- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

Asking about how to edit wiki?
by Shawn Aug. 2, 2011

Aug. 2, 2011

hi guys, I want to put an article onto website's wiki. I followed the steps, registered a user-name "Shawn the R0ck" on kernelnewbie website at first. Then sending a msg to mailinglist or irc? I want to know how to get the "trust-list" to edit wiki? Thanks! -- GNU powered it... GPL protect it... God blessing it... regards Shawn

3 3

Asynchronous read
by Da Zheng Aug. 2, 2011

Aug. 2, 2011

Hello, I'm trying to understand the read operation in VFS, and get confused by the asynchronous and synchronous operations. At the beginning, do_sync_read() invokes aio_read, which is generic_file_aio_read for ext4. generic_file_aio_read should be asynchronous read. But what really confuses me is do_generic_file_read, which is called by generic_file_aio_read. It seems to me do_generic_file_read implements synchronous read as this is the only function I can find that copy data to the user space by invoking the actor callback function. If do_generic_file_read is synchronous, how can generic_file_aio_read be asynchronous? In do_generic_file_read, if the data to be read isn't in the cache, normally page_cache_sync_readahead should be called. As far as I understand, when page_cache_sync_readahead returns, the pages will be ready in the cache, but the corresponding data in the disk isn't necessarily copied to the pages yet (because it eventually only invokes submit_bio to submit the IO requests to the block layer), so PageUptodate of the requested page might still return false, and then do_generic_file_read tries to invoke readpage to read the page again instead of waiting. Since the disk is always very slow, doesn't it just waste CPU time? Or do I miss something? Best, Da

2 2

Bond, round robin mis-configuration.
by Eduard Sinelnikov Aug. 2, 2011

Aug. 2, 2011

Hi, In the kernel 2.6.39.3 ( /drivers/net/bond/bond_main.c). In the function ‘bond_xmit_roundrobin’ The code check if the bond is active via ‘bond_is_active_slave(slave)’ Function call. Which actually checks if the slave is backup or active What is the meaning of slave being backup in round robin mode? Correct me if I wrong but in round robin every slave should send a packet, regardless of being active or backup. Thank you, Eduard

3 7

One more question about interrupts
by manish honap Aug. 1, 2011

Aug. 1, 2011

Interrupt gates are set in linux as set_intr_gate(i, interrupt[i-FIRST_EXTERNAL_VECTOR]); ^ but the 'interrupt' is declared as extern void (*__initconst interrupt[NR_VECTORS-FIRST_EXTERNAL_VECTOR])(void); can anyone possibly tell when are these enteries filled with valid pointers ???

1 0

Re: Asynchronous read
by Rajat Sharma Aug. 1, 2011

Aug. 1, 2011

On Mon, Aug 1, 2011 at 8:15 AM, Adam Cozzette <acozzette(a)cs.hmc.edu> wrote: > On Sun, Jul 31, 2011 at 03:58:55PM -0700, Da Zheng wrote: >> Hello, >> >> I'm trying to understand the read operation in VFS, and get confused by the >> asynchronous and synchronous operations. >> >> At the beginning, do_sync_read() invokes aio_read, which is >> generic_file_aio_read for ext4. generic_file_aio_read should be asynchronous >> read. But what really confuses me is do_generic_file_read, which is called by >> generic_file_aio_read. It seems to me do_generic_file_read implements >> synchronous read as this is the only function I can find that copy data to the >> user space by invoking the actor callback function. If do_generic_file_read is >> synchronous, how can generic_file_aio_read be asynchronous? >> >> In do_generic_file_read, if the data to be read isn't in the cache, normally >> page_cache_sync_readahead should be called. As far as I understand, when >> page_cache_sync_readahead returns, the pages will be ready in the cache, but the >> corresponding data in the disk isn't necessarily copied to the pages yet >> (because it eventually only invokes submit_bio to submit the IO requests to the >> block layer), so PageUptodate of the requested page might still return false, >> and then do_generic_file_read tries to invoke readpage to read the page again >> instead of waiting. Since the disk is always very slow, doesn't it just waste >> CPU time? Or do I miss something? > > This is a bit puzzling. I haven't figured it out but here are some things I came > across as I was trying to solve the problem. > > First of all, this article might shine some light on the problem: > > http://lwn.net/Articles/170954/ > > Essentially, a few years ago there was a simplification of the API and aio_read > and aio_write gained the ability to do vectored operations, making it possible > to eliminate readv and writev. This even made it possible for drivers and > filesystems to avoid implementing read() and write(), since the aio versions > could take care of that. > > So my point is that I suspect that aio_read and aio_write are now often used in > cases where they're not actually expected to be asynchronous, just because it > simplifies the API to be able to reuse those functions for synchronous > operations. In fact the LWN article says: > > Note that this change does not imply that asynchronous operations themselves > must be supported - it is entirely permissible (if suboptimal) for > aio_read() and aio_write() to operate synchronously at all times. > > So perhaps generic_file_aio_read is not actually asynchronous? My only other > guess is that whatever it does happens fast enough to count as asynchronous. > > -- > Adam Cozzette > Harvey Mudd College Class of 2012 > > _______________________________________________ > Kernelnewbies mailing list > Kernelnewbies(a)kernelnewbies.org > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > Linux libaio behave asynchronous only if file is opened with O_DIRECT mode. -Rajat

1 0

emneded ftp server which support IPV6
by Gang Lei (雷刚) Aug. 1, 2011

Aug. 1, 2011

Hello all: Could you please list any embedded ftp server which can support IPV6 nowadays? The image size should be less than 1M and it can run in standalone mode, I need such a server running on my arm926ej-s board. Thanks for your help in advance! Rui

2 7

Fw: Question about interrupts
by manish honap July 30, 2011

July 30, 2011

Well 2.6.39 seems to have changed and the flow as I understand for interrupt processing is do_IRQ -> irq_enter -> handle_irq -> ... but in above flow i cant find local_irq_enable anywhere; is there some different trick used to enable the interrupts ? ----- Forwarded Message ---- From: santhosh kumars <karuna.linux(a)gmail.com> To: manish honap <manish_honap_vit(a)yahoo.co.in> Cc: kernelnewbies(a)kernelnewbies.org Sent: Fri, 29 July, 2011 3:31:10 PM Subject: Re: Question about interrupts On Fri, Jul 29, 2011 at 3:16 PM, manish honap <manish_honap_vit(a)yahoo.co.in> wrote: > Hello all, > > When interrupts are handled through interrupt gate cpu automatically clears IF > flag. > > but to allow nesting kernel sets the IF flag in handle_irq_event - for 2.6.11 > > but in 2.6.39 how the interrupts are re-enabled so as to facilitate nesting of > interrupts > I am unable to find local_irq_enable in do_IRQ() I think in handle_IRQ_event event again interrupts are enabled if (!(action->flags & IRQF_DISABLED)) local_irq_enable_in_hardirq(); > > Thanks in advance. > > > _______________________________________________ > Kernelnewbies mailing list > Kernelnewbies(a)kernelnewbies.org > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >

3 2

How to get cpu utilization through code with least overhead?
by Vaibhav Jain July 30, 2011

July 30, 2011

Hi, I need to write code to get and record the cpu utilization in certain interval with the least overhead. Could anybody please advise me on how to proceed ? I tried looking at the code for "top" command to see how it works but couldn't figure out much. I don't want to write script that uses commands and pulls out the results as I want the overhead to be as small as possible. Thanks Vaibhav Jain

4 6

No printk to inform how much memory freed in ARM after booting
by anish singh July 30, 2011

July 30, 2011

Guess, this is trivial question but probably useful to know. void free_initmem(void) { #ifdef CONFIG_HAVE_TCM extern char __tcm_start, __tcm_end; totalram_pages += free_area(__phys_to_pfn(__pa(&__tcm_start)), __phys_to_pfn(__pa(&__tcm_end)), "TCM link"); #endif #ifndef CONFIG_HOTPLUG_CPU if (!machine_is_integrator() && !machine_is_cintegrator()) totalram_pages += free_area(__phys_to_pfn(__pa(__init_begin)), __phys_to_pfn(__pa(__init_end)), "init"); #endif } Here i don't see any printk information for specifying how much memory is freed. Other architectures do have this printk e.g kernel/arch/ia64/mm/init.c +169 While checking this excellent answer[1] given by rene i found that arm is missing this or may be i am missing something probably this information i can find it somewhere else also. Intention was to add __devinit in my probe function and out of curiosity how much i am saving. [1]http://www.spinics.net/lists/newbies/msg29197.html

1 0

Regarding connector netlink running out of buffer space.
by mindentropy July 29, 2011

July 29, 2011

Hi, I am trying to benchmark connector netlink and am passing a file and sending the same to listening userspace app's using connectors i.e. cat filename > /dev/myconnector and an app listening. When I do this I always run into ENOBUFS or "No buffer space available" for files greater than 2MB. Now how should I fix this? a) Increase the socket buffers? My rmem_max is 131071 and rmem_default is 126976 b) Should I check for the seq and ack and resend the packets? If yes how much packets should I not discard to do a resend? Are there any other ways? Also I am not sure why local loop would cause enobufs? I am running this on a quad core with 3G of RAM. Thanks.

2 2