Hi
The script for testing strace is
qemu_multiarch_testing/start-tests.sh, for testing on the current
architecture just run it (./start-tests.sh) if the current
architecture supports multiple personalities it will test on all of
them by building ltp for each one. For other architectures it will use
Hi all,
I updated all my work in my gist[0], my patch is based on git commit:
"57fac75 xtensa: sort values in struct_user_offsets". I found strace
4.9 is just released and it's about 20 commits ahead of my work. I
will continue my work and rebase my patch to the latest strace 4.9.
You can find a
Hi all,
here is my update to the JSON support for strace, I upload some test
output and diff patch to gist for your review, you can see it all
here:
https://gist.github.com/zym0017d
there are 3 kind files in my Gists:
1, you can see the test outputs with and without option '-j' for a
comparison.
Hi all,
I redesign all my framework last week because I found it's getting
more and more complicated,
and It seems we really do not need so much extra extensibility.
For example, the old framework allow us to change the separator of the
arguments in syscalls
by just modifying one place. In order t
On Mon, Jul 28, 2014 at 8:44 AM, yangmin zhu wrote:
> Hi Dmitry,
>
> I'm working to modify strace to support JSON, but I just encounter a
> big problem recently and I could not find
> out a suitable way to get through.
>
> the problem is:
> 1), as you know, I designed a framework to support the JS
Hi Dmitry,
I'm working to modify strace to support JSON, but I just encounter a
big problem recently and I could not find
out a suitable way to get through.
the problem is:
1), as you know, I designed a framework to support the JSON output and
the core of the framework is to hook the metadata pri
Hi
I was trying to run the parallel-build-hdc-img.sh script in the
qemu_multiarch_testing. According to the readme has to be run as:
./parallel-build-hdc-img.sh system-image-ARCH.
I got some system images but It needs a script called native-build.sh
but what is it about?
--
The JSON output format is used to make it easier for other program
to parse strace's results. We need to modify the existing code to
use this basic framework to support the JSON output.
* defs.h: Add the interface of the JSON output framework.
* json.c(newfile): The Implmentation of JSON output fr
nstall wdiff and corlordiff to make the output
look better.
Before the deadline of GSOC 2014(August 18), I will probably not
modify the base framework any more(or just some small
refactor), And I will put my energy to modify the existing strace code
to add JSON support for those sys_* functions i
A special test file for syscalls in io.c, I tried to make the trace
results as stable as each time we execute the program.
* test/.gitignore: Add json_io.
* test/.Makefile: Add json_io.c to compile list.
* test/json_io.c(newfile): contain test cases for syscalls in io.c,
The only random number wou
Each syscall in the output will be a object in a single line.
The object has at least 3 attributes:"type", "name", "args" and "ret".
the vaule of "args" is an array and the output now only support string
as the basic value.
* syscall.c (trace_syscall_entering, trace_syscall_exiting):
Add support
Modify strace to use the JSON framework.
* Makefile.am: Add json.c to compile list.
* strace.c(tprintf, tprints, usage, init): Add a new option '-j'
in init(), Modify tprintf/tprints to use the basic JSON framework,
Add 2 new function:_tprintf/_tprints for the original output.
* strace.1: Add docu
We will do 3 compare to check the output:
1) Compare the Unmodified strace output to the Normal outputs;
2) Compare the JSON parsed results to the Normal outputs(ignore space);
3) Compare the JSON parsed results to the Unmodified strace outputs(ignore
space);
The 1st comparison should be identica
Modify all the syscalls in io.c to support JSON output.
There are mainly two kind modification to strace:
1) simply modify sys_* functions to use the JSON framework;
2) modify those format/detail funcitons(such as printfd, printlval etc..)
to use the JSON framework;
* io.c (sys_read/write, sys_rea
There are only 3 kind of '+++' signals in strace.
Note: Currently do NOT support the '/' in
JSON format. And this will make the those JSON lines invalid(or you can
just ignore these lines). Need to find a way to solve this problem.
* process.c(sys_exit): slightly reformat the code.
* syscall.c(tra
Hi
I updated the script and changed the way the script calls the LTP
tests. I was using the LTP interface to call the tests but there was
no simple way to generate a strace output for each test. Now I the
tests are run directly and each strace test output is saved in a file.
I had to process the fi
> There was also similar code in the LTP project (as opposed to sandbox?)
Yes, I worked with the LTP and commited the tests to my repo [1]. The
script that starts testing is (surprisingly):
strace/qemu_multiarch_testing/start-tests.sh
with no args it just tests in the actual architecture and with
On Fri, Jul 4, 2014 at 3:45 PM, Edson Ticona wrote:
> I read this post [1] and I was thiking to adapt the syscalls tests in
> the LDP so that they could be run in a multiarch environment using
> qemu; based on the actual qemu_multiarch_test. The script could be
> something like this:
>
> $ test --
Hi Dmitry,
>
> In this example, since va_list processing is easier and more efficient
> than string parsing, I'd recommend
> json_printf(JSON_ARG, JSON_SEP, 0L);
> instead of
> json_printf("arg sepa");
>
> Also, when a function name ends with "printf", it is expected to take a
> pr
Hi
I read this post [1] and I was thiking to adapt the syscalls tests in
the LDP so that they could be run in a multiarch environment using
qemu; based on the actual qemu_multiarch_test. The script could be
something like this:
$ test --syscall=open,close,fstat --arch=i386,x32
[1]
http://source
Hi Edson,
On Mon, Jun 30, 2014 at 06:34:28PM -0500, Edson Ticona wrote:
> Hi
> I could convert i386 binary to x86_64 by using:
> objcopy -O elf64-x86-64 sigaltstack-i386.i sigaltstack-x86_64.i
Conversion of an elf32-i386 format object to an elf64-x86-64 format object
doesn't convert x86 code to x
Hi YangMin,
On Sun, Jun 29, 2014 at 09:23:12PM +0800, yangmin zhu wrote:
[...]
> 1) I had ever thought about to hide all JSON related details from
> syscall parsers, such as to use functions like
> print_arg_fd(tcp->u_arg[0]);
> and I even implemented a simple framework for that. But I changed not
On Tue, Jul 1, 2014 at 1:34 AM, Edson Ticona wrote:
> Hi
> I could convert i386 binary to x86_64 by using:
> objcopy -O elf64-x86-64 sigaltstack-i386.i sigaltstack-x86_64.i
> then, compiling goes fine like this:
> gcc -Wall -Wwrite-strings -g -O2 -o strace aio.o bjm.o block.o
> count.o desc.o fa
Hi
I could convert i386 binary to x86_64 by using:
objcopy -O elf64-x86-64 sigaltstack-i386.i sigaltstack-x86_64.i
then, compiling goes fine like this:
gcc -Wall -Wwrite-strings -g -O2 -o strace aio.o bjm.o block.o
count.o desc.o fanotify.o file.o inotify.o io.o ioctl.o ioprio.o ipc.o
kexec.o key
Hi Dmitry,
Thank you for your review.
On Fri, Jun 27, 2014 at 10:02 AM, Dmitry V. Levin wrote:
> Hi YangMin,
>
> Thanks for these examples. Let's have a look at the first one:
>
>> int
>> sys_read(struct tcb *tcp)
>> {
>> JSON_BEGIN_META_MODE;
>>
>> if (entering(tcp)) {
>>
Hi YangMin,
On Thu, Jun 19, 2014 at 12:29:49AM +0800, Zhu YangMin wrote:
> The basic means that all the arguments are simply wrapped in
> a string value even for those complicated arguments printed by
> specific functions such as printflags(). We need to do more work
> to change those functions to
Hi Edson,
On Thu, Jun 26, 2014 at 01:27:24PM -0500, Edson Ticona wrote:
> Hi
> I am refactoring in such a way that functions relaying on specific
> personalities are isolated on its own file. The patch has the concrete
> case of printllval function which relies on different architectures
> types;
On Thu, Jun 26, 2014 at 8:27 PM, Edson Ticona wrote:
> Hi
> I am refactoring in such a way that functions relaying on specific
> personalities are isolated on its own file. The patch has the concrete
> case of printllval function which relies on different architectures
> types; I removed it from u
Hi
I am refactoring in such a way that functions relaying on specific
personalities are isolated on its own file. The patch has the concrete
case of printllval function which relies on different architectures
types; I removed it from util.c and inserted the adequate version on
linux//autil.c
The Ma
---
Makefile.am | 3 ++-
defs.h | 11 ++-
util.c | 54 --
3 files changed, 8 insertions(+), 60 deletions(-)
diff --git a/Makefile.am b/Makefile.am
index a789b8c..94aec57 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -50,7 +50,
---
linux/aarch64/autil.c| 28
linux/alpha/autil.c | 19 +++
linux/arc/autil.c| 19 +++
linux/arm/autil.c| 25 +
linux/avr32/autil.c | 19 +++
linux/hppa/autil.c
---
linux/aarch64/autil.c| 28
linux/alpha/autil.c | 19 +++
linux/arc/autil.c| 19 +++
linux/arm/autil.c| 25 +
linux/avr32/autil.c | 19 +++
linux/hppa/autil.c
Hi Dmitry,
On Thu, Jun 19, 2014 at 2:01 AM, Dmitry V. Levin wrote:
> Hi YangMin,
>
> On Wed, Jun 11, 2014 at 11:10:18PM +0800, yangmin zhu wrote:
> [...]
>> As Google says; "Mentors and students can begin submitting mid-term
>> evaluations." in [2],
>> I'm not sure what should I do or prepare for
Hi YangMin,
On Wed, Jun 11, 2014 at 11:10:18PM +0800, yangmin zhu wrote:
[...]
> As Google says; "Mentors and students can begin submitting mid-term
> evaluations." in [2],
> I'm not sure what should I do or prepare for the mid-term evaluations?
There seems to be nothing special, just a list of q
The JSON format is used to make it more easy to parse the
trace results by other program. It is enabled by -j option.
* Makefile.am: Add json.c to compile list.
* strace.1: Add documents for -j option.
* strace.c(init, usage, tprintf, tprints, tvprintf): Add support
for the '-j' option. Modify tpr
We will do 3 compare to check the output:
1) Compare the Unmodified strace output to the Normal outputs;
2) Compare the JSON parsed results to the Normal outputs(ignore space);
3) Compare the JSON parsed results to the Unmodified strace outputs(ignore
space);
The 1st comparison should be identica
Now the JSON output is one dict object each line, representing a
the enter/exit of a syscall. The dict object contain key filed
such as 'type', 'name', 'ret' of a syscall. All the value are string.
* syscall.c(trace_syscall_entering, trace_syscall_exiting): Modify to support
in JSON format.
* proc
The basic means that all the arguments are simply wrapped in
a string value even for those complicated arguments printed by
specific functions such as printflags(). We need to do more work
to change those functions to be more like JSON-style.
* io.c(sys_read, sys_write, sys_readv, sys_writev,
sys_
Hi all,
I had cleaned and refactored my code and I think it's time to send the patches
to the mailing list for a review.
The code are still immature and need to change a lot and I send them
here mainly for
discussion before my next step, I think it's necessary to get your
feedback and suggestions,
Hi all,
I just finished a big refactor to my code and I had pushed them all
to my github[1].
And before the mid-term evaluation(23 June) I will try to work out a
usable prototype at least
supporting the basic JSON output of entering/exiting and syscalls in io.c.
As Google says; "Mentors and stude
On Wed, Jun 4, 2014 at 4:26 AM, Philippe Ombredanne
wrote:
> On Tue, Jun 3, 2014 at 5:43 PM, yangmin zhu wrote:
>> I'm not sure whether should I send my patches directly to this mailing
>> list or just put these patch in my github
>> and we discuss here?
>>
>> you can always find my latest commit
conver the output of stacktrace?
>
I'm now working on the relatively stable parts of strace, for example
I'm modifying those functions in io.c which didn't change rapidly.
I'm not sure whether I have enough time to cover the output of
stackstrace before the pencil down time o
Hi, YangMin
I have not read the thread about your work, JSON format because I
have concentrated on unwind(stacktrace) patch. (The patch is finally
merge into the master tree!)
Will your work conver the output of stacktrace?
If yes, please, read the rest of this mail.
Currently the stacktrace
On Tue, Jun 3, 2014 at 5:43 PM, yangmin zhu wrote:
> I'm not sure whether should I send my patches directly to this mailing
> list or just put these patch in my github
> and we discuss here?
>
> you can always find my latest commits here:
> https://github.com/zym0017d/strace_GSOC/commits/master
Y
On Sat, May 31, 2014 at 8:21 AM, Dmitry V. Levin wrote:
>
> YangMin, please enable pre-commit hook on your side, it will stop
> whitespace errors that make all your commits look red:
>
> $ mv .git/hooks/pre-commit{.sample,}
>
I had cleared all the white space error in my commit history when made
On Sat, May 31, 2014 at 8:21 AM, Dmitry V. Levin wrote:
> On Fri, May 30, 2014 at 12:36:25AM +0800, yangmin zhu wrote:
>> Hi everyone,
>> I learned a lot about git from the great book [1] and I had cleared my
>> repository in github[2].
>
> YangMin, please enable pre-commit hook on your side, it w
On Fri, May 30, 2014 at 12:36:25AM +0800, yangmin zhu wrote:
> Hi everyone,
> I learned a lot about git from the great book [1] and I had cleared my
> repository in github[2].
YangMin, please enable pre-commit hook on your side, it will stop
whitespace errors that make all your commits look red:
Hi everyone,
I learned a lot about git from the great book [1] and I had cleared my
repository in github[2].
I splited my current work into 3 commits:
1. Add basic hook framework for JSON output and the support for '-j' option. [3]
2. Modify sys_read(),sys_write() in io.c to use this hook framework
On Wednesday, 21 May 2014, Philippe Ombredanne wrote:
> On Wed, May 21, 2014 at 2:27 PM, Zubin Mithra
> >
> wrote:
> > Hello everyone!
> >
> > My name is Zubin Mithra and my tasks as part of GSoC include the
> following
> > :-
> > - Improving path decoding
> > - Improving security in the strace
On Wed, May 21, 2014 at 2:27 PM, Zubin Mithra wrote:
> Hello everyone!
>
> My name is Zubin Mithra and my tasks as part of GSoC include the following
> :-
> - Improving path decoding
> - Improving security in the strace codebase
>
> The repository I will be pushing code to can be found here[1]. I'
Hello everyone!
My name is Zubin Mithra and my tasks as part of GSoC include the following
:-
- Improving path decoding
- Improving security in the strace codebase
The repository I will be pushing code to can be found here[1]. I'll push
updates about weekly progress onto the wiki page, and provid
On Tue, May 20, 2014 at 9:19 AM, Philippe Ombredanne
wrote:
> On Tue, May 20, 2014 at 5:59 PM, yangmin zhu wrote:
>>I am really new to open source project
>> and feel some kind lost about how the open source development work.
>
> Hang in there, you are here to learn... ;)
>
>> I found the strace
On Tue, May 20, 2014 at 5:59 PM, yangmin zhu wrote:
>I am really new to open source project
> and feel some kind lost about how the open source development work.
Hang in there, you are here to learn... ;)
> I found the strace code changing every week. And I will merge the
> latest strace code to
On Mon, May 19, 2014 at 6:15 PM, Philippe Ombredanne
wrote:
> On Sun, May 18, 2014 at 6:24 PM, yangmin zhu wrote:
>
> To ensure that your code can be reviewed and eventually submitted as
> patches you would need to start from a clone of the strace git repo.
> You seem to have no history in github
On Sun, May 18, 2014 at 6:24 PM, yangmin zhu wrote:
> I am now doing all my work on my own github
> ( https://github.com/zym0017d/strace_GSOC2014 ) and I am getting
> familiar with git.
To ensure that your code can be reviewed and eventually submitted as
patches you would need to start from a clo
On Sun, May 18, 2014 at 6:24 PM, yangmin zhu wrote:
[...]
> I am now doing all my work on my own github
> ( https://github.com/zym0017d/strace_GSOC2014 ) and I am getting
> familiar with git.
Sorry, in my previous post I had not seen that part Using this
repository is fine.
> So should I sub
On Sun, May 18, 2014 at 6:24 PM, yangmin zhu wrote:
> Hi all,
Hi Yang-Min!
You are new to the list: we do not top post on this list. This means
that you should quote the subset of text your reply is about; and put
the reply _after_ the quoted text you are replying to. See the lkml
style FAQ: we u
o use open_memstream().
On Fri, May 16, 2014 at 5:02 AM, Philippe Ombredanne
wrote:
> Sorry for having been a tad quiet lately ...
>
> We are rather spoiled by having three projects allocated to strace for
> the GSOC 2014.
> The program starts in earnest now and kicks in officially on
On Mar 21, 2014, at 8:25 AM, Philippe Ombredanne wrote:
> On Fri, Mar 21, 2014 at 11:50 AM, Zev Weiss wrote:
>> On Mar 20, 2014, at 12:54 PM, yangmin zhu wrote:
>>> Hi,
>>> I'm yangmin zhu. I'm a master student from University of Chinese Academy of
>>> Sciences and now I'm participating in th
Expanding specifically on the JSON streaming idea;
(Sorry if I'm rehearsing ideas already stated, I just subscribed
yesterday and only glanced at the recent archives)
2014-03-21 6:50 GMT-04:00 Zev Weiss :
> (Though w.r.t another aspect of Marc-Antoine's comment -- JSON doesn't
> necessarily have
On Fri, Mar 21, 2014 at 11:50 AM, Zev Weiss wrote:
> On Mar 20, 2014, at 12:54 PM, yangmin zhu wrote:
>> Hi,
>> I'm yangmin zhu. I'm a master student from University of Chinese Academy of
>> Sciences and now I'm participating in the Google Summer of Code 2014.
>> I'm working for the strace pro
On Fri, Mar 21, 2014 at 1:22 PM, yangmin zhu wrote:
> Hi Ombredanne,
> Sorry for the HTML email. it's a mistake.
> I had submitted my proposal to GSOC and you can find a basic prototype
> in my proposal. In order to make it better, I contacted some strace's
> output parser's authors to collect the
On Tue, Mar 18, 2014 at 9:09 PM, Philippe Ombredanne
wrote:
> yangmin zhu:
> I think you are the right track!
> This is a very thorough approach you are taking there which is very good!
>
> Small note: try avoid using HTML emails and stick to plain text on the list.
>
>
>> From this mail
>> (http:
(CCing strace-devel as well, hope that's OK...)
On Mar 20, 2014, at 12:54 PM, yangmin zhu wrote:
> Hi,
> I'm yangmin zhu. I'm a master student from University of Chinese Academy of
> Sciences and now I'm participating in the Google Summer of Code 2014.
> I'm working for the strace project abo
On Thu, Mar 13, 2014 at 6:29 PM, Marcin Grzywaczewski
wrote:
> Hello!
>
> My name is Marcin Grzywaczewski, and I'm studying computer science on
> University of Wroclaw.
Hi Marcin and welcome to strace!
> I'm very interested in working with Strace organisation within Google
> Summer of Code 2014.
On Thu, Mar 13, 2014 at 4:55 PM, yangmin zhu wrote:
> I had spent some day reading the code of strace to grasp the big picture
> of strace.my main purpose is first to find out how did strace dispatch the
> syscall to each printing function. I found that the main function ( in
> file:strace.c) firs
Hello!
My name is Marcin Grzywaczewski, and I'm studying computer science on
University of Wrocław.
I'm very interested in working with Strace organisation within Google
Summer of Code 2014. It's a fantastic tool which I'm using often (it's
great when it comes to finding strange glitches on se
Hi all,
I had spent some day reading the code of strace to grasp the big picture
of strace.my main purpose is first to find out how did strace dispatch the
syscall to each printing function. I found that the main function ( in
file:strace.c) first call init(argc, argv) to initialize some importa
I find
> strace which I used a lot in my undergraduate graduation design. I wish to
> participate GSoC 2014 of the strace project.
>
> I ever implemented a small tool to intercept(by using LD_PRELOAD method) a
> program’s syscall (read()/write()/dup() and etc.) and pass it to a rem
wish to
participate GSoC 2014 of the strace project.
I ever implemented a small tool to intercept(by using LD_PRELOAD method) a
program's syscall (read()/write()/dup() and etc.) and pass it to a remote
host(yes, another server program is running at that remote host and
listening on a port sp
On Sun, Mar 2, 2014 at 12:43 PM, ChenQin wrote:
> HERE IS MY REQUEST FOR GSOC 2014~
> Name:ChenQin
> Title:Downstream patches roundup and ushering
>
> Abstract : The goal of this project would to review available patches and
> bugs from the key downstream maintainers and eithe
HI~HERE IS MY REQUEST FOR GSOC 2014~Name:ChenQinTitle:Downstream patches roundup and usheringAbstract : The goal of this project would to review available patches and bugs from the key downstream maintainers and either contribute these as patches to strace when relevant or work with the downstream
On Sun, Mar 2, 2014 at 4:26 AM, Sanidhya Kashyap
wrote:
> My name is Sanidhya, and I am Masters student at IIIT Hyderabad. I have been
> working in the area of OS and Virtualization from the past two and half
> years.
>
> This year, I am interested in participating in Google Summer of Code and I
>
Hi,
My name is Sanidhya, and I am Masters student at IIIT Hyderabad. I have
been working in the area of OS and Virtualization from the past two and
half years.
This year, I am interested in participating in Google Summer of Code and I
came across 'strace' which I happen to use regularly. I have g
On Wed, Feb 26, 2014 at 7:32 AM, Rohan Saxena wrote:
> I am Rohan Saxena. Currently I am pursuing my B.E.(Hons) in computer science
> from BITS Pilani India. I am deeply interested in system programming and
> want to contribute to strace as part of GSOC 2014.
> Can someone please ge
Hi,
On Wed, Feb 26, 2014 at 12:02:09PM +0530, Rohan Saxena wrote:
> Hi Everyone
>
> I am Rohan Saxena. Currently I am pursuing my B.E.(Hons) in computer
> science from BITS Pilani India. I am deeply interested in system
> programming and want to contribute to strace as pa
Hi Everyone
I am Rohan Saxena. Currently I am pursuing my B.E.(Hons) in computer
science from BITS Pilani India. I am deeply interested in system
programming and want to contribute to strace as part of GSOC 2014.
Can someone please get me started and show me the right direction
77 matches
Mail list logo
