April 2013 - kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

the github repo, and bugzilla.kernel.org
by Haix64 April 10, 2013

April 10, 2013

Hi. I was wondering how the "torvalds/linux" github repo, and the bugzilla.kernel.org are involved, if we can simply clone the repo from kernel.org and discuss things in the mailing lists. thank you.

2 1

open_softirq do?
by Robert Clove April 10, 2013

April 10, 2013

Hello All, I am clear all that the softirq is a software interrupt : Whenever a system call is about to return to the user space or h/w interrupt exit,ant s/w interrupt which are marked pending are run. Now in networking code i found open_softirq(NET_TX_SOFTIRQ, net_tx_action, NULL); What does this do? Thanks

2 1

At which Network Layer
by Robert Clove April 10, 2013

April 10, 2013

Hello All, There is a field in sk_buff structure name clone which tell whether the sk_buff structure has been cloned or not? I want to know at which layer does this clonning happen at L2,L3 or L4. Thanks

2 1

Linux from scratch
by Alexandru Juncu April 10, 2013

April 10, 2013

Hello! I am not a professional kernel developer but I do like learning how things work..especially operating systems like linux. I know the basis stuff about the linux kernel (enough to write a very simple kernel module). But I wanted to learn how things like a simple scheduler or a memory manager would work. Only reading the current kernel code is sometimes difficult because all the layers of abstraction implemented in order to make Linux code scale. So I was thinking taking a much much older version of the kernel and starting to analyze it. Something like...v0.01. Seems like I'm on a good track since sched.c in v0.01 has 250 lines while shed.h in the current kernel has 1200 :P. But I couldn't get v0.01 to compile and I've search the Internet and found that it needs a _lot_ of hacking to get it to work. So I moved to try 1.0 which seems to have all of the basic features implemented yet simple enough to understand. But I am still having trouble compiling (having trouble with header files...). Although I could just read the code and try to figure out what it does, I would like to be able to modify it and compile and run it. In order to avoid useless trial and error, I wanted to ask other people that maybe did this: What is the oldest kernel version that can be compiled on a current system and then pun on a virtual machine like qemu-kvm or vmware? On a side note, I did look into the "Linux from scratch" documentation project, but that seems to tell you how to build a distribution, not a kernel. Thank you.

4 5

Passthrough device driver performance is low on reads compared to writes
by neha naik April 9, 2013

April 9, 2013

Hi All, I have written a passthrough block device driver using 'make_request' call. This block device driver simply passes any request that comes to it down to lvm. However, the read performance for my passthrough driver is around 65MB/s (measured through dd) and write performance is around 140MB/s for dd block size 4096. The write performance matches with lvm's write performance more or less but, the read performance on lvm is around 365MB/s. I am posting snippets of code which i think are relevant here: static int passthrough_make_request( struct request_queue * queue, struct bio * bio) { passthrough_device_t * passdev = queue->queuedata; bio->bi_bdev = passdev->bdev_backing; generic_make_request(bio); return 0; } For initializing the queue i am using following: blk_queue_make_request(passdev->queue, passthrough_make_request); passdev->queue->queuedata = sbd; passdev->queue->unplug_fn = NULL; bdev_backing = passdev->bdev_backing; blk_queue_stack_limits(passdev->queue, bdev_get_queue(bdev_backing)); if ((bdev_get_queue(bdev_backing))->merge_bvec_fn) { blk_queue_merge_bvec(sbd->queue, sbd_merge_bvec_fn); } Now, I browsed through dm code in kernel to see if there is some flag or something which i am not using which is causing this huge performance penalty. But, I have not found anything. If you have any ideas about what i am possibly doing wrong then please tell me. Thanks in advance. Regards, Neha

1 0

Building multiple .ko files from a single module
by Sankar P April 9, 2013

April 9, 2013

Hi, I have some source files: a.c b.c c.c etc. Now I need to generate a few .ko files such as, 1.ko, 2.ko, 3.ko etc. I know the correspondence between the source files and the .ko file. Is there a way I could achieve this with a single Makefile ? All the tutorials and the help pages seem to focus on creating a single .ko file from a Makefile, as that is the usual case. However, I need to have different modules generated from my sources. For some shareability reasons, I cannot move the sources to different folders either. Can someone please show me how it should be done ? For example, consider the following association between source files and the .ko files. a.c a.h common.c common.h => 1.ko b.c b.h common.c common.h => 2.ko Thank you. -- Sankar P http://psankar.blogspot.com

3 3

Kernel generates vmImage instead of zImage
by Onur Cem Çelebi April 9, 2013

April 9, 2013

Hi all, I want to run my DE2-115 board for NIOSII processor with uClinux but stucked on somewhere. After i compiled the kernel, I got this message. Image Name: Linux-3.7.0-01377-g1061bd7 Created: Mon Mar 25 15:41:30 2013 Image Type: NIOS II Linux Kernel Image (gzip compressed) Data Size: 1445018 Bytes = 1411.15 kB = 1.38 MB Load Address: c8000000 Entry Point: c8000000 *Kernel: arch/nios2/boot/vmImage is ready* How to obtain zImage ? Am i missing something ? Any help is greatly appreciated. Cem Celebi

5 7

Interrupt Handler of Ethernet Device
by Robert Clove April 8, 2013

April 8, 2013

Hello All, I am new here. I want to know the interrupt handler of the ethernet card and where can i find the definition of it so as i can clear the flow of packet reception. Thanks Robert

8 12

Using vmalloc in device driver code
by neha naik April 5, 2013

April 5, 2013

Hi All, From what i have learnt vmalloc requires a minor page fault to sync the process page tables with the reference page tables. So, is it safe to use memory already allocated by vmalloc in io context/device driver code (especially when the io call has returned back to the device driver after finishing io on the device). I know that in general it is better to use kmalloc (for better efficiency and so on....) but, i want to know if it is inaccurate to use vmalloc in this particular case and why (in terms of cannot service page faults in interrupt context etc)? Thanks in advance for your help. Regards, Neha

3 3

NMIs, Locks, linked lists, barriers, and rculists, oh my!
by Arlie Stephens April 5, 2013

April 5, 2013

Hi Folks, I've got another "what's the natural way to do this in the linux kernel" question. As always, I'm attempting something bizarre, that doesn't seem to be covered in the newbie examples, and discussion with coworkers produced no consensus. I got such useful ideas last time I asked one of those question here, that I'm coming back for more. Here's the situation: - I've got a linked list, with entries occassionally added from normal contexts. - Entries are never deleted from the list. - I've got data structures which save pointers to various members of the list. - I've got routines which want to update a saved pointer to the next item in the list, or walk the whole list, or do other simple list operations. This would be simple, except that the routines which READ the list may get called from panic(), or inside an oops, or from an NMI. It's important that they succeed in doing their jobs in that case, even if they managed to interrupt a list addition in progress. Put another way, my problem is that I want the readers to be able to do their thing regardless of the state of the lock that serializes writers. In all cases, writers are protected from each other with a mutex. Here are some choices for making sure readers see something sensible: 1) Use an rculist, and hope that it really does cover this situation. 2) Be thankful I'm only running on x86, hope that the linux kernel isn't built with aggressive optimizations that reorder writes to pointers, fix __list_add() so it updates pointers in the new entry before updating pointers in its neighbours, and do the reads with no synchronization whatsoever. 3) Add my own explicit use of compiler directives, either to a copy of list.h or to a simpler, home-grown singly linked tail queue implementation. Because I'm on x86 all I really care about is compiler optimizations, but it looks like the linux way to do this may be with barrier macros. 4) Give up on always getting this to work from an NMI. Use something like a reader-writer spinlock, but if I'm in an NMI, make only a conditional attempt to acquire it - and don't try to walk the list if that acquisition fails (presumed self-deadlock) As always, my questions are: - Will these even work? In particular, do I need to deal with compiler optimizations on x86? And is using RCU semantics sufficient to deal with interrupting the writing context? - Which of these make sense from the POV of someone whose been in the linux kernel for a while? As a generic kernel person who's spent a lot of time in industry, my instincts are to avoid roll-your-own unless no existing primitive can do what I want. I also have a particular dislike of home grown concurrency control; even if the original author gets it right, it's tricky enough that some maintainer is almost certain to break it. And as for my team, one guy favours the rculist (#1), but isn't sure it's adequate; one guy wants to use explicit barriers (#3); and I've optimistically implemented #2, which I know would work on FreeBSD, and which could fairly easily be converted to using rculists. None of us want #4. Thanks for any insight, -- Arlie

3 5