MSI-X Vectors and NUMA
Hi all, I want to understand how MSI-X Vectors are allocated between Processors for SMP systems for NUMA operations? Also , How much MSI-X vectors are allocated Per -CPU? Can anyone Please redirect to me a good source of MSI-X vectors in Linux Kernel? -- Warm Regards, Ravi . "I don't Know is not a excuse , its an opportunity." ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Storage and I/O subsystem
Hello all, I want to start learning the Linux Block I/O stack about how it handles I/O requests and about the storage subsystem .anybody has any documents with that ? any pointers for where to start ? and how to start? thanks. -- Warm Regards, Ravi . "I don't Know is not a excuse , its an opportunity to know." ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: Measuring time in range of microseconds.
Hello Satish, On Wed, Dec 7, 2011 at 12:41 PM, Sathishkumar Duraisamy wrote: > Hi Fredrick, > > On Wed, Dec 7, 2011 at 8:12 AM, Fredrick wrote: >> I think if the specific ARM platform you are working supports a oneshot >> clock_device, you would get the high res timer support. >> >> You can check for clock_devices defined in your platform having feature - >> CLOCK_EVT_FEAT_ONESHOT. > My guess is that u want to measure the time taken to execute a function in microseconds. something like start_time YOUR_FUNCTION end_time for this i use something like this declare struct timeval tv in your function and insert printk statements before and after calling to your function do_gettimeofday(&tv); printk ("Before %s (): Sec = %dmicro Sec = %d\n", __FUNCTION__, tv.tv_sec, tv.tv_usec); call_to_your_function((); printk ("After %s (): Sec = %dmicro Sec = %d\n", __FUNCTION__, tv.tv_sec, tv.tv_usec); -- Warm Regards, Ravi . "I don't Know is not a excuse , its an opportunity." ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: understanding the SCSI Specification.
Hi On Tue, Aug 2, 2011 at 9:36 PM, amit mehta wrote: > Linux SCSI layer, specifically SCSI mid layer is the place where > all the conversion happens yeah im trying to understand how the abstraction is provided . . There is a Linux - SCSI mailing list > (linux-scsi) where you'll find relevant discussions. I have posted there also but no luck:( > > On Tue, Aug 2, 2011 at 12:15 PM, RKK wrote: >> Hi Amit >>> download sg3-utils and its source code. good place to begin with. >> >> Thanks for the reply. >> I believe the sg3-utils uses the linux sg driver which does not >> provide any abstraction and sends commands directly to the device. >> what i was looking was more of protocol specific information and >> explaination of the flow like how block device read , write requests >> are converted into Commands and how those SCSI commands are passed to >> the lower layer. My target is FPGA board which has a pci interface for >> host controller . can u give me some pointers to understand the below >> ? >> >> Warm Regards, >> Ravi Kulkanri. >> >>> >>> On Mon, Aug 1, 2011 at 2:41 PM, RKK wrote: >>>> HI all, >>>> Im working on a storage protocol which uses SCSI commands set as >>>> the command set and also the SCSI Task Management functions for >>>> processing the Task Management. >>>> The Architectural Model is based on SAM-5. The Command set is based on >>>> SPC-4 and SBC-3. I have some doubts in trying to understanding some >>>> SCSI terms and definitions. >>>> >>>> For using the send command transport protocol service request to >>>> request a initiator the following API is defined >>>> >>>> [1]Send SCSI Command (IN (I_T_L_Q Nexus, CDB, Task Attribute, [Data-in >>>> Buffer Size], [Data- >>>> out Buffer], [Data-out Buffer Size], [CRN], [Command Priority], [First >>>> Burst Enabled])) >>>> >>>> Here I want to know what relation does the I_T_L_Q tries to convey. >>>> from what it says in SAM-5 Specification section 4.8 tries to define >>>> the Nexus as " nexus represents a relationship between a SCSI >>>> initiator port, a SCSI target port, optionally a logical unit, >>>> and optionally a command." >>>> how do we convey this information in the above [1]. >>>> Also can someone help me in how to map the initiator-target relation >>>> to host-device structure? thanks. >>>> >>>> Warm Regards, >>>> Ravi Kulkarni. >>>> >>>> >>>> >>>> >>>> >>>> Warm Regards, >>>> Ravi Kulkarni. >>>> >>>> ___ >>>> Kernelnewbies mailing list >>>> Kernelnewbies@kernelnewbies.org >>>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >>>> >>> >> >> >> >> -- >> Warm Regards, >> Ravi . >> "I don't Know is not a excuse , its an opportunity." >> > -- Warm Regards, Ravi . "I don't Know is not a excuse , its an opportunity." ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: understanding the SCSI Specification.
Hi Amit > download sg3-utils and its source code. good place to begin with. Thanks for the reply. I believe the sg3-utils uses the linux sg driver which does not provide any abstraction and sends commands directly to the device. what i was looking was more of protocol specific information and explaination of the flow like how block device read , write requests are converted into Commands and how those SCSI commands are passed to the lower layer. My target is FPGA board which has a pci interface for host controller . can u give me some pointers to understand the below ? Warm Regards, Ravi Kulkanri. > > On Mon, Aug 1, 2011 at 2:41 PM, RKK wrote: >> HI all, >> Im working on a storage protocol which uses SCSI commands set as >> the command set and also the SCSI Task Management functions for >> processing the Task Management. >> The Architectural Model is based on SAM-5. The Command set is based on >> SPC-4 and SBC-3. I have some doubts in trying to understanding some >> SCSI terms and definitions. >> >> For using the send command transport protocol service request to >> request a initiator the following API is defined >> >> [1]Send SCSI Command (IN (I_T_L_Q Nexus, CDB, Task Attribute, [Data-in >> Buffer Size], [Data- >> out Buffer], [Data-out Buffer Size], [CRN], [Command Priority], [First >> Burst Enabled])) >> >> Here I want to know what relation does the I_T_L_Q tries to convey. >> from what it says in SAM-5 Specification section 4.8 tries to define >> the Nexus as " nexus represents a relationship between a SCSI >> initiator port, a SCSI target port, optionally a logical unit, >> and optionally a command." >> how do we convey this information in the above [1]. >> Also can someone help me in how to map the initiator-target relation >> to host-device structure? thanks. >> >> Warm Regards, >> Ravi Kulkarni. >> >> >> >> >> >> Warm Regards, >> Ravi Kulkarni. >> >> ___ >> Kernelnewbies mailing list >> Kernelnewbies@kernelnewbies.org >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >> > -- Warm Regards, Ravi . "I don't Know is not a excuse , its an opportunity." ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
understanding the SCSI Specification.
HI all, Im working on a storage protocol which uses SCSI commands set as the command set and also the SCSI Task Management functions for processing the Task Management. The Architectural Model is based on SAM-5. The Command set is based on SPC-4 and SBC-3. I have some doubts in trying to understanding some SCSI terms and definitions. For using the send command transport protocol service request to request a initiator the following API is defined [1]Send SCSI Command (IN (I_T_L_Q Nexus, CDB, Task Attribute, [Data-in Buffer Size], [Data- out Buffer], [Data-out Buffer Size], [CRN], [Command Priority], [First Burst Enabled])) Here I want to know what relation does the I_T_L_Q tries to convey. from what it says in SAM-5 Specification section 4.8 tries to define the Nexus as " nexus represents a relationship between a SCSI initiator port, a SCSI target port, optionally a logical unit, and optionally a command." how do we convey this information in the above [1]. Also can someone help me in how to map the initiator-target relation to host-device structure? thanks. Warm Regards, Ravi Kulkarni. Warm Regards, Ravi Kulkarni. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: Query regarding dma_alloc()
> Hi all, > What is the difference between the following memory allocations. > > 1) Memory allocation by doing, > char *buf; > buf = kmalloc(BUF_LEN, GFP_DMA | GFP_KERNEL); > > 2) And by > char *buf; > buf = dma_alloc(...); > > in the link http://www.arm.linux.org.uk/developer/memory.txt > it says that dma_alloc() returns virtual address in "ffc0-fffe" > this region, > > while in the first case it is returning Virtual address in "low-memory" > region. > > For DMA mappable memory which one we should use..? > Hi Sandeep Please have a look at these two API;s below. http://lxr.linux.no/#linux+v3.0/Documentation/DMA-API-HOWTO.txt http://lxr.linux.no/#linux+v3.0/Documentation/DMA-API.txt -- Warm Regards, Ravi Kulkarni ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: how to apply patches from git ?
Hi Prashant , > >>> >>> If its a public mailing list then I use a newsgroup reader that will >>> allow me to download the entire patch email as file and then use the >>> git apply on it. >>> >> may i know which newsgroup reader u use and how to setup that? thanks. >> > > I use Pan News reader > > http://pan.rebelbase.com/ > > Server : news.gmane.org > Linux Kernel : http://news.gmane.org/gmane.linux.kernel > thanks for that i have set up the pan news reader. i tried applying patch but it dint work. guess have to check git apply for that. also, how do u clone a particular git tree for a particular tag for example if i use git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git linux-2.6 it will clone the linus latest tree but if i want to clone only till rc4 or rc5 of his tree how to do it. Warm Regards, Ravi Kulkarni. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: how to apply patches from git ?
Hi Prashant, On Thu, Jul 14, 2011 at 1:46 PM, Prashant Shah wrote: > Hi Ravi, > > On Wed, Jul 13, 2011 at 12:52 PM, RKK wrote: >> Hi all, >> >> Sorry if this is a basic question. >> How do i take patches from e-mail? >> for example git send mail sends patches as e-mail >> >> if someone wants to apply that patch to his branch then how do i get >> that in . patch format and then use >> patch -p1 *.patch or is there something im missing here? thanks. > > If its a public mailing list then I use a newsgroup reader that will > allow me to download the entire patch email as file and then use the > git apply on it. > may i know which newsgroup reader u use and how to setup that? thanks. -- Warm Regards, Ravi Kulkarni ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: how to apply patches from git ?
Hi >>> Hi all, >>> >>> Sorry if this is a basic question. >>> How do i take patches from e-mail? >>> for example git send mail sends patches as e-mail >>> >>> if someone wants to apply that patch to his branch then how do i get >>> that in . patch format and then use >>> patch -p1 *.patch or is there something im missing here? thanks. >>> -- >>> Warm Regards, >>> Ravi . >>> >> >> Save the email message in mbox format and then you can use git am >> /path/to/mbox/file to apply the email patch to your tree. > > well, actually you don't need to do the whole mbox thing with git. You > only need to get the patches into .patch files in a directory. I believe > you can then use git am /path/to/patches and it will apply the patches > in the order the ls command displays them. are there any git tools to set up this so that i can use them ? > I'm assuming that would be helpful since most people who don't know how > to apply patches from mail also don't have an mbox. > makes sense -- Warm Regards, Ravi Kulkarni. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
how to apply patches from git ?
Hi all, Sorry if this is a basic question. How do i take patches from e-mail? for example git send mail sends patches as e-mail if someone wants to apply that patch to his branch then how do i get that in . patch format and then use patch -p1 *.patch or is there something im missing here? thanks. -- Warm Regards, Ravi . ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
how to implement a SCSI commands through a driver?
Hi all , Im trying to implement a custom new storage protocl which uses SCSI for storage in Linux . I have been looking around for some files in the Linux SCSI subsystems to understand its three layered design. The idea is to develop a separate stack which will internally use SCSI . so to get started all i want to do is write a module(kernel space) which sends SCSI commands to target device and retrieve some information from it. any idea like to how to proceed and which are files/books that can help me in this regard. I'm reading SCSI SPC-4 and SBC-3 specifications . ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies