steps. and it executed successfully. no kernel panic. is it the right way .? But i helplessly do it as last resort everytime when after my all efforts new kernel panics on boot everytime. Nidhi On Thu, Jan 28, 2010 at 2:53 PM, nidhi mittal hada nidhimitta...@gmail.com wrote: hi all i compiled latest kernel linux2.6.32.6 after 4 steps of make menuconfig make make modules make modules_install make install when i rebooted these error messages came Unable to access resume device (/dev/sda3) mount: could not find filesystem /dev/root' setuproot : moving /dev failed : No such file or directory setuproot : error mounting /proc : No such file or directory setuproot : error mounting /Sys : No such file or directory switchroot: mount failed No such file or directory earlier also whenever i compiled new kernel .. i used to copy the .config of old running kernel into new liinux folder as everytime if i create .config file with make menuconfig .. some unknown mistake gives kernel panic i need to know how to set proper .config so that it caters to all requirement of my present machine hardware i got to know make cloneconfig but it gives error as make[1]: *** No rule to make target `cloneconfig'. Stop. make: *** [cloneconfig] Error 2 will make oldconfig work ?? what to do Nidhi -- Thanks Regards Nidhi Mittal Hada -- Thanks Regards Nidhi Mittal Hada -- Waqar Afridi Android Developer SERG IM|Sciences -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- Thanks Regards Nidhi Mittal Hada -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Needs the complete definition user_path_walk()
On Wed, Jan 27, 2010 at 1:35 PM, waqar afridi afridi.wa...@gmail.com wrote:
I wasnt expecting that much quick reply, Zellux and specially Manish for for
sharing this great Ocean of Knowledge (Google)...
On Wed, Jan 27, 2010 at 12:33 PM, Manish Katiyar mkati...@gmail.com wrote:
On Wed, Jan 27, 2010 at 12:36 PM, waqar afridi afridi.wa...@gmail.com
wrote:
Hello Every One
This is my First mail to this community, so dont know much about the
Rules
and Regulations.
In general doing google is a good idea.
I want the Definition of user_path_walk() function, If any one had,
Please
send it and if not, can any one tell me where could I find it.
I guess it was removed after 2.6.6 . And now this function is called
as user_path. Look at the following commit which can give you some
idea.
commit 2d8f30380ab8c706f4e0a8f1aaa22b5886e9ac8a
Author: Al Viro v...@zeniv.linux.org.uk
Date: Tue Jul 22 09:59:21 2008 -0400
[PATCH] sanitize __user_walk_fd() et.al.
...
...
+asmlinkage long compat_sys_statfs(const char __user *pathname, struct
compat_statfs __user *buf)
{
- struct nameidata nd;
+ struct path path;
int error;
- error = user_path_walk(path, nd);
+ error = user_path(pathname, path);
if (!error) {
struct kstatfs tmp;
- error = vfs_statfs(nd.path.dentry, tmp);
+ error = vfs_statfs(path.dentry, tmp);
if (!error)
error = put_compat_statfs(buf, tmp);
- path_put(nd.path);
+ path_put(path);
}
...
Thanks -
Manish
This is the first link in google.
http://www.developerweb.net/forum/archive/index.php/t-3466.html
I already had tried this link but it did gave me anything, The reason might
be that I am newbie, But things will improve...
Thanx Alot
Thanx In advance
--
Waqar Afridi
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
Waqar Afridi
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: dentry path name lookup
On Wed, Jan 27, 2010 at 1:37 PM, er krishna erkris...@gmail.com wrote: Dear All, Can anybody please help me how does dentry does path resolution. I want to undersatnd the whole process in detail. Any link or comments are most welcome. Just asking for my understanding, Have a look at function link_path_walk(). Thanks - Manish Dentry maintains a relationship between file object inode for path name lookup ? Is it right ? If the pathname is /home/krishna/myfile, how it will be resolve ? Best Regards, Krishna -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Needs the complete definition user_path_walk()
On Wed, Jan 27, 2010 at 12:36 PM, waqar afridi afridi.wa...@gmail.com wrote: Hello Every One This is my First mail to this community, so dont know much about the Rules and Regulations. In general doing google is a good idea. I want the Definition of user_path_walk() function, If any one had, Please send it and if not, can any one tell me where could I find it. This is the first link in google. http://www.developerweb.net/forum/archive/index.php/t-3466.html Thanx In advance -- Waqar Afridi -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Kernel Build fail on Ubuntu
On Tue, Jan 19, 2010 at 5:56 PM, Ramya Desai ramya.de...@gmail.com wrote: Dear Experts, I am not able to compile the required kernel on Ubuntu system. When I do the make modules command it showed the following messages. Due to these messages, it does not creates the initrd image file in the /boot directory to that specified kernel. WARNING: modpost: Found 2 section mismatch (es) To see full details build your kernel with: make CONFIG_DEBUG_SECTION_MISMATCH=y I have tried in different ways as suggested in some of the google links as well. However, there is no luck. Could you please any one suggest me how can I resolve the issue? You can ignore this and run mkinitramfs your kernel version in /boot. Thanks - Manish Thanks in advance. RD -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: copy-on-write overhead
On Mon, Jan 11, 2010 at 3:01 PM, Pete greg.p...@gmail.com wrote:
On Mon, Jan 11, 2010 at 6:22 AM, Mulyadi Santosa
mulyadi.sant...@gmail.com wrote:
On 1/11/10, Joel Fernandes agnel.j...@gmail.com wrote:
Oh I'm sorry, if you were talking about copying of the address space
information that can be avoided, that does not happen because it
would've already been copied before exec() in the child gets a chance
to execute.. the fork system call calls do_fork somewhere which calls
copy_process which does this copying so it can't be avoided in any
case. The book says copy-on-write itself has more overhead that is
avoided with exec() in the child, but I'm trying to figure how.
-Joel
Hi Joel...
Manish is right. Please notice that he talked about why do we do copy
on write (COW) if soon after child is forked, it quickly does exec().
So yes, COW has overhead, but imagine if parent ran first. COW will be
triggered for parent address space, then child soon runs too. Then it
issues exec(). Clearly, this waste certain amout of memory which can
be fairly avoided if child runs first.
After going through this thread, I just tried out the following simple
code:
int main (void) {
int pid;
int *testVar = (int *) malloc (sizeof (int));
*testVar = 10;
printf (%d [%d] Main \n, *testVar, testVar);
pid=vfork(); // works fine if we use fork instead.
if (pid==0) {
printf (Child %d [%d]\n, *testVar, testVar);
return 1;
} else if (pid 0) {
printf (Parent %d [%d]\n, *testVar, testVar);
*testVar=11; // segfault if we use vfork, as vfork blocks until child
returns call exec/exits.
wait(NULL);
printf (Parent %d [%d]\n, *testVar, testVar);
return 1;
}
exit(0);
}
can someone let me know why this segfaults with vfork and not with fork?
Man page says most of the stuff :-
The use of vfork() was tricky: for example, not modifying data
in the parent process depended on knowing which variables are held in
a register.
In particular, the programmer cannot rely on the parent remaining blocked
until the child either terminates or calls execve(2), and cannot
rely on any specific behavior with respect to shared memory.
Thanks -
Manish
From my understanding - it is because the parent is blocked until the
child exec's/exits AND in the mean time when the program is being
executed the child/parent process is trying to change the *testVar is
causing to modify the parents read-only memory. CMIAW.
Another thing I noticed is on linux the child always gets to run first
in case of fork() and vfork()?
Cheers
~Pete
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: compiling ext2 outside source tree
On Sun, Jan 10, 2010 at 10:37 PM, Onkar onkar@gmail.com wrote: Hi , I want to take the folder ext2 from linux-2.6.31.5/fs/ext2 compile it outside the tree and insmod ext2.ko. What changes should I make to the Makefile ? /tmp cp -r /home/mkatiyar/personal/kernel/linux-2.6.32.2/fs/ext2 . /tmp cd ext2 /tmp/ext2 make -C /lib/modules/`uname -r`/build M=$PWD make: Entering directory `/home/mkatiyar/personal/kernel/linux-2.6.32.2' LD /tmp/ext2/built-in.o CC [M] /tmp/ext2/balloc.o CC [M] /tmp/ext2/dir.o CC [M] /tmp/ext2/file.o CC [M] /tmp/ext2/ialloc.o CC [M] /tmp/ext2/inode.o CC [M] /tmp/ext2/ioctl.o CC [M] /tmp/ext2/namei.o CC [M] /tmp/ext2/super.o CC [M] /tmp/ext2/symlink.o CC [M] /tmp/ext2/xattr.o CC [M] /tmp/ext2/xattr_user.o CC [M] /tmp/ext2/xattr_trusted.o LD [M] /tmp/ext2/ext2.o Building modules, stage 2. MODPOST 1 modules CC /tmp/ext2/ext2.mod.o LD [M] /tmp/ext2/ext2.ko make: Leaving directory `/home/mkatiyar/personal/kernel/linux-2.6.32.2' Thanks - Manish Please help me get going with this. Regards, Onkar -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: copy-on-write overhead
On Mon, Jan 11, 2010 at 12:02 AM, Joel Fernandes agnel.j...@gmail.com wrote: Hi, The book Linux Kernel Development mentions a certain overhead which is reduced in linux because the child is made to run first after it is forked (instead of the parent) and if the child executes the exec system call. Quoting Pg 32 (2nd edition): In do_fork() the child is woken up and run. Deliberately the kernel runs the child process first. In the common case of the child simply calling exec() immediately, this eliminates any copy-on-write overhead that would occur if the parent ran first and began writing to the address space How does running exec() in child more efficient than having copy-on-write trigger in the parent? Hi, In order to create a new process (fork), you will copy the whole of relevant info of address space from the parent. Now if the address space is shared and parent runs first for any change it makes, it will need to copy and then modify because it doesn't know if child needs it or not. Since mostly after fork, if you do an exec it is guaranteed that child will not need this address space and the whole of copy can be avoided by making child to run first. (In simple terms you can take decision early whether this address space is useful or not for child). Thanks - Manish Let me know if my question is clear, or if I should rephrase. Thanks! -Joel -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Out of memory vs memory leak
On Thu, Jan 7, 2010 at 8:34 PM, Anand Raj Manickam anan...@gmail.com wrote:
I have a few doubts regarding Out of Memory .
1.Does memory leak in the kernel always lead to OOM ?
2. I wrote a kernel test module testing the memory leak , which on
exit kmallocs
exit_module ()
{
while(1)
kmalloc((8*1024),GFP_KERNEL);
}
so when i rmmod test , the kernel never panicked on OOM , but it
panicked on the page allocation stat (sorry for not attaching the
dump)
Does the kernel differentiate between a Memory Leak and a Out of
memory situation ?
You don't panic immediately. Memory leaks eventually leads to out of
memory because the memory manager has given all the memory to someone
and hasn't got back (leaked !!). So if a new requestor comes he
doesn't have anything to allocate and you get out of memory.
Thanks -
Manish
Google didnt help me in the releation between memory leak and OOM .
Thanks,
Anand
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: question about __v annotation
On Sat, Jan 2, 2010 at 6:40 PM, Shawn cit...@gmail.com wrote:
hello guys,
I got a newbie confused when I was looking into the source code of
s3c2440's RTC driver.I dont know what is __v excatly means
A local variable name.
..anyone can
tell?thanks anyway!
#define readb(c) ({ __u8 __v = __raw_readb(__mem_pci(c)); __v; })
--
GNU powered it...
GPL protect it...
God blessing it...
regards
HFG--Shawn the R0ck
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Kernel compilation problem.........
On Mon, Dec 21, 2009 at 11:07 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: Hi... On Tue, Dec 22, 2009 at 12:10 AM, bajrangi pandey bajrangi12pan...@gmail.com wrote: Hi , suppose i am making changes in kernel file fs.h Is it necessary to perform all three steps 1 make I think you can just do make fs 2 make modules_install 3 make install to rebuilt the kernel (compiling same kernel version source code 2nd time)? can i skip step make modules_install My best guess is yes, you need to do all the above step. The reason is to make sure all the related code pick your modification. You don't need to worry if compilation time is your concern. make will make sure only .c files that depends on this fs.h that get recompiled. So theoritically, it should be done relatively faster. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How does ls command work in Linux in detail?
On Thu, Dec 3, 2009 at 9:18 AM, Daniel (Youngwhan) Song
breadn...@gmail.com wrote:
Thanks, krushnaal and Greg,
By looking at the Greg's strace, it gives me great insight of the command.
So, It looks like it calls a library which can be in glib(?), and the
library actually calls kernel device drivers like console or something like
that, and they were communicating each other back and forth, and finally, it
outputs the result of ls onto the screen. Is my understanding correct?
Actually what you have told is true for any command and is not limited
to ls. If you want to know how ls works you will need to understand
a little about how files are stored on disk and how they are indexed.
And then you can try to read about readdir(), opendir(),stat() system
calls and finally implement your own ls :-)
Best Regards,
Daniel (Youngwhan) Song
On Wed, Dec 2, 2009 at 7:41 PM, Greg Freemyer greg.freem...@gmail.com
wrote:
On Wed, Dec 2, 2009 at 9:12 PM, Daniel (Youngwhan) Song
breadn...@gmail.com wrote:
Hi,
Could somebody explain me how exactly ls command work in Unix/Linux?
When we type ls command in linux shell, what does process/procedure
work
with linux library or linux kernel, and how exactly does it show
directory
information to the standard output (1)?
Thanks in advance.
Best Regards,
Daniel Song
read the source of course, but a very handy shortcut is:
# strace ls
The key part with a couple comments is:
# Get the directory entries
open(., O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY|O_CLOEXEC) = 3
fcntl64(3, F_GETFD) = 0x1 (flags FD_CLOEXEC)
getdents64(3, /* 53 entries */, 32768) = 1744
getdents64(3, /* 0 entries */, 32768) = 0
close(3) = 0
# Verify stdout is a character device
fstat64(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 0), ...}) = 0
# map stdin into memory. (Not sure why, see the source)
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0xb73ff000
# write the directory entries to stdout and wrap-up
write(1, bin Desktop Documents Downloa..., 91bin Desktop
Documents Download Music Pictures Public public_html Templates
Videos
) = 91
close(1) = 0
munmap(0xb73ff000, 4096) = 0
close(2) = 0
exit_group(0) = ?
Greg
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How does ls command work in Linux in detail?
On Thu, Dec 3, 2009 at 10:23 AM, Daniel (Youngwhan) Song
breadn...@gmail.com wrote:
Thanks, Greg and Manish,
What I am trying to understand is not only for ls itself. I just wanted to
know how general command such like ps, top, or something like that works
in linux. Also, I wanted to know the path from user-level application to how
it accesses to kernel resources.
I guess I could understand roughly about the mechanism of an application in
linux, but I feel to study more. If you know some resources I need to dig
into and share of it, it would be great helpful to me.
Hi Daniel,
Then you would need to learn about system calls and do google search
about how they work. Each of the unix commands call some set of system
calls underneath which actually does the job in kernel space. What you
see in the output of strace are the system calls made to kernel.
searching for system calls in google will give you plethora of results.
HTH
Thanks,
Best Regards,
Daniel (Youngwhan) Song
On Wed, Dec 2, 2009 at 8:36 PM, Manish Katiyar mkati...@gmail.com wrote:
On Thu, Dec 3, 2009 at 9:18 AM, Daniel (Youngwhan) Song
breadn...@gmail.com wrote:
Thanks, krushnaal and Greg,
By looking at the Greg's strace, it gives me great insight of the
command.
So, It looks like it calls a library which can be in glib(?), and the
library actually calls kernel device drivers like console or something
like
that, and they were communicating each other back and forth, and
finally, it
outputs the result of ls onto the screen. Is my understanding correct?
Actually what you have told is true for any command and is not limited
to ls. If you want to know how ls works you will need to understand
a little about how files are stored on disk and how they are indexed.
And then you can try to read about readdir(), opendir(),stat() system
calls and finally implement your own ls :-)
Best Regards,
Daniel (Youngwhan) Song
On Wed, Dec 2, 2009 at 7:41 PM, Greg Freemyer greg.freem...@gmail.com
wrote:
On Wed, Dec 2, 2009 at 9:12 PM, Daniel (Youngwhan) Song
breadn...@gmail.com wrote:
Hi,
Could somebody explain me how exactly ls command work in
Unix/Linux?
When we type ls command in linux shell, what does process/procedure
work
with linux library or linux kernel, and how exactly does it show
directory
information to the standard output (1)?
Thanks in advance.
Best Regards,
Daniel Song
read the source of course, but a very handy shortcut is:
# strace ls
The key part with a couple comments is:
# Get the directory entries
open(., O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY|O_CLOEXEC) = 3
fcntl64(3, F_GETFD) = 0x1 (flags FD_CLOEXEC)
getdents64(3, /* 53 entries */, 32768) = 1744
getdents64(3, /* 0 entries */, 32768) = 0
close(3) = 0
# Verify stdout is a character device
fstat64(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 0), ...}) = 0
# map stdin into memory. (Not sure why, see the source)
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0xb73ff000
# write the directory entries to stdout and wrap-up
write(1, bin Desktop Documents Downloa..., 91bin Desktop
Documents Download Music Pictures Public public_html Templates
Videos
) = 91
close(1) = 0
munmap(0xb73ff000, 4096) = 0
close(2) = 0
exit_group(0) = ?
Greg
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Query on File Creation time in Linux
On Thu, Nov 19, 2009 at 5:59 PM, Indraneel Mukherjee indro...@gmail.com wrote:
Hi,
It seems that Linux does NOT provide a File Creation Time field.
There is atime(file access time) ctime(inode change time) mtime(file
modification time) but NO field for File Creation Time. Why is it so?
Actually if you see the comments in ext2_inode structure in kernel
sources i_ctime says it is creation time. However everywhere it is
changed frequently along with the mtime, so from the code looks like
c stands for change time and not creation time.
242 struct ext2_inode {
243 __le16 i_mode; /* File mode */
244 __le16 i_uid; /* Low 16 bits of Owner Uid */
245 __le32 i_size; /* Size in bytes */
246 __le32 i_atime;/* Access time */
247 __le32 i_ctime;/* Creation time */
A bit of googling showed that creation time is supported by some
non-posix filesystems.
Regards,
Indro
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: kernel panic about kernel unaligned access
On Thu, Oct 29, 2009 at 11:13 AM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On Thu, Oct 29, 2009 at 10:48 AM, loody milo...@gmail.com wrote: Hi: thanks for all your kind help. Take 8709ed20 writeback_inodes+0xb4/0x160 for example, what does 0x160, the last hex mean? The value of parameter? 0x160 means the total size of your disassembled function and 0xb4 is the offset in it where it crashed. If you do disas writeback_inodes you should look around (writeback_inodes+0xb4) to find the instruction where it crashed. appreciate your help, miloody It's the size (hexadecimal number) of the related function (in bytes, I think) -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How to print Superblock(ext2) data
On Wed, Oct 21, 2009 at 1:24 PM, mahesh sobale msobale...@gmail.com wrote: HI, I am new to kernelnewbies. I am trying to print meta data in kernel space for superblock. for that i have to fire ioctl and write char device driver(This part i know), But i don't know how create loop device and mount ext2 file system on that and then print superblock info. You can just do dumpe2fs -h device_name Can anybody help on this. Thanx. Mahesh. -- Thanks - Manish == [$\*.^ -- I miss being one of them == -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Mounting block device fails.
On Fri, Oct 9, 2009 at 11:16 AM, Prasad Joshi prasadjoshi...@gmail.com wrote:
Thanks Manish,
The issue was with filesystem creation. It was not getting created
properly. I used dumpe2fs command to verify the file system. Actually
Superblock was written correctly but, few other fields like free block
and free inode list etc were incorrect
It seems to be working properly now
# dumpe2fs /dev/ftl
...
...
Group 0: (Blocks 0-65527)
Primary superblock at 0, Group descriptors at 1-1
Reserved GDT blocks at 2-8
Block bitmap at 9 (+9), Inode bitmap at 10 (+10)
Inode table at 11-1034 (+11)
64489 free blocks, 32757 free inodes, 2 directories
Free blocks: 1039-65527
Free inodes: 12-32768
Group 1: (Blocks 65528-131055)
Backup superblock at 65528, Group descriptors at 65529-65529
Reserved GDT blocks at 65530-65536
Block bitmap at 65537 (+9), Inode bitmap at 65538 (+10)
Inode table at 65539-66562 (+11)
64493 free blocks, 32768 free inodes, 0 directories
Free blocks: 66563-131055
Free inodes: 32769-65536
# mount
...
...
/dev/ftl on /mntpt type ext2 (rw,relatime,errors=continue)
The other problem which I am facing is with the blocksize.
mkfs allows ext2 file system be created with blocksize=8192,
overriding the default 4096. But, when I try to mount the file system
it fails
# mkfs -t ext2 -b 8192 /dev/ftl
Warning: blocksize 8192 not usable on most systems.
mke2fs 1.41.4 (27-Jan-2009)
mkfs.ext2: 8192-byte blocks too big for system (max 4096)
Proceed anyway? (y,n) y
Warning: 8192-byte blocks too big for system (max 4096), forced to continue
warning: 16 blocks unused.
Filesystem label=
OS type: Linux
Block size=8192 (log=3)
Fragment size=8192 (log=3)
65536 inodes, 131056 blocks
# mount -t ext2 /dev/ftl /mntpt
[54658.092031] EXT2-fs: blocksize too small for device.
mount: mounting /dev/ftl on /mntpt failed: Invalid argument
I tried specifying blocksize with the mount command, but it also fails
# mount -t ext2 /dev/ftl /mntpt -o blocksize=8192
mount: mounting /dev/ftl on /mntpt failed: Invalid argument
The above command fails in parsing the options passed to the command
The section Mount options for ext2 of mount manpage does not mention
blocksize option.
Yes, there is no blocksize option in ext2. For a list of valid
options see tokens table in ext2/super.c
Is there anyway I can create ext2 filesystem with blocksize 8192?
No ... as far as I know.
Man page of mke2fs says :
-b block-size
Specify the size of blocks in bytes. Valid block-size
values are 1024, 2048 and 4096 bytes per block.
And the kernel code says :
int sb_set_blocksize(struct super_block *sb, int size)
{
if (set_blocksize(sb-s_bdev, size))
return 0;
/* If we get here, we know size is power of two
* and it's value is between 512 and PAGE_SIZE */
sb-s_blocksize = size;
sb-s_blocksize_bits = blksize_bits(size);
return sb-s_blocksize;
}
and the PAGE_SIZE is 4096.
Reiserfs supports blocksizes of 8192 , but its man page says :-
-b | --block-size N
N is block size in bytes. It may only be set to a power
of 2 within the 512-8192 interval. Note
that current versions of the kernel ( 2.6.19 ) do NOT
support any size other than 4096.
But is there any specific reason you want to use block sizes of 8192 ?
Thanks and Regards,
Prasad
On Thu, Oct 8, 2009 at 6:41 PM, Manish Katiyar mkati...@gmail.com wrote:
On Thu, Oct 8, 2009 at 12:38 PM, Prasad Joshi prasadjoshi...@gmail.com
wrote:
Hi,
I am working on translation layer for flash. For the purpose testing I am
using a disk file to simulate the flash behavior. So every read and write on
the block device finally goes to a disk file, something similar to loop
device, but in addition maintaining the flash property.
I am able to create the file system on the device (/dev/ftl) but when I try
to mount the device it fails with error Invalid Argument.
From the code I can see -EINVAL can be returned at lot of places. You
need to see dmesg to see the message which can help tracing.
I thought it is failing to read the super block of the file systems, but it
seems like it did not fail in superblock read
# strace -o out mount -t ext2 /dev/ftl /mntpt/
[ 175.192033] Opening the device.
[ 175.193599] IftlIoRead: offset=0x400 Length=0x400
[ 175.193875] OffsetWithinPage: 400
[ 175.195104] Read pos=1002535424, length=8192
[ 175.195474] 0001 0004 eea5 0003 fff5
0002 0002 8000 8000 2000 8a9e 4acd
0020 ef53 0001 0001 8a9e 4acd 4e00 00ed 0001
000b 0100 0038 0002 0003 a037 8e0c d6b3 de11 abb8
0635 bff9 2b36
Re: Mounting block device fails.
On Thu, Oct 8, 2009 at 12:38 PM, Prasad Joshi prasadjoshi...@gmail.com wrote:
Hi,
I am working on translation layer for flash. For the purpose testing I am
using a disk file to simulate the flash behavior. So every read and write on
the block device finally goes to a disk file, something similar to loop
device, but in addition maintaining the flash property.
I am able to create the file system on the device (/dev/ftl) but when I try
to mount the device it fails with error Invalid Argument.
From the code I can see -EINVAL can be returned at lot of places. You
need to see dmesg to see the message which can help tracing.
I thought it is failing to read the super block of the file systems, but it
seems like it did not fail in superblock read
# strace -o out mount -t ext2 /dev/ftl /mntpt/
[ 175.192033] Opening the device.
[ 175.193599] IftlIoRead: offset=0x400 Length=0x400
[ 175.193875] OffsetWithinPage: 400
[ 175.195104] Read pos=1002535424, length=8192
[ 175.195474] 0001 0004 eea5 0003 fff5
0002 0002 8000 8000 2000 8a9e 4acd
0020 ef53 0001 0001 8a9e 4acd 4e00 00ed 0001
000b 0100 0038 0002 0003 a037 8e0c d6b3 de11 abb8
0635 bff9 2b36
003f
a037 8f0c d6b3 de11 abb8 0635 bff9 2b36 0001
8a9e 4acd
001c 001c 0001
[ 175.215929] Closing the
device.
mount: mounting /dev/ftl on /mntpt/ failed: Invalid argument
I tried printing the data in hex format just to make sure proper ext2 magic
code in present in the superblock read from the flash device. ef53 is the
magic code of ext2 superblock, ftl device is correctly reading the magic
code
# cat out
execve(/bin/mount, [mount, -t, ext2, /dev/ftl, /mntpt/], [/* 4
vars */]) = 0
uname({sys=Linux, node=(none), ...}) = 0
brk(0) = 0x72f000
brk(0x72ff40) = 0x72ff40
arch_prctl(ARCH_SET_FS, 0x72f880) = 0
open(/dev/urandom, O_RDONLY) = -1 ENOENT (No such file or
directory)
brk(0x750f40) = 0x750f40
brk(0x751000) = 0x751000
getuid() = 0
getuid() = 0
geteuid() = 0
stat(/dev/ftl, {st_mode=S_IFBLK|0644, st_rdev=makedev(254, 0), ...}) = 0
mount(/dev/ftl, /mntpt/, ext2, MS_SILENT, ) = -1 EINVAL (Invalid
argument)
vfork() = 869
--- SIGCHLD (Child exited) @ 0 (0) ---
write(2, mount: mounting /dev/ftl on /mntp..., 61) = 61
exit_group(-1) = ?
What could be the reason of failure of mount?
Thanks and Regards,
Prasad
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: C question
On Thu, Oct 8, 2009 at 7:42 AM, Rick Brown rick.brow...@gmail.com wrote:
Hello list,
As far as I recall from KR, isn't pointer arithmetic on a void
pointer banned? And any effort to do that results in an error -
because the compiler won't know by how much size to increment the
pointer for a statement like ptr++? But then how about this:
But in the program, you aren't actually trying to dereference the
value. Just adding means it becomes normal arithmetic and that is why
you get result as 1. You will see the error if you try to dereference
it.
/tmp gcc a.c
a.c: In function ‘main’:
a.c:5: warning: format ‘%d’ expects type ‘int’, but argument 2 has type ‘void *’
a.c:6: warning: dereferencing ‘void *’ pointer
a.c:6: error: invalid use of void expression
/tmp cat a.c
#include stdio.h
int main()
{
void *ptr = 0;
printf(%d \n, ptr+1);
printf(%d \n, *(ptr+1));
}
[r...@linux rick]$ cat t.c
#include stdio.h
int main()
{
void *ptr = 0;
printf(%d \n, ptr+1);
}
[r...@linux rick]$ gcc t.c
[r...@linux rick]$ ./a.out
1
[r...@linux rick]$
It compiles and runs fine ... !
TIA,
Rick
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
==
[$\*.^ -- I miss being one of them
==
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Per cpu data
On Fri, Aug 28, 2009 at 1:17 PM, Pharaoh . pharaoh...@gmail.com wrote: On Thu, Aug 27, 2009 at 12:02 PM, Pharaoh . pharaoh...@gmail.com wrote: On Thu, Aug 27, 2009 at 12:34 AM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On Wed, Aug 26, 2009 at 11:06 PM, Pharaoh .pharaoh...@gmail.com wrote: Hi list, 1. A cpu can access its own per cpu data atomically but if it has to access other cpus per cpu data it might not be atomically accessed. Is this correct? IMO, assuming your per CPU data is something simple like word length (4 byte in x86 32 bit), then yes local access to per CPU data is atomic. But it can not be atomic if it's accessing complex data structure. And IIRC, per CPU API are written to store and retrieve simple type of data. If you're accessing other CPU data, I guess it still might done atomically by doing lock (e.g spinlock). But yes, without lock, you have to careful about race condition If my driver makes sure that no cpu accesses other cpu's per cpu data my code can be entirely lockless? Correct? IMO yes. Just make sure you use provided per CPU API. IIRC, when you're accessing per CPU data using the APIs, it will also disable preemption..thus preventing your code to be migrated to other core/CPU. 2. Can I have a rather complex data structure like a radix tree as a per cpu data structure? Ehm, I don't think so. Well, theoritically you can but IIRC the kernel doesn't provide such API to store and retrieve such data structure by default. 3. Is it guaranteed that all the accesses happening to this tree from a cpu which allocated it are atomic? I think they are atomic since the per cpu data access functions disable preemption. My main objective is to have a radix tree which will be accessed on multiple cores, I want to avoid the locking overhead of keeping a central lock for accessing this tree. I have just started investigating the possibilities. Hi, I was reading about lockless pagecache and seems it tries to do what you want. You can have a look at its code. See the link ftp://ftp.kernel.org/pub/linux/kernel/people/npiggin/patches/lockless/2.6.16-rc5/radix-intro.pdf . Also search google for lockless pagecache protocol Thanks - Manish -pharaoh. -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com More generically put, how is a highly concurrent data structure designed in kernel space? If I keep a central lock for accessing the radix tree, wont it cause too much latency? I am sure there must be some other way which I am not aware of. Using read/write locks might be better idea so that atleast reads can happen concurrently. E.g. userspace concurrent programs have a concept of thread local storage, so each thread acccesses its own copy. If Linux is able to run on hundreds of cores, I am sure there must be a way to an optimal way to access these concurrent data structures. -pharaoh. How about RCU mechanism? I think using this I can design an efficient concurrently accessible rb tree. Any thoughts? -phraoh. -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Search module in linux
On Sat, Oct 3, 2009 at 8:29 PM, Deepak Mishra dpux4li...@gmail.com wrote: Hello everyone, I am very new to this list and kernel programming. Maybe this is too early, but I am actually interested in knowing the file search module (inherently the source code for find program) in linux Hi Deepak, find is a userspace program and there is no kernel component in it. filesystem. I believe there would be seperate search module for every filesytem. The only filesystem specific code which find needs to call is how to traverse directories, and how to get the file names given that each filesystem will have its own layout. However this is abstracted using the syscall readdir(). so you don't need to worry about the underlying layout of the filesystem. You can look at ext2_readdir() for example . I feel looking at the codes may help me appreciate the filesystems better. I searched the kernel tree and even googled for such code, but I feel I am yet not proficient enough to get anything. Hope you guys can help me out. You can try to write your own find program in userspace. See the usage of following syscalls. readdir(), opendir() Best of luck. Thanks - Manish Thanks in anticipation, Deepak Mishra -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How to build DocBook in Linux kernel?
On Wed, Sep 30, 2009 at 4:13 PM, NAHieu nah...@gmail.com wrote: Hi, There are a lot of DocBook files under Documentation/DocBook/. I tried to build them with make xmldocs, but that only produces .xml file, which cannot be viewed in Firefox. make pdfdocs So how to view those resulted .xml file, or if I want to buid those DocBook to .pdf files, what should I do? Thanks, Hieu -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: ext4 extents: How to determine if an extent points to a hole.
On Mon, Sep 14, 2009 at 11:59 AM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi Greg, Thanks a lot for such nice explanation. But I still have the same query lingering... On Mon, Sep 14, 2009 at 9:36 AM, Greg Freemyer greg.freem...@gmail.com wrote: On Sun, Sep 13, 2009 at 1:49 PM, SandeepKsinha sandeepksi...@gmail.com wrote: On Sun, Sep 13, 2009 at 11:08 PM, Manish Katiyar mkati...@gmail.com wrote: On Sun, Sep 13, 2009 at 10:12 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi, On Sun, Sep 13, 2009 at 10:07 PM, Manish Katiyar mkati...@gmail.com wrote: On Sun, Sep 13, 2009 at 5:44 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi all, Looking at the flags in the extent info, Is there any specifc flags which indicates an extent to be a HOLE?? I am not sure if I understand the question correctly .. why would you need that ? Can you give an example where it should be used ?? Look at e4defrag.c, it checks the file size and allocates the same number of blocks for donor inode. Which will eventually make a holey file into a normal one. Any tool/application should make sure that they leave a sparse file as sparse. I think, as suggested by Greg Freemyer, we can use BMAP ioctl to get such information. Yes, but I think bmap would be costly if the file is large and is not holey :-( but that would be probably same calling fiemap if the file is completely fragmented such that each extent size is 1. Since, ext2/ext3 did not have mutli block allocation thats why this is the only way that we might have. But generally most of the new features work on with extent based files on ext4. I am still wondering that how to we represent a hole using extents in a extent based file. Just like we had a convention of having the block number 0 in case of holes. Similarly, what do we look at to figure out if its a hole or not. BMAP is one way. But since, in a extent based file, we have only extents, there should be some flag to indicate the same. Sandeep, If you look at e4defrag, it first gets a list of all the extents. I'm pretty sure extents only exist for allocated extents. Holes do not have any associated extents. I did a extent dump of a holey file. Inode: 12 Type: regularMode: 0644 Flags: 0x8 Generation: 4284390079Version: 0x0001 User: 5572 Group: 5573 Size: 20877312000 File ACL: 0Directory ACL: 0 Links: 1 Blockcount: 16000 Fragment: Address: 0Number: 0Size: 0 ctime: 0x4aae0975 -- Mon Sep 14 14:44:29 2009 atime: 0x4aad34e9 -- Sun Sep 13 23:37:37 2009 mtime: 0x4aae0975 -- Mon Sep 14 14:44:29 2009 EXTENTS: (0-3999): 10241-14240, (20384000-20387999): 14337-18336 Level Entries LogicalPhysical Length Flags 0/ 0 1/ 20 - 3999 10241 - 14240 4000 0/ 0 2/ 2 20384000 - 20387999 14337 - 18336 4000 I am looking at the source, but as you can see it is not printing the extents for the hole part. This is what my actual question is, think of this situation... |-|-|-|--| 0 500 700 1200 3000 [Logical block numbers for an inode] In this situation you will have four extents for sure. ext_1 - 0 -- 500 ext_2 - 501 --700 [ This will be an initialized extent] ext_3 - 701 -- 1200 ext_4 - 1200 -- 3000 After looking at the sources and some comments in the ext4 source code, I could figure out that holes would be having an initialized extent. Reference: http://lxr.linux.no/#linux+v2.6.31/fs/ext4/extents.c#L2843 I think we cannot have a mixture of both a BMAP and an EMAP, it will be either. Then it calls merge_extents to create extent groups. In e4defrag terminology, an extent group is a collection of all the logically contiguous extents. I don't know if the kernel uses that terminology or not. Hope they are not merging together any initialized and uninitialized extents together, since they can be logically contiguous. Or rather they will be. In other words in e4defrag terminology a sparse file is a series of: extent group - hole - extent group - hole - extent group - etc. Then e4defrag creates a donor file with exactly the same allocated block areas by calling fallocate on the donor file for each extent group with the same starting offset and length as the extent group. This is true and should be applicable to initialized extents as well. I fear if they are Thus the donor file ends up have exactly the same holes as the original file. Then the donor blocks are used to defrag the original file by calling move_extent. In the kernel, the move_extent logic looks for holes and only replaces blocks that are allocated in the original file. This is true. I am sure of the kernel logic. Greg -- Regards, Sandeep. “To learn is to change. Education is a process that changes the learner.” -- Thanks - Manish
Re: ext4 extents: How to determine if an extent points to a hole.
On Mon, Sep 14, 2009 at 3:09 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi Manish, On Mon, Sep 14, 2009 at 2:57 PM, Manish Katiyar mkati...@gmail.com wrote: On Mon, Sep 14, 2009 at 11:59 AM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi Greg, Thanks a lot for such nice explanation. But I still have the same query lingering... On Mon, Sep 14, 2009 at 9:36 AM, Greg Freemyer greg.freem...@gmail.com wrote: On Sun, Sep 13, 2009 at 1:49 PM, SandeepKsinha sandeepksi...@gmail.com wrote: On Sun, Sep 13, 2009 at 11:08 PM, Manish Katiyar mkati...@gmail.com wrote: On Sun, Sep 13, 2009 at 10:12 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi, On Sun, Sep 13, 2009 at 10:07 PM, Manish Katiyar mkati...@gmail.com wrote: On Sun, Sep 13, 2009 at 5:44 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi all, Looking at the flags in the extent info, Is there any specifc flags which indicates an extent to be a HOLE?? I am not sure if I understand the question correctly .. why would you need that ? Can you give an example where it should be used ?? Look at e4defrag.c, it checks the file size and allocates the same number of blocks for donor inode. Which will eventually make a holey file into a normal one. Any tool/application should make sure that they leave a sparse file as sparse. I think, as suggested by Greg Freemyer, we can use BMAP ioctl to get such information. Yes, but I think bmap would be costly if the file is large and is not holey :-( but that would be probably same calling fiemap if the file is completely fragmented such that each extent size is 1. Since, ext2/ext3 did not have mutli block allocation thats why this is the only way that we might have. But generally most of the new features work on with extent based files on ext4. I am still wondering that how to we represent a hole using extents in a extent based file. Just like we had a convention of having the block number 0 in case of holes. Similarly, what do we look at to figure out if its a hole or not. BMAP is one way. But since, in a extent based file, we have only extents, there should be some flag to indicate the same. Sandeep, If you look at e4defrag, it first gets a list of all the extents. I'm pretty sure extents only exist for allocated extents. Holes do not have any associated extents. I did a extent dump of a holey file. Inode: 12 Type: regular Mode: 0644 Flags: 0x8 Generation: 4284390079 Version: 0x0001 User: 5572 Group: 5573 Size: 20877312000 File ACL: 0 Directory ACL: 0 Links: 1 Blockcount: 16000 Fragment: Address: 0 Number: 0 Size: 0 ctime: 0x4aae0975 -- Mon Sep 14 14:44:29 2009 atime: 0x4aad34e9 -- Sun Sep 13 23:37:37 2009 mtime: 0x4aae0975 -- Mon Sep 14 14:44:29 2009 EXTENTS: (0-3999): 10241-14240, (20384000-20387999): 14337-18336 Level Entries Logical Physical Length Flags 0/ 0 1/ 2 0 - 3999 10241 - 14240 4000 0/ 0 2/ 2 20384000 - 20387999 14337 - 18336 4000 I am looking at the source, but as you can see it is not printing the extents for the hole part. I suspect the debugfs code masking that out for the output. yes... and that is the flag/method that we are looking for ... isn't it :-) ??? If you remember, for a normal block based file also, it does something similar, it does not prints the values for holes. It can be a case here as well. This is what my actual question is, think of this situation... |-|-|-|--| 0 500 700 1200 3000 [Logical block numbers for an inode] In this situation you will have four extents for sure. ext_1 - 0 -- 500 ext_2 - 501 --700 [ This will be an initialized extent] ext_3 - 701 -- 1200 ext_4 - 1200 -- 3000 After looking at the sources and some comments in the ext4 source code, I could figure out that holes would be having an initialized extent. Reference: http://lxr.linux.no/#linux+v2.6.31/fs/ext4/extents.c#L2843 I think we cannot have a mixture of both a BMAP and an EMAP, it will be either. Then it calls merge_extents to create extent groups. In e4defrag terminology, an extent group is a collection of all the logically contiguous extents. I don't know if the kernel uses that terminology or not. Hope they are not merging together any initialized and uninitialized extents together, since they can be logically contiguous. Or rather they will be. In other words in e4defrag terminology a sparse file is a series of: extent group - hole - extent group - hole - extent group - etc. Then e4defrag creates a donor file with exactly the same allocated block areas by calling fallocate on the donor file for each extent group with the same starting offset and length as the extent group. This is true and should be applicable to initialized extents as well. I fear
Re: ext4 extents: How to determine if an extent points to a hole.
On Sun, Sep 13, 2009 at 5:44 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi all, Looking at the flags in the extent info, Is there any specifc flags which indicates an extent to be a HOLE?? I am not sure if I understand the question correctly .. why would you need that ? Can you give an example where it should be used ?? I could only find the following ones... 39#define FIEMAP_MAX_OFFSET (~0ULL) 40 41#define FIEMAP_FLAG_SYNC 0x0001 /* sync file data before map */ 42#define FIEMAP_FLAG_XATTR 0x0002 /* map extended attribute tree */ 43 44#define FIEMAP_FLAGS_COMPAT (FIEMAP_FLAG_SYNC | FIEMAP_FLAG_XATTR) 45 46#define FIEMAP_EXTENT_LAST 0x0001 /* Last extent in file. */ 47#define FIEMAP_EXTENT_UNKNOWN 0x0002 /* Data location unknown. */ 48#define FIEMAP_EXTENT_DELALLOC 0x0004 /* Location still pending. 49 * Sets EXTENT_UNKNOWN. */ 50#define FIEMAP_EXTENT_ENCODED 0x0008 /* Data can not be read 51 * while fs is unmounted */ 52#define FIEMAP_EXTENT_DATA_ENCRYPTED 0x0080 /* Data is encrypted by fs. 53 * Sets EXTENT_NO_BYPASS. */ 54#define FIEMAP_EXTENT_NOT_ALIGNED 0x0100 /* Extent offsets may not be 55 * block aligned. */ 56#define FIEMAP_EXTENT_DATA_INLINE 0x0200 /* Data mixed with metadata. 57 * Sets EXTENT_NOT_ALIGNED.*/ 58#define FIEMAP_EXTENT_DATA_TAIL 0x0400 /* Multiple files in block. 59 * Sets EXTENT_NOT_ALIGNED.*/ 60#define FIEMAP_EXTENT_UNWRITTEN 0x0800 /* Space allocated, but 61 * no data (i.e. zero). */ 62#define FIEMAP_EXTENT_MERGED 0x1000 /* File does not natively 63 * support extents. Result 64 * merged for efficiency. */ -- Regards, Sandeep. “To learn is to change. Education is a process that changes the learner.” -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: e4defrag: file extent map : redundant loop???
On Sun, Sep 13, 2009 at 5:26 PM, SandeepKsinha sandeepksi...@gmail.com wrote:
Hi all,
Here is a code snippet from e4defrag.c, can be located @
http://git.kernel.org/?p=fs/ext2/e2fsprogs.git;a=blob_plain;f=misc/e4defrag.c;hb=next
/*
* get_file_extents() - Get file's extent list.
*
* @fd: defrag target file's descriptor.
* @ext_list_head: the head of the extent list.
*/
static int get_file_extents(int fd, struct fiemap_extent_list **ext_list_head)
{
__u32 i;
int ret;
int ext_buf_size, fie_buf_size;
__u64 pos = 0;
struct fiemap *fiemap_buf = NULL;
struct fiemap_extent *ext_buf = NULL;
struct fiemap_extent_list *ext_list = NULL;
/* Convert units, in bytes.
* Be careful : now, physical block number in extent is 48bit,
* and the maximum blocksize for ext4 is 4K(12bit),
* so there is no overflow, but in future it may be changed.
*/
/* Alloc space for fiemap */
ext_buf_size = EXTENT_MAX_COUNT * sizeof(struct fiemap_extent);
fie_buf_size = sizeof(struct fiemap) + ext_buf_size;
fiemap_buf = malloc(fie_buf_size);
if (fiemap_buf == NULL)
return -1;
ext_buf = fiemap_buf-fm_extents;
memset(fiemap_buf, 0, fie_buf_size);
fiemap_buf-fm_length = FIEMAP_MAX_OFFSET;
fiemap_buf-fm_flags |= FIEMAP_FLAG_SYNC;
fiemap_buf-fm_extent_count = EXTENT_MAX_COUNT;
do {
fiemap_buf-fm_start = pos;
memset(ext_buf, 0, ext_buf_size);
ret = ioctl(fd, FS_IOC_FIEMAP, fiemap_buf);
if (ret 0)
goto out;
for (i = 0; i fiemap_buf-fm_mapped_extents; i++) {
ext_list = NULL;
ext_list = malloc(sizeof(struct fiemap_extent_list));
if (ext_list == NULL)
goto out;
ext_list-data.physical = ext_buf[i].fe_physical
/ block_size;
ext_list-data.logical = ext_buf[i].fe_logical
/ block_size;
ext_list-data.len = ext_buf[i].fe_length
/ block_size;
ret = insert_extent_by_physical(
ext_list_head, ext_list);
if (ret 0) {
FREE(ext_list);
goto out;
}
}
/* Record file's logical offset this time */
pos = ext_buf[EXTENT_MAX_COUNT-1].fe_logical +
ext_buf[EXTENT_MAX_COUNT-1].fe_length;
/*
* If fm_extents array has been filled and
* there are extents left, continue to cycle.
*/
} while (fiemap_buf-fm_mapped_extents
== EXTENT_MAX_COUNT
!(ext_buf[EXTENT_MAX_COUNT-1].fe_flags
FIEMAP_EXTENT_LAST));
FREE(fiemap_buf);
return 0;
out:
FREE(fiemap_buf);
return -1;
}
Why do we have a while loop here, we have already made sure that we
are initially itself allocating memory for EXTENT_MAX_COUNT?
This is the maximum number of extents supported by the file system.
Sandeep,
EXTENT_MAX_COUNT doesn't look to be the max extents supported by the
filesystem. It is the max chunk supported by e4defrag. Max supported
is FIEMAP_MAX_EXTENTS which is much larger.
Is
this loop not redundant ?
No, the loop just tries to get more extents as long as you haven't hit
the last one.
--
Regards,
Sandeep.
“To learn is to change. Education is a process that changes the learner.”
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: ext4 extents: How to determine if an extent points to a hole.
On Sun, Sep 13, 2009 at 10:12 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi, On Sun, Sep 13, 2009 at 10:07 PM, Manish Katiyar mkati...@gmail.com wrote: On Sun, Sep 13, 2009 at 5:44 PM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi all, Looking at the flags in the extent info, Is there any specifc flags which indicates an extent to be a HOLE?? I am not sure if I understand the question correctly .. why would you need that ? Can you give an example where it should be used ?? Look at e4defrag.c, it checks the file size and allocates the same number of blocks for donor inode. Which will eventually make a holey file into a normal one. Any tool/application should make sure that they leave a sparse file as sparse. I think, as suggested by Greg Freemyer, we can use BMAP ioctl to get such information. Yes, but I think bmap would be costly if the file is large and is not holey :-( but that would be probably same calling fiemap if the file is completely fragmented such that each extent size is 1. I could only find the following ones... 39#define FIEMAP_MAX_OFFSET (~0ULL) 40 41#define FIEMAP_FLAG_SYNC 0x0001 /* sync file data before map */ 42#define FIEMAP_FLAG_XATTR 0x0002 /* map extended attribute tree */ 43 44#define FIEMAP_FLAGS_COMPAT (FIEMAP_FLAG_SYNC | FIEMAP_FLAG_XATTR) 45 46#define FIEMAP_EXTENT_LAST 0x0001 /* Last extent in file. */ 47#define FIEMAP_EXTENT_UNKNOWN 0x0002 /* Data location unknown. */ 48#define FIEMAP_EXTENT_DELALLOC 0x0004 /* Location still pending. 49 * Sets EXTENT_UNKNOWN. */ 50#define FIEMAP_EXTENT_ENCODED 0x0008 /* Data can not be read 51 * while fs is unmounted */ 52#define FIEMAP_EXTENT_DATA_ENCRYPTED 0x0080 /* Data is encrypted by fs. 53 * Sets EXTENT_NO_BYPASS. */ 54#define FIEMAP_EXTENT_NOT_ALIGNED 0x0100 /* Extent offsets may not be 55 * block aligned. */ 56#define FIEMAP_EXTENT_DATA_INLINE 0x0200 /* Data mixed with metadata. 57 * Sets EXTENT_NOT_ALIGNED.*/ 58#define FIEMAP_EXTENT_DATA_TAIL 0x0400 /* Multiple files in block. 59 * Sets EXTENT_NOT_ALIGNED.*/ 60#define FIEMAP_EXTENT_UNWRITTEN 0x0800 /* Space allocated, but 61 * no data (i.e. zero). */ 62#define FIEMAP_EXTENT_MERGED 0x1000 /* File does not natively 63 * support extents. Result 64 * merged for efficiency. */ -- Regards, Sandeep. “To learn is to change. Education is a process that changes the learner.” -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- Regards, Sandeep. “To learn is to change. Education is a process that changes the learner.” -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: decoding kernel oops
On Thu, Sep 3, 2009 at 6:45 AM, Mulyadi Santosamulyadi.sant...@gmail.com wrote: On Wed, Sep 2, 2009 at 6:08 PM, Manish Katiyarmkati...@gmail.com wrote: Hi, I was going through this post http://kerneltrap.org/mailarchive/linux-kernel/2008/1/8/546623 and couldn't understand the steps around the below paragraph. and run it under gdb, and then when it gets the SIGSEGV (due to the obvious NULL pointer dereference), I can just ask gdb to disassemble around the array that contains the code[] stuff. Try a few offsets, to see when the disassembly makes sense (and gives the reported EIP as the beginning of one of the disassembled instructions). Has anyone tried this ?? I don't see any useful disas instructions when I do so ... Need help in interpreting the procedures of the above pos Hi Manish... I am not so clear either, but perhaps by exchanging each of our knowledge, we can teach each other. Specificly, which part that confuses you? The disas part? That's what the x/20i does toward the start address of array array. It turns that sequence of hex numbers into instructions. I think the real art here is spotting the offending instruction. I think Linus find it out actually by lookin that there is prefetch instruction in array+49. And since it's prefetching at the address stored in %eax, so the last instruction that modifies its content is surely the root of the problem. In the other cases, decoding might be not that easy. So IMO that's why stack trace provides code context. It's not just the offending instruction which is dumped, but ones sorrounding it. thanks a lot.. now if I go back and read the post again, it is much more clear. BTW probably a stupid question, why does the op code in char string generates back the assembly ?? -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
decoding kernel oops
Hi, I was going through this post http://kerneltrap.org/mailarchive/linux-kernel/2008/1/8/546623 and couldn't understand the steps around the below paragraph. and run it under gdb, and then when it gets the SIGSEGV (due to the obvious NULL pointer dereference), I can just ask gdb to disassemble around the array that contains the code[] stuff. Try a few offsets, to see when the disassembly makes sense (and gives the reported EIP as the beginning of one of the disassembled instructions). Has anyone tried this ?? I don't see any useful disas instructions when I do so ... Need help in interpreting the procedures of the above post. -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: ext3 structure mapping to block device
On Mon, Aug 17, 2009 at 11:06 AM, ajit motemail2black...@gmail.com wrote: Hello, I would like to know how file system structures are mapped to hard disk sector. Each filesystem deals in the granularity of block size which is generally configurable (typically 4K) during fs creation time. So each block maps to 8 sectors (which are generally 512 bytes). Now given a block number you can multiply it by 8 to get your hard disk sector. My goal is to understand how file system is created on block device and how superblocks are mapped to hard disk sectors. Filesystem is created using a userspace binary (mke2fs for example in case of ext2). Appropriate datastructures are written at appropiate places during the FS creation time. See the sources of mkfs.c in e2fsprogs. Superblock will have a predefined block number where it resides. From the above calculation you can know which sector to read. I am not sure whether I stated the problem precisely or not but if I did not described it properly then let me know. I am not sure if that answers your question correctly :-) . If not please let me know. Thanks - Manish This is just an understanding sort of issue which I am looking for teaching purpose. Note: Please free to change the subject line if not proper. -- Thanks Regards, Ajit Subhash Mote -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: some question about open fail on /dev/sda nodes
On Sun, Aug 16, 2009 at 3:09 PM, loodymilo...@gmail.com wrote:
Dear all:
I try to open /dev/sda1 with system call, open(/dev/sda1, O_RDONLY).
But each time it will return negative value, that means fail.
Have you checked for the errno ? I guess it is permission denied.
/home/mkatiyar dd if=/dev/sda1 of=/dev/zero
dd: opening `/dev/sda1': Permission denied
Thanks -
Manish
my program is like below:
(gdb) list
1 #includestdio.h
2 #includefcntl.h
3 int main(void)
4 {
5 int fd=0;
6 fd = open(/dev/sda1, O_RDONLY);
7
8 printf(%d\n,fd);
9 }
10
(gdb) b 8
Breakpoint 1 at 0x8048423: file test.c, line 8.
(gdb) r
Starting program: /home/cc/a.out
Breakpoint 1, main () at test.c:8
8 printf(%d\n,fd);
(gdb) p fd
$1 = -1
(gdb)
But I am quite sure I really get /dev/sda1 on my linux system.
below are the dmesg:
$ dmesg |grep sda
[ 4.304624] sd 0:0:0:0: [sda] 117210240 512-byte hardware sectors (60012
MB)
[ 4.304642] sd 0:0:0:0: [sda] Write Protect is off
[ 4.304646] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 4.304672] sd 0:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 4.304742] sd 0:0:0:0: [sda] 117210240 512-byte hardware sectors (60012
MB)
[ 4.304757] sd 0:0:0:0: [sda] Write Protect is off
[ 4.304760] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 4.304786] sd 0:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 4.304790] sda:4Driver 'sr' needs updating - please use bus_type
methods
[ 4.319499] sda1 sda2 sda3 sda56usb 4-2: configuration #1
chosen from 1 choice
[ 4.359729] sda6 sda4
[ 4.359951] sd 0:0:0:0: [sda] Attached SCSI disk
[ 22.637589] EXT3 FS on sda4, internal journal
[ 23.586600] EXT3 FS on sda6, internal journal
$
Does that mean I use the wrong systemcall or we cannot open /dev/sda1 like
this?
appreciate your help,
miloody
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Question about lock ordering.
On Wed, Aug 12, 2009 at 11:15 AM, shailesh
jaincoolworldofsh...@gmail.com wrote:
I haven't looked the code but ...
You can unlock in any order, right ? because whatever happens you will
always be able to unlock and it the locks are acquired in the same order
throughout the system, then your system is guaranteed to make forward
progress.
Hmmm ok ... same order throughout is the key.
Thanks -
Manish
As an exercise try coming up with an example that leads to you deadlock when
you unlock in incorrect order.
Shailesh Jain
On Wed, Aug 12, 2009 at 12:32 AM, Manish Katiyar mkati...@gmail.com wrote:
Hi,
I was going through this code below and was confused why do we need
lock ordering here. What is the problem that it is trying to prevent
given that the corresponding unlocking doesn't care for the order. I
assume lock orders are primarily needed to avoid deadlocks, and if
they are then they need to be unlocked in the same order too.
File : fs/ext4/move_extent.c
/**
* mext_double_down_write - Acquire two inodes' write semaphore
*
* @orig_inode: original inode structure
* @donor_inode: donor inode structure
* Acquire write semaphore of the two inodes (orig and donor) by i_ino
order.
*/
static void
mext_double_down_write(struct inode *orig_inode, struct inode
*donor_inode)
{
struct inode *first = orig_inode, *second = donor_inode;
BUG_ON(orig_inode == NULL || donor_inode == NULL);
/*
* Use the inode number to provide the stable locking order instead
* of its address, because the C language doesn't guarantee you can
* compare pointers that don't come from the same array.
*/
if (donor_inode-i_ino orig_inode-i_ino) {
first = donor_inode;
second = orig_inode;
}
down_write(EXT4_I(first)-i_data_sem);
down_write(EXT4_I(second)-i_data_sem);
}
/**
* mext_double_up_read - Release two inodes' read semaphore
*
* @orig_inode: original inode structure to be released its lock
first
* @donor_inode: donor inode structure to be released its lock
second
* Release read semaphore of two inodes (orig and donor).
*/
static void
mext_double_up_read(struct inode *orig_inode, struct inode *donor_inode)
{
BUG_ON(orig_inode == NULL || donor_inode == NULL);
up_read(EXT4_I(orig_inode)-i_data_sem);
up_read(EXT4_I(donor_inode)-i_data_sem);
}
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Question about lock ordering.
Hi,
I was going through this code below and was confused why do we need
lock ordering here. What is the problem that it is trying to prevent
given that the corresponding unlocking doesn't care for the order. I
assume lock orders are primarily needed to avoid deadlocks, and if
they are then they need to be unlocked in the same order too.
File : fs/ext4/move_extent.c
/**
* mext_double_down_write - Acquire two inodes' write semaphore
*
* @orig_inode: original inode structure
* @donor_inode:donor inode structure
* Acquire write semaphore of the two inodes (orig and donor) by i_ino order.
*/
static void
mext_double_down_write(struct inode *orig_inode, struct inode *donor_inode)
{
struct inode *first = orig_inode, *second = donor_inode;
BUG_ON(orig_inode == NULL || donor_inode == NULL);
/*
* Use the inode number to provide the stable locking order instead
* of its address, because the C language doesn't guarantee you can
* compare pointers that don't come from the same array.
*/
if (donor_inode-i_ino orig_inode-i_ino) {
first = donor_inode;
second = orig_inode;
}
down_write(EXT4_I(first)-i_data_sem);
down_write(EXT4_I(second)-i_data_sem);
}
/**
* mext_double_up_read - Release two inodes' read semaphore
*
* @orig_inode: original inode structure to be released its lock first
* @donor_inode:donor inode structure to be released its lock second
* Release read semaphore of two inodes (orig and donor).
*/
static void
mext_double_up_read(struct inode *orig_inode, struct inode *donor_inode)
{
BUG_ON(orig_inode == NULL || donor_inode == NULL);
up_read(EXT4_I(orig_inode)-i_data_sem);
up_read(EXT4_I(donor_inode)-i_data_sem);
}
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Avoid uml opening virtual consoles
[ I didn't get any response from UML mailing list, trying my luck here ] -- Forwarded message -- From: Manish Katiyar mkati...@gmail.com Date: Thu, Aug 6, 2009 at 9:07 PM Subject: Avoid uml opening virtual consoles To: user-mode-linux-user user-mode-linux-u...@lists.sourceforge.net Hi, I have downloaded fedora-10 uml images from http://uml.nagafix.co.uk/filesystems/ . After booting it up it opens up 6 virtual consoles for me. What I would like to have is just one terminal and that also in the same terminal from where I am invoking the UML kernel. (I used to get that with a different OS image that I had downloaded from UML homepage). What settings change or parameters I can pass to make this happen. My current UML inittab looks like below. == /home/mkatiyar/personal/uml/mount/etc cat inittab|grep -v ^# id:3:initdefault: si::sysinit:/etc/rc.d/rc.sysinit l0:0:wait:/etc/rc.d/rc 0 l1:1:wait:/etc/rc.d/rc 1 l2:2:wait:/etc/rc.d/rc 2 l3:3:wait:/etc/rc.d/rc 3 l4:4:wait:/etc/rc.d/rc 4 l5:5:wait:/etc/rc.d/rc 5 l6:6:wait:/etc/rc.d/rc 6 ca::ctrlaltdel:/sbin/shutdown -t3 -r now pf::powerfail:/sbin/shutdown -f -h +2 Power Failure; System Shutting Down pr:12345:powerokwait:/sbin/shutdown -c Power Restored; Shutdown Cancelled x:5:once:/etc/X11/prefdm -nodaemon c0:1235:respawn:/sbin/mingetty tty0 = -- Thanks - Manish -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Avoid uml opening virtual consoles
On Sat, Aug 8, 2009 at 5:03 AM, Mulyadi
Santosamulyadi.sant...@gmail.com wrote:
On Sat, Aug 8, 2009 at 12:44 AM, Manish Katiyarmkati...@gmail.com wrote:
[ I didn't get any response from UML mailing list, trying my luck here ]
-- Forwarded message --
From: Manish Katiyar mkati...@gmail.com
Date: Thu, Aug 6, 2009 at 9:07 PM
Subject: Avoid uml opening virtual consoles
To: user-mode-linux-user user-mode-linux-u...@lists.sourceforge.net
Hi,
I have downloaded fedora-10 uml images from
http://uml.nagafix.co.uk/filesystems/ . After booting it up it opens
up 6 virtual consoles for me. What I would like to have is just one
terminal and that also in the same terminal from where I am invoking
the UML kernel. (I used to get that with a different OS image that I
had downloaded from UML homepage). What settings change or parameters
I can pass to make this happen.
Hi Manish...
Looks like now things are a bit different in initscripts handling.
From what I read directly from /etc/inittab comment in my F9 box, I
think you might try to rename /etc/event.d/tty{1,2,3,4,5,6} into
something else or simply move them to other directory. From their
names, you can guess it's configuration file to spawn console on
certain tty.
Thanks mulyadi, I will try that. As of now adding con0=null,fd:2
con1=fd:0,fd:1 con=/dev/null to command line solves my issue. Only
thing is I get lot of error msgs for other ttys.
Thanks -
Manish
That away, I guess it will prevent the creation of virtual console on
UML. IMO, just leave tty1 as is, so you will only have 1 console open
and that console sits in the same console session you currently work
on.
--
regards,
Mulyadi Santosa
Freelance Linux trainer
blog: the-hydra.blogspot.com
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Compiling modules on UML
Hi, I am trying to compile a module on UML. The kernel with which I am booting UML is my custom kernel built on i386 machine with the command line as make defconfig ARCH=um; make linux ARCH=um. After the kernel boots up my uname shows something like below :- bash-3.2# uname -a Linux localhost.localdomain 2.6.30.4 #6 Wed Aug 5 15:12:08 IST 2009 i686 i686 i386 GNU/Linux Now when I compile a module, by default it resolves the ARCH to x86. And some of the compilation fails because somethings are expected to be always defined on x86 but not on UML (e.g. CONFIG_TRACE_IRQFLAGS_SUPPORT). because my asm directory resolves to asm-um rather than asm-x86. It throws error like below. In file included from include/linux/prefetch.h:14, from include/linux/list.h:6, from include/linux/wait.h:22, from include/linux/fs.h:359, from include/linux/buffer_head.h:11, from /home/mkatiyar/personal/ohsm/test/kernel/admin.c:23: /mnt/host/home/mkatiyar/personal/uml/linux-2.6.30.4/arch/x86/include/asm/processor.h: In function 'load_cr3': /mnt/host/home/mkatiyar/personal/uml/linux-2.6.30.4/arch/x86/include/asm/processor.h:191: error: 'CONFIG_PAGE_OFFSETUL' undeclared (first use in this function) /mnt/host/home/mkatiyar/personal/uml/linux-2.6.30.4/arch/x86/include/asm/processor.h:191: error: (Each undeclared identifier is reported only once /mnt/host/home/mkatiyar/personal/uml/linux-2.6.30.4/arch/x86/include/asm/processor.h:191: error: for each function it appears in.) /mnt/host/home/mkatiyar/personal/uml/linux-2.6.30.4/arch/x86/include/asm/processor.h: In function 'wbinvd_halt': /mnt/host/home/mkatiyar/personal/uml/linux-2.6.30.4/arch/x86/include/asm/processor.h:771: error: implicit declaration of function 'halt' My question is how do people build modules which are specific to a particular architecture on UML ? Is it possible ? Once I export ARCH=um before doing a make, I can do a successful build. -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: [uml-user] Compiling modules on UML
On Wed, Aug 5, 2009 at 4:46 PM, Flaviofbcyb...@gmail.com wrote: 2009/8/5 Manish Katiyar mkati...@gmail.com: My question is how do people build modules which are specific to a particular architecture on UML ? Is it possible ? Sure! Are you compiling the kernel modules from the host system? If not, try to do it. The exact procedure is the following: # mount -o loop rootfs loop_dir/ # cd /uml/src/linux # make modules ARCH=um # make modules_install INSTALL_MOD_PATH=/uml/loop_dir ARCH=um # umount /uml/loop_dir Then, boot the UML again. Hi Flavio, I have done that step and my UML kernel boots fine. My question is in UML environment what will the ARCH resolve to ? will it be um or x86 ? Thanks - Manish Flavio -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: [uml-user] Compiling modules on UML
On Wed, Aug 5, 2009 at 8:44 PM, H M Thalibhmtha...@gmail.com wrote: Hi Manish, Manish Katiyar wrote: On Wed, Aug 5, 2009 at 4:46 PM, Flaviofbcyb...@gmail.com wrote: 2009/8/5 Manish Katiyar mkati...@gmail.com: My question is how do people build modules which are specific to a particular architecture on UML ? Is it possible ? Sure! Are you compiling the kernel modules from the host system? If not, try to do it. The exact procedure is the following: # mount -o loop rootfs loop_dir/ # cd /uml/src/linux # make modules ARCH=um # make modules_install INSTALL_MOD_PATH=/uml/loop_dir ARCH=um # umount /uml/loop_dir Then, boot the UML again. Hi Flavio, I have done that step and my UML kernel boots fine. My question is in UML environment what will the ARCH resolve to ? will it be um or x86 ? Thanks - Manish Flavio It will um. Actualy UML is a port of the Linux kernel as a normal Intel Linux process. So that would mean that any new kernel module that I compile has to be prefixed with ARCH=um ??? because my kernel source resides on host and has been configured as um. If I compile any x86 module from UML it will fail because the asm directory in kernel source points to asm-um and not asm-x86. Isn't it ?? Or I am just talking senseless :-( ... Below is how my directory structures look like. -bash-3.2# ls -l /lib/modules/2.6.30.4/source/include/ total 120 -rw-r--r-- 1 mkatiyar 5573 218 Jul 30 18:34 Kbuild drwxr-xr-x 3 mkatiyar 5573 4096 Jul 30 18:34 acpi lrwxrwxrwx 1 mkatiyar 5573 6 Aug 5 06:16 asm - asm-um -bash-3.2# ls -l /lib/modules/2.6.30.4/ total 64 lrwxrwxrwx 1 root root 42 Aug 4 13:36 build - /home/mkatiyar/personal/uml/linux-2.6.30.4 drwxr-xr-x 6 root root 4096 Aug 5 06:18 kernel -rw-r--r-- 1 root root 265 Aug 5 11:01 modules.alias .. ... -rw-r--r-- 1 root root 189 Aug 5 11:01 modules.usbmap lrwxrwxrwx 1 root root 42 Aug 5 06:18 source - /home/mkatiyar/personal/uml/linux-2.6.30.4 -bash-3.2# ls -l /home/mkatiyar total 0 lrwxrwxrwx 1 root root 32 Aug 4 14:48 personal - /mnt/host/home/mkatiyar/personal -bash-3.2# mount /dev/ubda on / type auto (rw) none on /proc type proc (rw) none on /dev/pts type devpts (rw,gid=5,mode=620) none on /mnt/host type hostfs (rw) Thanks - Manish -- Thanks Regards, H M Thalib. -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Changing path in kernel object Debug of Kernel module with gdb
On Wed, Aug 5, 2009 at 11:18 PM, Microbit_Ubuntumicro...@virginbroadband.com.au wrote: Hello Lin/all, Following up on my reply to the u32 issue and Kernel compilation (in Eclipse/CDT), I am wondering about 2 issues someone here might have figured before : 1. How do you (re) change path to set the module output when building ? - The hierarchy of my project (ideally) requires that my Kernel Module source resides in a sub-dir called 'src'. Thus, my (test) module is defined as such : obj-m := ../src/my_test.o OK, all is good, no problem. But now my module output ends up in the /src/ sub directory. However, I want the executable code to land in my /Debug folder Can anyone give a hint how to accomplish this ? A make install target ??? I can't seem to readily find a solution to this... So, basically, the folders are like this : My Project : ... src/ Debug/ My Makefile is in the Debug folder - the module source is in /src. But the module output needs to go into Debug/ folder ? 2. As referred to in the u32 previous post, I'm a bit stumped how to invoke GBD on debugging kernel modules. The commands would depend on what environment are yoiu developing your modules. A vmware based kernel can easily talk to gdb via some predefined port as specified in vmx file. A UML kernel can be debugged by just starting the kernel in gdb . Thanks - Manish Do I need a separate pre-command from the debugger, launching insmod, or some such ? I look forward to any help or suggestions I can get... -- Best regards, Kris -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: a couple questions about gdb vmlinux /proc/kcore
On Sun, Aug 2, 2009 at 7:58 PM, Robert P. J. Dayrpj...@crashcourse.ca wrote:
On Sun, 2 Aug 2009, Manish Katiyar wrote:
Hi Robert,
May be you can add how to see the already loaded modules and traverse
or find a particular module. Something like
(gdb) p modules
$1 = {next = 0x3502cea4, prev = 0x3502cea4}
(gdb) p *((struct module *)0x3502cea0)
$3 = {state = MODULE_STATE_LIVE, list = {next = 0x81e0dec, prev = 0x81e0dec},
name = loop, '\0' repeats 55 times, mkobj = {kobj = {
k_name = 0x3502ceec loop, name = loop, '\0' repeats 15 times,
kref = {refcount = {counter = 2}}, entry = {next = 0x81e0ac8,
prev = 0x34a486c8}, parent = 0x81e0ad0, kset = 0x81e0ac0, ktype =
0x0,
dentry = 0x3297277c, poll = {lock = {raw_lock = {No data fields}},
task_list = {next = 0x3502cf1c, prev = 0x3502cf1c}}},
mod = 0x3502cea0}, param_attrs = 0x0, modinfo_attrs = 0x9598620,
version = 0x0, srcversion = 0x0, syms = 0x3502bc20, num_syms = 2,
crcs = 0x0, gpl_syms = 0x0, num_gpl_syms = 0, gpl_crcs = 0x0,
unused_syms = 0x0, num_unused_syms = 0, unused_crcs = 0x0,
unused_gpl_syms = 0x0, num_unused_gpl_syms = 0, unused_gpl_crcs = 0x0,
gpl_future_syms = 0x0, num_gpl_future_syms = 0, gpl_future_crcs = 0x0,
num_exentries = 0, extable = 0x0, init = 0x3502f000, module_init = 0x0,
module_core = 0x3502a000, init_size = 0, core_size = 12328,
init_text_size = 0, core_text_size = 6468, unwind_info = 0x0,
arch = {No data fields}, unsafe = 0, license_gplok = 1, ref = {{count = {
counter = 0}}}, modules_which_use_me = {next = 0x3502cfe0,
prev = 0x3502cfe0}, waiter = 0x94cf7c0, exit = 0x3502b8b6,
symtab = 0x3502bc30, num_symtab = 166, strtab = 0x3502c690 ,
sect_attrs = 0x346e14d8, percpu = 0x0, args = 0x94c3698 }
And similarly traversing the chain and printing the list . Also
to add that the loading address == $3.module_core and may be highlight
on other important members.
possibly, but to start with, i'm looking for simple variables that
beginners can print out, just for the entertainment value, to convince
themselves that they really are getting to examine kernel data values.
by way of clarification, let me explain what i'm trying to do with my
kernel newbie column over at linux.com.
my main beef with documentation as long as i can remember is that it
so often doesn't carefully explain how to get *started* doing
something. most of the time, it blasts through the initial install
and setup, then launches straight into how to do involved and advanced
stuff. but a lot of the time, if so much as a single setup
instruction is mistyped or forgotten, the poor beginner is left
sitting there, not even able to get the program *started*. and if
only the newbie could get past that point to do the *simple* stuff,
everything after that follows nicely. so the critical issue is to get
past that startup phase.
that's what i'm trying to do -- get newbies past that introductory
phase.
Hi Robert,
I understand and appreciate your effort and motive. My motive behind
suggesting that was that most of the beginners look kernel code as
black box with some smart or complex data structures and are afraid to
delve into it. Assuming most of them are already pretty good in
userspace, by traversing you are just trying to say that this is
nothing more than a simple linked list program and adding a loadable
modules is just adding a node in that list.
Thanks -
Manish
as long as a beginner can write a simple module and get it
compiled and loaded, that's a *huge* victory because, after that, they
can simply start adding new features a little at a time to see how
they work. but if they can't even get that first hello, world
module loaded and running, they get frustrated and walk away. *my*
focus in my columns is to get total beginners over that initial hump,
at which point a lot of them will take off from there.
so while printing out the loaded modules is cool (and i might even
use that), i don't want to get into such advanced examples that the
beginners lose sight of what they're trying to do. anyway, for better
or worse, that's the philosophy behind the weekly column -- to get
people with *no* kernel programming experience into kernel
programming.
rday
--
Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca
Twitter: http://twitter.com/rpjday
Kernel Newbie Corner column @ linux.com: http://cli.gs/WG6WYX
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http
Re: a couple questions about gdb vmlinux /proc/kcore
On Sun, Aug 2, 2009 at 2:00 AM, Robert P. J. Dayrpj...@crashcourse.ca wrote: for an upcoming column on how to debug the kernel (and modules) with gdb, i'm showing how to use the above and, since it's been a while, i need a refresher. once i'm in gdb, i can list the section breakdown of the vmlinux file with: (gdb) info files ... Local exec file: `/home/rpjday/k/f11/vmlinux', file type elf64-x86-64. Entry point: 0x100 0x8100 - 0x81009000 is .text.head 0x81009000 - 0x813f2d3f is .text 0x813f2d40 - 0x813f2ebc is .notes 0x813f2ec0 - 0x813f6f80 is __ex_table 0x813f7000 - 0x8158e10a is .rodata 0x8158e110 - 0x8159417c is __bug_table 0x81594180 - 0x81595810 is .pci_fixup 0x81595810 - 0x81595838 is .tracedata 0x81595840 - 0x815a26c0 is __ksymtab 0x815a26c0 - 0x815a91d0 is __ksymtab_gpl 0x815a91d0 - 0x815a91e0 is __ksymtab_unused 0x815a91e0 - 0x815bf743 is __ksymtab_strings 0x815bf750 - 0x815eacd0 is __init_rodata 0x815eacd0 - 0x815ed000 is __param 0x815ed000 - 0x8165ae60 is .data ... and so on. if i try to list the variables (obviously a long list) with (gdb) info variables gdb just goes away (perhaps I'm just asking for too much?). is there a way to ask for the variable/symbol list within a single section? i don't see the gdb command for that. also (and a simpler question), as an example of how to print kernel variables once you're in gdb, i always start with one of: (gdb) p jiffies (gdb) p jiffies_64 depending on your architecture. are there some other kernel vars that are informative and obvious to print? there's loops_per_jiffy as well, but off the top of peoples' heads, given the stunning number of possible kernel variables to print, which ones would be good examples that would be meaningful for a newbie column? how about p current ?? Thanks - Manish thanks. rday -- Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Annoying Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday Kernel Newbie Corner column @ linux.com: http://cli.gs/WG6WYX -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Tmpfs nr_inodes
On Wed, Jul 29, 2009 at 10:08 AM, Mulyadi Santosamulyadi.sant...@gmail.com wrote: On Wed, Jul 29, 2009 at 4:06 AM, Darvin Denmiandarvin.denm...@gmail.com wrote: Hello, My question is about tmpfs: - If i increase the number of inodes avaliable in a tmpfs mounted directory, will i have some performance impact? # Before increase number of inodes df -i /media/queue tmpfs 129384 4229 125155 4% /media/queue # After increase number of inodes tmpfs 2048000 4229 2043771 1% /media/queue Sorry for my Brazilian English :( I think no. AFAIK, number of inodes just tell you how many file you can make (since one inode represents the metadata of a file). However, IMO, depending on average file size you will create there, having more inodes doesn't mean you can cram more files into your tmpfs based directory. So it's more like balancing between inodes, predicted number of file and storage size. By default, mount do this for you. Yes... In general increasing number of inodes shouldn't affect performance unless you are going to pre-allocate space for that many inodes on disk (things like expanding your bitmaps and other metadata stuff). Thanks - Manish CMIIW people -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Accessing File system directories
On Mon, Jul 27, 2009 at 2:49 PM, Dan Danillodandani...@ymail.com wrote: Hello, Hi Dan, On Mon, Jul 27, 2009 at 11:54 AM, Dan Danillowrote: Hello, I'm trying to find out how to browse the file system from the linux kernel. What exactly do you mean by that. Do you mean that you want to traverse the namespace being inside the kernel through some kernel module or something? Yes, I want to write a module which gets as parameter the path to a directory. In the module I want to be able to find all files and subfolders in that directory. Especially the directory content is interesting to me. Can anyone give me a hint? Are there even some libc like functions to do that ? The libc functions like these map to the kernel code eventually, so mostly there will be one. So far I have found the filp_open, vfs_read and vfs_readdir functions. However, the vfs_readdir function expects a function pointer to somehow format the directory information. My assumption is that this is the wrong function. So I guess that's not the function I'm after. Do you want it at vfs layer or at the file system level Mh, I don't really have a clue. What are the differences? I'm not interested in all the inode information or some funky stuff. I just need to know if it is a file or directory, filesize and where to find the data of the file. Hi, Have a look at vfs_path_lookup() . In particular probably you want path_walk(). Thanks - Manish What all information do you have before you could call this function which will traverse the file system hierarchy for you? Just the path to the directory Dan -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Checking Installed library
On Sat, Jul 18, 2009 at 2:37 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Sat, Jul 18, 2009 at 2:25 PM, Manish Katiyar mkati...@gmail.com wrote: On Sat, Jul 18, 2009 at 1:45 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Sat, Jul 18, 2009 at 1:05 PM, Manish Katiyar mkati...@gmail.com wrote: On Sat, Jul 18, 2009 at 9:48 AM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 7:22 PM, Manish Katiyar mkati...@gmail.com wrote: On Fri, Jul 17, 2009 at 10:34 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:40 PM, Rishi Agrawal rishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:27 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On 7/17/09, Rishi Agrawal rishi.b.agra...@gmail.com wrote: Is autoconfigure the only way to find this out.?? On Wed, Jul 15, 2009 at 6:52 AM, Pei Lin telent...@gmail.com wrote: Please don't do top posting... Although it's OOT, I can help a bit. IMO, you can directly check /usr/lib or parse /etc/ld.so.conf and so on to find the related DSO and header files. However, doing that will require more time and probably not portable across system, in fact it's kinda reinventing the wheel. So, I personally suggest to adopt autoconfigure. -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com I will have a detailed look at the autoconfigure utility and see. Thanks for all the help. -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ I have used it and it is working very good. I wanted to know that How to check the presence of libxml2 library using autoconfigure ? First link in google. http://www.kdevelop.org/index.html?filename=3.0/doc/tutorial_autoconf.html Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish Actually I tried using that information earlier but somehow it did not work. Can you please tell us a little more why it didn't work ? Any errors or anything else ?? Currently I am using a macro which checks wether a function is present in the mentioned library. I didn't understand this ? are you using something like nm ?? like printf can be checked in libc I used a function xmlParsefile and libxml2 library and this works fine. The logic behind this is that when the function is present in the library then the library is also present on the system (is it correct ??) Should be... but by any chance are you relying the library to be present in a particular path ?? If that's the case it is not going to work anywhere ? Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish I am using this particular line: AC_CHECK_LIB(xml2,xmlParseFile,,AC_MSG_ERROR(oops! no function xmlParseFile function in xml?!?),) in my configure.in for autoconf You can try something like below too ... it doesn't depend on any particular function. /tmp/test ./configure checking for gcc... gcc checking for C compiler default output file name... a.out checking whether the C compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... checking for suffix of object files... o checking whether we are using the GNU C compiler... yes checking whether gcc accepts -g... yes checking for gcc option to accept ISO C89... none needed checking for -pkg-config... no checking for pkg-config... /usr/bin/pkg-config checking for libxml2... yes configure: creating ./config.status /tmp/test cat configure.ac # -*- Autoconf -*- # Process this file with autoconf to produce a configure script. AC_PREREQ([2.63]) AC_INIT([Sample test package], [1.1.0], [mkati...@gmail.com]) AC_CONFIG_SRCDIR([a.c]) AC_PROG_CC AC_PATH_PROGS([PACKAGE_CONFIG], [$host_alias-pkg-config pkg-config], [none]) if test x$PACKAGE_CONFIG = xnone; then AC_MSG_ERROR([*** pkg-config, needed to check for libxml2 existence has not been found.]) fi AC_MSG_CHECKING(for libxml2) if $PACKAGE_CONFIG libxml-2.0 --libs /dev/null 21; then AC_MSG_RESULT(yes) else AC_MSG_RESULT(no) fi AC_OUTPUT = Thanks - Manish The use of macro is AC_CHECK_LIB : Checks whether a function exists in the given library (library names without the leading lib, e.g., for libxml, use just xml here) It seems that it is not path dependant
Re: Checking Installed library
On Sat, Jul 18, 2009 at 9:48 AM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 7:22 PM, Manish Katiyar mkati...@gmail.com wrote: On Fri, Jul 17, 2009 at 10:34 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:40 PM, Rishi Agrawal rishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:27 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On 7/17/09, Rishi Agrawal rishi.b.agra...@gmail.com wrote: Is autoconfigure the only way to find this out.?? On Wed, Jul 15, 2009 at 6:52 AM, Pei Lin telent...@gmail.com wrote: Please don't do top posting... Although it's OOT, I can help a bit. IMO, you can directly check /usr/lib or parse /etc/ld.so.conf and so on to find the related DSO and header files. However, doing that will require more time and probably not portable across system, in fact it's kinda reinventing the wheel. So, I personally suggest to adopt autoconfigure. -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com I will have a detailed look at the autoconfigure utility and see. Thanks for all the help. -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ I have used it and it is working very good. I wanted to know that How to check the presence of libxml2 library using autoconfigure ? First link in google. http://www.kdevelop.org/index.html?filename=3.0/doc/tutorial_autoconf.html Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish Actually I tried using that information earlier but somehow it did not work. Can you please tell us a little more why it didn't work ? Any errors or anything else ?? Currently I am using a macro which checks wether a function is present in the mentioned library. I didn't understand this ? are you using something like nm ?? like printf can be checked in libc I used a function xmlParsefile and libxml2 library and this works fine. The logic behind this is that when the function is present in the library then the library is also present on the system (is it correct ??) Should be... but by any chance are you relying the library to be present in a particular path ?? If that's the case it is not going to work anywhere ? Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Checking Installed library
On Sat, Jul 18, 2009 at 1:45 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Sat, Jul 18, 2009 at 1:05 PM, Manish Katiyar mkati...@gmail.com wrote: On Sat, Jul 18, 2009 at 9:48 AM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 7:22 PM, Manish Katiyar mkati...@gmail.com wrote: On Fri, Jul 17, 2009 at 10:34 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:40 PM, Rishi Agrawal rishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:27 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On 7/17/09, Rishi Agrawal rishi.b.agra...@gmail.com wrote: Is autoconfigure the only way to find this out.?? On Wed, Jul 15, 2009 at 6:52 AM, Pei Lin telent...@gmail.com wrote: Please don't do top posting... Although it's OOT, I can help a bit. IMO, you can directly check /usr/lib or parse /etc/ld.so.conf and so on to find the related DSO and header files. However, doing that will require more time and probably not portable across system, in fact it's kinda reinventing the wheel. So, I personally suggest to adopt autoconfigure. -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com I will have a detailed look at the autoconfigure utility and see. Thanks for all the help. -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ I have used it and it is working very good. I wanted to know that How to check the presence of libxml2 library using autoconfigure ? First link in google. http://www.kdevelop.org/index.html?filename=3.0/doc/tutorial_autoconf.html Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish Actually I tried using that information earlier but somehow it did not work. Can you please tell us a little more why it didn't work ? Any errors or anything else ?? Currently I am using a macro which checks wether a function is present in the mentioned library. I didn't understand this ? are you using something like nm ?? like printf can be checked in libc I used a function xmlParsefile and libxml2 library and this works fine. The logic behind this is that when the function is present in the library then the library is also present on the system (is it correct ??) Should be... but by any chance are you relying the library to be present in a particular path ?? If that's the case it is not going to work anywhere ? Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish I am using this particular line: AC_CHECK_LIB(xml2,xmlParseFile,,AC_MSG_ERROR(oops! no function xmlParseFile function in xml?!?),) in my configure.in for autoconf You can try something like below too ... it doesn't depend on any particular function. /tmp/test ./configure checking for gcc... gcc checking for C compiler default output file name... a.out checking whether the C compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... checking for suffix of object files... o checking whether we are using the GNU C compiler... yes checking whether gcc accepts -g... yes checking for gcc option to accept ISO C89... none needed checking for -pkg-config... no checking for pkg-config... /usr/bin/pkg-config checking for libxml2... yes configure: creating ./config.status /tmp/test cat configure.ac # -*- Autoconf -*- # Process this file with autoconf to produce a configure script. AC_PREREQ([2.63]) AC_INIT([Sample test package], [1.1.0], [mkati...@gmail.com]) AC_CONFIG_SRCDIR([a.c]) AC_PROG_CC AC_PATH_PROGS([PACKAGE_CONFIG], [$host_alias-pkg-config pkg-config], [none]) if test x$PACKAGE_CONFIG = xnone; then AC_MSG_ERROR([*** pkg-config, needed to check for libxml2 existence has not been found.]) fi AC_MSG_CHECKING(for libxml2) if $PACKAGE_CONFIG libxml-2.0 --libs /dev/null 21; then AC_MSG_RESULT(yes) else AC_MSG_RESULT(no) fi AC_OUTPUT = Thanks - Manish The use of macro is AC_CHECK_LIB : Checks whether a function exists in the given library (library names without the leading lib, e.g., for libxml, use just xml here) It seems that it is not path dependant. -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Checking Installed library
On Fri, Jul 17, 2009 at 10:34 PM, Rishi Agrawalrishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:40 PM, Rishi Agrawal rishi.b.agra...@gmail.com wrote: On Fri, Jul 17, 2009 at 12:27 PM, Mulyadi Santosa mulyadi.sant...@gmail.com wrote: On 7/17/09, Rishi Agrawal rishi.b.agra...@gmail.com wrote: Is autoconfigure the only way to find this out.?? On Wed, Jul 15, 2009 at 6:52 AM, Pei Lin telent...@gmail.com wrote: Please don't do top posting... Although it's OOT, I can help a bit. IMO, you can directly check /usr/lib or parse /etc/ld.so.conf and so on to find the related DSO and header files. However, doing that will require more time and probably not portable across system, in fact it's kinda reinventing the wheel. So, I personally suggest to adopt autoconfigure. -- regards, Mulyadi Santosa Freelance Linux trainer blog: the-hydra.blogspot.com I will have a detailed look at the autoconfigure utility and see. Thanks for all the help. -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ I have used it and it is working very good. I wanted to know that How to check the presence of libxml2 library using autoconfigure ? First link in google. http://www.kdevelop.org/index.html?filename=3.0/doc/tutorial_autoconf.html Thanks - Manish -- Regards, Rishi B. Agrawal http://www.linkedin.com/in/rishibagrawal http://code.google.com/p/fscops/ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Tue, Jul 7, 2009 at 2:41 PM, Donato Capitellad.capite...@tin.it wrote:
On Tue, 2009-07-07 at 00:10 +0530, Manish Katiyar wrote:
Hmm.. Can you try to write something like ext3_get_journal() and tell
us the errors you are enountering.
Here's the function I wrote:
//
struct inode *secdel_get_journal(struct super_block *sb)
{
struct inode *journal;
journal = ext3_iget(sb, 3);
if (IS_ERR(journal)) {
printk(KERN_ERR ext3secdel - no journal found.\n);
return NULL;
}
if (!S_ISREG(journal-i_mode)) {
printk(KERN_ERR ext3secdel - invalid journal inode.\n);
iput(journal);
return NULL;
}
return journal;
}
//
It doesn't crash, but stops on the first error, after the ext3_iget()
call and returns NULL.
Can you also add print the value of journal after it is returned from
ext3_iget() in debug statement. It will have the error code and will
help to trace the problem.
And also could you post what is the
output of stat 3 after you have initialized your reserved inode.
Here's the output:
May be you should also initialize i_nlinks = 1.
Thanks -
Manish
Inode: 3 Type: regular Mode: 0600 Flags: 0x0 Generation: 0x0
User: 0 Group: 0 Size: 0
File ACL: 0 Directory ACL: 0
Links: 0 Blockcount: 0
Fragment: Address: 0 Number: 0 Size: 0
ctime: 0x4622bea1 -- Mon Apr 16 08:09:05 2007
atime: 0x462a2136 -- Sat Apr 21 22:35:34 2007
mtime: 0x4622325e -- Sun Apr 15 22:10:38 2007
BLOCKS:
ps. never mind the dates, I put random values...
Instead of trying to reference any pointer and do any operation, could
you just try to print success or the errno from the ext3_iget()
Thanks -
Manish
Thank you,
Donato
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Tue, Jul 7, 2009 at 6:16 PM, Donato Capitellad.capite...@tin.it wrote:
On Tue, 2009-07-07 at 15:46 +0530, Manish Katiyar wrote:
On Tue, Jul 7, 2009 at 3:21 PM, Donato Capitellad.capite...@tin.it wrote:
On Tue, 2009-07-07 at 15:10 +0530, Manish Katiyar wrote:
On Tue, Jul 7, 2009 at 2:41 PM, Donato Capitellad.capite...@tin.it
wrote:
It doesn't crash, but stops on the first error, after the ext3_iget()
call and returns NULL.
Can you also add print the value of journal after it is returned from
ext3_iget() in debug statement. It will have the error code and will
help to trace the problem.
journal after calling ext3_iget() has value -5.
OK I think I understand your problem. My guess is that you chose
the inode number to be 3 , but forgot to update the below function
static inline int ext3_valid_inum(struct super_block *sb, unsigned
long ino)
{
return ino == EXT3_ROOT_INO ||
ino == EXT3_JOURNAL_INO ||
ino == EXT3_RESIZE_INO ||
(ino = EXT3_FIRST_INO(sb)
ino = le32_to_cpu(EXT3_SB(sb)-s_es-s_inodes_count));
}
Thanks, this helped. I just modified ext3_valid_inum() to accept inode
3. Unfortunatelly, now I get error -116.
And this time you are probably hitting this code in ext3_iget()
because you forgot to initialize inode-i_nlink.
if (inode-i_nlink == 0) {
if (inode-i_mode == 0 ||
!(EXT3_SB(inode-i_sb)-s_mount_state EXT3_ORPHAN_FS)) {
/* this inode is deleted */
brelse (bh);
ret = -ESTALE;
goto bad_inode;
}
/* The only unlinked inodes we let through here have
* valid i_mode and are being read by the orphan
* recovery code: that's fine, we're about to complete
* the process of deleting those. */
}
How do I understand what this
error corresponds to? Is it documented somewhere?
This error corresponds to -ESTALE. See the file
include/asm-generic/errno.h in kernel source tree.
Hope that helps
Thanks -
Manish
--
Thanks -
Manish
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Mon, Jul 6, 2009 at 8:45 PM, Donato Capitellad.capite...@tin.it wrote: [ Please keep kernelnewbies in cc while replying ] Well, I tried to initialize the inode 3 with debugfs, but nothing changed. I just set the file mode and the creation but when I try to read the inode with the ext3_iget() function, somehow it fails and dereferences a NULL pointer :( What pointer is NULL ? ext3_iget doesn't return a NULL pointer (either it returns a valid inode pointer or an errno typecasted as ptr). Where is it crashing ? What are you using the reserved inode for ? Maybe I am missing something. Unless you tell us the complete picture it might be hard to identify. A snippet of code would definitely help, but I am pretty sure that it is because you don't have a valid inode structure on disk and you need to initialize it appropriately. Which fields do I need to initialize in order for this to work? Have you tried the same thing on ext2 first ? I would suggest doing that because ext2_iget() is much simpler than ext3_iget(). While setting i_mode you also need to take care that mode also consists of filetype apart from permissions. You also may have to initialize i_nlinks appropriately Shouldn't be this hard to just read an inode... It isn't but you need to create a valid inode before that. Thanks - Manish Thanks again, Donato On Sun, 2009-07-05 at 22:37 +0530, Manish Katiyar wrote: On Sun, Jul 5, 2009 at 10:28 PM, Donato Capitellad.capite...@tin.it wrote: Well, as it's a reserved inode, I didn't create it. stat 3 shows that the inode has bad type... so this might be the problem. But how am I supposed to initialize the inode fields if I can't read it? What are you reserving it for ? Note that the reserved inodes today (2 for root) (8 for journal commonly) both are initialized during mkfs. You also need to do something similar. Depending on what you are using it for you may need to initialize it properly. To do that you can do the following :- a) Either write a simple C program to write the fields properly at correct offset. b) Hack the code of mkfs to include your inode too. c) Use debugfs to change the contents of inode 3. This is the easiest option but I am not sure if you can modify all the fields of the inode, so it depends on what you need. Thanks - Manish On Sun, 2009-07-05 at 22:06 +0530, Manish Katiyar wrote: On Sun, Jul 5, 2009 at 7:30 PM, Donato Capitellad.capite...@tin.it wrote: Hi there, I need some help with a patch to the ext3 filesystem I'm developing. I'm completelly new to kernel programming and I need to understand practically how to use one of the ext3 reserved inode for storing a log. I thought it would be just as easy as getting the inode with the ext3_iget() function, but it doesn't seem to be the case. For example, if I try to load the third reserved inode like this : inode = ext3_iget(sb, 3); ext3_iget() fails somehow and crashes the process with a NULL pointer reference. What am I missing here? Have you created the inode 3 ? What entries does it have ? What does stat 3 show from the debugfs ? Thanks - Manish How do you use a reserved inode? Can anybody point to some documentation or project that uses reserved inodes? I'm really stuck :( Thanks a lot, Donato Capitella -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Mon, Jul 6, 2009 at 11:50 PM, Donato Capitellad.capite...@tin.it wrote:
Ok, this is what I want to do.
I'm developing a patch to ext3; this patch uses a queue of structures
representing tasks to be performed and a separated kernel thread that
reads the structures from the queue as they become available and
performs the required task on the filesystem.
So, if the system crashes when there are still things to do in the
queue, those tasks won't be accomplished. That's why I want to keep this
queue in a reserved inode so that I can restart the tasks in case of
system crash.
Now, what am I trying to develop? Basically this function:
write_to_log(my_struct *task) {
1. get reserved inode
2. append the struct pointed to by task to the reserved inode
}
Hmm.. Can you try to write something like ext3_get_journal() and tell
us the errors you are enountering. And also could you post what is the
output of stat 3 after you have initialized your reserved inode.
Instead of trying to reference any pointer and do any operation, could
you just try to print success or the errno from the ext3_iget()
Thanks -
Manish
I tried to be clear, my English doesn't help I guess :P.
Thanks again for your time,
Donato :D
On Mon, 2009-07-06 at 22:47 +0530, Manish Katiyar wrote:
On Mon, Jul 6, 2009 at 8:45 PM, Donato Capitellad.capite...@tin.it wrote:
[ Please keep kernelnewbies in cc while replying ]
Well, I tried to initialize the inode 3 with debugfs, but nothing
changed. I just set the file mode and the creation but when I try to
read the inode with the ext3_iget() function, somehow it fails and
dereferences a NULL pointer :(
What pointer is NULL ? ext3_iget doesn't return a NULL pointer (either
it returns a valid inode pointer or an errno typecasted as ptr). Where
is it crashing ? What are you using the reserved inode for ?
Maybe I am missing something.
Unless you tell us the complete picture it might be hard to identify.
A snippet of code would definitely help, but I am pretty sure that it
is because you don't have a valid inode structure on disk and you need
to initialize it appropriately.
Which fields do I need to initialize in
order for this to work?
Have you tried the same thing on ext2 first ? I would suggest doing
that because ext2_iget() is much simpler than ext3_iget(). While
setting i_mode you also need to take care that mode also consists of
filetype apart from permissions. You also may have to initialize
i_nlinks appropriately
Shouldn't be this hard to just read an inode...
It isn't but you need to create a valid inode before that.
Thanks -
Manish
Thanks again,
Donato
On Sun, 2009-07-05 at 22:37 +0530, Manish Katiyar wrote:
On Sun, Jul 5, 2009 at 10:28 PM, Donato Capitellad.capite...@tin.it
wrote:
Well, as it's a reserved inode, I didn't create it. stat 3 shows that
the inode has bad type... so this might be the problem. But how am I
supposed to initialize the inode fields if I can't read it?
What are you reserving it for ? Note that the reserved inodes today (2
for root) (8 for journal commonly) both are initialized during mkfs.
You also need to do something similar. Depending on what you are using
it for you may need to initialize it properly. To do that you can do
the following :-
a) Either write a simple C program to write the fields properly at
correct offset.
b) Hack the code of mkfs to include your inode too.
c) Use debugfs to change the contents of inode 3. This is the easiest
option but I am not sure if you can modify all the fields of the
inode, so it depends on what you need.
Thanks -
Manish
On Sun, 2009-07-05 at 22:06 +0530, Manish Katiyar wrote:
On Sun, Jul 5, 2009 at 7:30 PM, Donato Capitellad.capite...@tin.it
wrote:
Hi there,
I need some help with a patch to the ext3 filesystem I'm developing.
I'm
completelly new to kernel programming and I need to understand
practically how to use one of the ext3 reserved inode for storing a
log.
I thought it would be just as easy as getting the inode with the
ext3_iget() function, but it doesn't seem to be the case. For
example,
if I try to load the third reserved inode like this :
inode = ext3_iget(sb, 3);
ext3_iget() fails somehow and crashes the process with a NULL pointer
reference. What am I missing here?
Have you created the inode 3 ? What entries does it have ? What does
stat 3 show from the debugfs ?
Thanks -
Manish
How do you use a reserved inode? Can
anybody point to some documentation or project that uses reserved
inodes?
I'm really stuck :(
Thanks a lot,
Donato Capitella
--
To unsubscribe from this list: send an email with
unsubscribe kernelnewbies to ecar...@nl.linux.org
Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Sun, Jul 5, 2009 at 10:28 PM, Donato Capitellad.capite...@tin.it wrote: Well, as it's a reserved inode, I didn't create it. stat 3 shows that the inode has bad type... so this might be the problem. But how am I supposed to initialize the inode fields if I can't read it? What are you reserving it for ? Note that the reserved inodes today (2 for root) (8 for journal commonly) both are initialized during mkfs. You also need to do something similar. Depending on what you are using it for you may need to initialize it properly. To do that you can do the following :- a) Either write a simple C program to write the fields properly at correct offset. b) Hack the code of mkfs to include your inode too. c) Use debugfs to change the contents of inode 3. This is the easiest option but I am not sure if you can modify all the fields of the inode, so it depends on what you need. Thanks - Manish On Sun, 2009-07-05 at 22:06 +0530, Manish Katiyar wrote: On Sun, Jul 5, 2009 at 7:30 PM, Donato Capitellad.capite...@tin.it wrote: Hi there, I need some help with a patch to the ext3 filesystem I'm developing. I'm completelly new to kernel programming and I need to understand practically how to use one of the ext3 reserved inode for storing a log. I thought it would be just as easy as getting the inode with the ext3_iget() function, but it doesn't seem to be the case. For example, if I try to load the third reserved inode like this : inode = ext3_iget(sb, 3); ext3_iget() fails somehow and crashes the process with a NULL pointer reference. What am I missing here? Have you created the inode 3 ? What entries does it have ? What does stat 3 show from the debugfs ? Thanks - Manish How do you use a reserved inode? Can anybody point to some documentation or project that uses reserved inodes? I'm really stuck :( Thanks a lot, Donato Capitella -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Sun, Jul 5, 2009 at 7:30 PM, Donato Capitellad.capite...@tin.it wrote: Hi there, I need some help with a patch to the ext3 filesystem I'm developing. I'm completelly new to kernel programming and I need to understand practically how to use one of the ext3 reserved inode for storing a log. I thought it would be just as easy as getting the inode with the ext3_iget() function, but it doesn't seem to be the case. For example, if I try to load the third reserved inode like this : inode = ext3_iget(sb, 3); ext3_iget() fails somehow and crashes the process with a NULL pointer reference. What am I missing here? Have you created the inode 3 ? What entries does it have ? What does stat 3 show from the debugfs ? Thanks - Manish How do you use a reserved inode? Can anybody point to some documentation or project that uses reserved inodes? I'm really stuck :( Thanks a lot, Donato Capitella -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: help with ext3 reserved inode
On Sun, Jul 5, 2009 at 11:08 PM, Donato Capitellad.capite...@tin.it wrote: Ok, thanks, now it is clear to me. I thought that as the first 11 inodes were reserved, they could be used right away. Instead, if I want to use one I have to first initialize it. I'll try and let you know. Another question: do reserved inodes need to be referenced by a directory entry? I would say it depends on how you are using it (btw you haven't told us yet why you need a reserved inode :-) ). For example root inode needs to be referenced by a directory entry but journal inode doesn't Thanks - Manish Thanks a lot, Donato On Sun, 2009-07-05 at 22:37 +0530, Manish Katiyar wrote: On Sun, Jul 5, 2009 at 10:28 PM, Donato Capitellad.capite...@tin.it wrote: Well, as it's a reserved inode, I didn't create it. stat 3 shows that the inode has bad type... so this might be the problem. But how am I supposed to initialize the inode fields if I can't read it? What are you reserving it for ? Note that the reserved inodes today (2 for root) (8 for journal commonly) both are initialized during mkfs. You also need to do something similar. Depending on what you are using it for you may need to initialize it properly. To do that you can do the following :- a) Either write a simple C program to write the fields properly at correct offset. b) Hack the code of mkfs to include your inode too. c) Use debugfs to change the contents of inode 3. This is the easiest option but I am not sure if you can modify all the fields of the inode, so it depends on what you need. Thanks - Manish On Sun, 2009-07-05 at 22:06 +0530, Manish Katiyar wrote: On Sun, Jul 5, 2009 at 7:30 PM, Donato Capitellad.capite...@tin.it wrote: Hi there, I need some help with a patch to the ext3 filesystem I'm developing. I'm completelly new to kernel programming and I need to understand practically how to use one of the ext3 reserved inode for storing a log. I thought it would be just as easy as getting the inode with the ext3_iget() function, but it doesn't seem to be the case. For example, if I try to load the third reserved inode like this : inode = ext3_iget(sb, 3); ext3_iget() fails somehow and crashes the process with a NULL pointer reference. What am I missing here? Have you created the inode 3 ? What entries does it have ? What does stat 3 show from the debugfs ? Thanks - Manish How do you use a reserved inode? Can anybody point to some documentation or project that uses reserved inodes? I'm really stuck :( Thanks a lot, Donato Capitella -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How the kernel printk works before do console_setup.
On Wed, Jun 24, 2009 at 9:20 AM, Johnny Hungjohnny.hack...@gmail.com wrote: Hi All: I have a powerpc arch platform. The default console is ttyS1 not ttyS0 in the platform. My question is how the kernel print debug message like DBG before parse kernel command line and do console_setup function. The command line passed to kernel about console info is console=ttyS1. So kernel can not print anything before parse cmd line and initial console but the result is against. Anyone point me or give me a clue is appreciate. I think you are looking for early_printk. http://mail.nl.linux.org/kernelnewbies/2005-10/msg00176.html Thanks - Manish BR, H. Johnny -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: how to make system calls from kernel module?
On Sat, Jun 20, 2009 at 10:29 PM, krushnaal paikrisonea...@gmail.com wrote: how to make system calls from kernel module? or how to include library headers in modules so as to use open,read write Why do you need that ? What are you trying to do ? Isn't this a repeated question ?? Thanks - Manish functtions? -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Recovering deleted file by editing inode..
On Tue, May 26, 2009 at 8:03 AM, ranjith kannikara ranjithkannik...@gmail.com wrote: Hi, First we were using the tool logdump in debugfs, and now we use the jls and jcat. The recovery works fine now it shows the recover file atonce and is available also. Can you post what was the problem and what was the fix. This would help others too -Manish rgards ranju. -- http://www.ranjithkannikara.blogspot.com/ -- Thanks - Manish
Re: Recovering deleted file by editing inode..
On Mon, May 25, 2009 at 9:04 AM, ranjith kannikara ranjithkannik...@gmail.com wrote: Hi, We are a team of prefinal year computer science engineering students from kerala.We are trying to design an application which can recover deleted data from the ext3 filesystem. And we are doing it by editing the inode of the deleted file with the help of debugfs. As you told the 'modify_inode' in debugfs will be help ful we have written code to recover data . We could recover files of fairly larger size, we tried recovering files over 1Gb and we are sure to recover files of 4Gb in size if its not over-written. But in the middle we are having little doubts and little problems in recovery. Like , after we recover the file,it appears in the disk as not accessible but when we unmout and remount the device the file is available. Btw, i think debugfs is not safe to use on a mounted file system since it interacts with raw devices . Someone will correct me if i am wrong. I shallbrief what we are doing, in the following lines. Please do go through it if you see it interesting. We have regestered the project in sourceforge and we will be uploading the code soon so that you can have your advices if you are interested. * useing debugfs list the deleted files and their inode and select the file to be recovered. * using logdump the details of the file inode, journal entry, size, links , blockcount. *if logdump yields a number of entries of none-zero size, the appropriate one is selected. *then the inode is set using command 'seti' *the inode is modified with the direct and indirect pointers which are taken from the journal. *now the inode is linked to a file in name of the deletd one. Here when the file is recovered it is appearing in the device but when we click on it, it will disappear Before clicking can you try few commands like ls, ls -l ls dir if it is a dir etc. on it ?? but if the device is unmounted and remounted again, the file will behave as a usual file itself. And if we ever delete a file which is recovered like this then all other files in the device will become read-only , untill it is remounted. do you get some error message in dmesg ?? I suspect this is because you have mounted as errors=remount-ro You must look at the sources of ext3grep or extundelete on sourceforge. Thanks - Manish Regards, Ranju. -- http://www.ranjithkannikara.blogspot.com/ -- Thanks - Manish
Re: using git
On Mon, May 18, 2009 at 11:20 AM, SandeepKsinha sandeepksi...@gmail.com wrote: Hi all, I have recently started using git... I know quite late. Say, I make some changes and create a patch. I want to revert back the files to the original/initial revision( when I created git ), how to do that? man git-reset Thanks - manish I tried git checkout -f This didn;t work for me. -- Regards, Sandeep. “To learn is to change. Education is a process that changes the learner.” -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: Getting input from USB HID foot pedal. Where to get started?
On Mon, May 18, 2009 at 8:02 AM, thzxcyl thzx...@gmail.com wrote: I have a foot pedal that is implemented as a USB HID device, and it sends one input: in binary, 0001, 0010, or 0100 corresponds to each pedal, respectively; combination of pedal presses sends combination of those binary number, and releases of pedal trigger input of whatever pedal still being pressed (if all pedal is up, is sent). My intention is to have the pedal buttons behave as normal keyboard keys (ctrl, alt, and pgdn). I've read many writeup on related topics, some of them a few years old, and I'm still confused as to where to start: do I need to patch the kernel? write a driver translate the input to keycode? do both? I would really appreciate a pointer to some tutorials on how to get started and example codes if there are any. Will this help ? http://www.linuxjournal.com/article/7353 Thanks - Manish Many thanks, T. -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How to set env variable for an already running shell from other shell ?
On Wed, May 6, 2009 at 2:55 PM, yogeshwar sonawane yog...@gmail.com wrote: Hi all, Can one export/set some variables for an already running shell from other shell ? Is there any command where i can specify env=value shell process id ? This is similar to sending signals from one to another process. Assuming it is C program, why can't you send the signal. just write a signal handler to enable the debug variable when the signal is received. We have one program, which prints some debugging information when a env variable is set. During execution, first a new shell is created then program is executed. Sometimes that program gets stuck somewhere without printing anything. If i set the env variable then execute, the messages will come from start of execution. I want some mechanism so that when program is stuck, You can set the variable by attaching the process from gdb, isn't it ?? i can set that env variable which will enable printing debugging information. Is this possible using some command ? OR how to do this ? This question is somewhat out of topic(not related to kernel). But, still help will be appreciated. I guess more details about the program (language, what it is trying to do) would help others to answer the question better. Thanks - Manish TIA, Yogeshwar -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- Thanks - Manish -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: how to kill defunct process ?
On Thu, Apr 9, 2009 at 6:48 PM, yogeshwar sonawane yog...@gmail.com wrote: Hi all, If a process is stuck and in defunct state, how to kill it ? Finally, we had to reboot the node. Aren't defunct processes already killed and just waiting for their parents to pick up their state ? We faced such problem and kill -9 pid was not working. Obviously killing a dead person :-) any other way out ? IIRC init used to do the cleanups for all such processes. Thanks - Manish Thanks, Yogeshwar -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
Re: filesystem project idea
On Wed, Mar 25, 2009 at 8:45 AM, Onkar onkar@gmail.com wrote: Hi All, Can anybody help me with project ideas on Filesystems,RAID,LVM etc. to which I can contribute. TUX3 Thanks - Manish regards, Onkar -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to ecar...@nl.linux.org Please read the FAQ at http://kernelnewbies.org/FAQ
