- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

who calls tx callback of the networking device
by devendra.aaru April 15, 2013

April 15, 2013

I have understood that a packet is transformed into skb in kernel and passed through the different layers of the stack and then given to Qdisc. from here will the pkt scheduler calls the device tx callback?

1 0

Forum for asking questions related to block device drivers
by neha naik April 15, 2013

April 15, 2013

Hi All, Nobody has replied to my query here. So i am just wondering if there is a forum for block device driver where i can post my query. Please tell me if there is any such forum. Thanks, Neha ---------- Forwarded message ---------- From: neha naik <nehanaik27(a)gmail.com> Date: Tue, Apr 9, 2013 at 10:18 AM Subject: Passthrough device driver performance is low on reads compared to writes To: kernelnewbies(a)kernelnewbies.org 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

7 12

How will Linux community support the coming Intel new chip Bay Trail?
by Peter Xu April 11, 2013

April 11, 2013

Hi, all, It seems that Intel will publish a nice chip called Bay Trail (or plus, I don't quick sure, which is for smartphones/tablets, also some lower ends of laptops in the future). It was said publically that Intel will support Linux platform on that chip. I just want to know something from the communitiy side, that is there a plan or something related to that chip? I think the most important part should be how to make sure Linux will have a good power managerment on the Bay Trail machines? Thanks in advance. Yours, Peter

3 2

Re: Linux from scratch
by Mulyadi Santosa April 11, 2013

April 11, 2013

On Wed, Apr 10, 2013 at 8:01 PM, Pietro Paolini <pulsarpietro(a)aol.com> wrote: > Did you achieve in compile and run the 1.0 Linux kernel on your system ? I am really curios ! > Best Regards > Pietro . > > > > I haven't tested that by myself. IIRC I only tried to run the provided disk image inside Qemu, but that's all. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com

2 1

Kexec sys call code
by leo kirotawa April 10, 2013

April 10, 2013

Hi, I'm working in a project and there I need use kexec code to do somethings. I read a lot of the kexec-tools code (user-land) and kexec-sys call (k-land) and I'd like to be sure about some arguments there. In kexec we have that kexec_image and into it that nr_segments and the segments itself. I know that kernel image passed by kexec in user land (bzImage) is copy in a buffer and in kexe_load it receives the entry point for this new kernel (bzImage) and the start position where (I'm supposing ) start the kernel image. So , my question is, using that nr_segments or segemtns by itself can I determine the size of the bzImage/image in kexec-sys call? And these guesses about entry point and the address of start kernel image, is it correct? This is the kexec.c file (where we have the sys call and a bunch of another code). http://lxr.free-electrons.com/source/kernel/kexec.c -- ---------------------------------------------- Leônidas S. Barbosa (Kirotawa)

1 0

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