Re: [patch] speed up / fix the new generic semaphore code (fix AIM7 40% regression with 2.6.26-rc1)

view
thread

!MAILaRCHIVE_VOTE_RePLACE

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

To: Alan Cox <alan@...>

Cc: Ingo Molnar <mingo@...>, Zhang, Yanmin <yanmin_zhang@...>, Andi Kleen <andi@...>, Matthew Wilcox <matthew@...>, LKML <linux-kernel@...>, Alexander Viro <viro@...>, Andrew Morton <akpm@...>, Thomas Gleixner <tglx@...>, H. Peter Anvin <hpa@...>

Subject: Re: [patch] speed up / fix the new generic semaphore code (fix AIM7 40% regression with 2.6.26-rc1)

Date: Thursday, May 8, 2008 - 7:33 pm

On Fri, 9 May 2008, Alan Cox wrote:
quoted text> 
> I don't think fixing n_tty is now a big job if someone wants to take a
> swing at it. The driver write/throttle/etc routines below the n_tty ldisc
> layer are now BKL clean so it should just be the internal locking of the
> buffers, window and the like to tackle.

Well, it turns out that Ingo's fixed statistics actually put the real cost 
in fcntl/ioctl/open/release:

    310  down <= lock_kernel <= sys_fcntl <= system_call_after_swapgs <
    332  down <= lock_kernel <= vfs_ioctl <= do_vfs_ioctl <
    380  down <= lock_kernel <= tty_release <= __fput <
    422  down <= lock_kernel <= chrdev_open <= __dentry_open <

rather than the write routines. But it may be that Ingo was just profiling 
two different sections, and it's really all of them.

			Linus
--
unsubscribe noticeTo unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:

AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Tue May 6, 1:48 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Tue May 6, 7:44 am)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Tue May 6, 10:11 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Tue May 6, 11:41 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Andi Kleen, (Wed May 7, 7:00 am)

Re: AIM7 40% regression with 2.6.26-rc1, Alan Cox, (Wed May 7, 9:59 am)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Wed May 7, 7:46 am)

Re: AIM7 40% regression with 2.6.26-rc1, Andi Kleen, (Wed May 7, 8:21 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 10:36 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 12:20 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 12:35 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 1:05 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 1:24 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 1:36 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 1:55 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Wed May 7, 1:59 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 2:17 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 2:49 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 11:19 am)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Wed May 7, 10:44 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Thu May 8, 2:43 am)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Thu May 8, 3:14 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Thu May 8, 3:39 am)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Thu May 8, 4:44 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Thu May 8, 5:21 am)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Thu May 8, 5:30 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Thu May 8, 5:29 am)

Re: AIM7 40% regression with 2.6.26-rc1, Andrew Morton, (Thu May 8, 2:48 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 11:29 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Thu May 8, 12:08 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Thu May 8, 12:17 am)

[patch] speed up / fix the new generic semaphore code (fix A..., Ingo Molnar, (Thu May 8, 8:01 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Arjan van de Ven, (Thu May 8, 9:56 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Matthew Wilcox, (Thu May 8, 9:20 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Thu May 8, 11:01 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Thu May 8, 8:28 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 12:02 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 2:30 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Thu May 8, 4:19 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 4:27 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Thu May 8, 5:45 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 6:55 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Alan Cox, (Thu May 8, 7:16 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 7:33 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Andi Kleen, (Fri May 9, 4:29 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Fri May 9, 2:50 am)

Re: [patch] speed up / fix the new generic semaphore code (f..., Alan Cox, (Thu May 8, 7:27 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 7:07 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Linus Torvalds, (Thu May 8, 7:14 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Thu May 8, 6:02 pm)

Re: [patch] speed up / fix the new generic semaphore code (f..., Ingo Molnar, (Thu May 8, 10:43 am)

[git pull] scheduler fixes, Ingo Molnar, (Thu May 8, 11:10 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 7:03 am)

Re: [git pull] scheduler fixes, Sven Wegener, (Sun May 11, 10:10 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 9:01 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 9:06 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 9:45 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 7:48 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 8:50 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 8:52 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 9:02 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 9:54 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 10:22 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 10:32 am)

Re: [git pull] scheduler fixes, Linus Torvalds, (Sun May 11, 12:47 pm)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 10:46 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 9:26 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 10:00 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 10:18 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 10:42 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 10:48 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Sun May 11, 11:19 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 11:29 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Tue May 13, 10:11 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Tue May 13, 10:21 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Tue May 13, 10:42 am)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Tue May 13, 11:28 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Tue May 13, 1:13 pm)

Re: [git pull] scheduler fixes, Linus Torvalds, (Tue May 13, 1:22 pm)

Re: [git pull] scheduler fixes, Ingo Molnar, (Tue May 13, 5:05 pm)

Re: [git pull] scheduler fixes, Matthew Wilcox, (Sun May 11, 7:14 am)

Re: [git pull] scheduler fixes, Adrian Bunk, (Thu May 8, 11:33 am)

Re: [git pull] scheduler fixes, Ingo Molnar, (Thu May 8, 11:41 am)

Re: [git pull] scheduler fixes, Adrian Bunk, (Thu May 8, 3:42 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 1:14 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Andi Kleen, (Wed May 7, 10:57 am)

Re: AIM7 40% regression with 2.6.26-rc1, Andrew Morton, (Wed May 7, 11:31 am)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Wed May 7, 12:22 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Alan Cox, (Wed May 7, 10:35 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 11:00 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 11:02 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 2:26 am)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 2:28 am)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Wed May 7, 3:05 am)

Re: AIM7 40% regression with 2.6.26-rc1, Andrew Morton, (Tue May 6, 11:59 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Wed May 7, 12:46 am)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Tue May 6, 8:09 am)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Tue May 6, 12:23 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Wed May 7, 11:24 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 11:34 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Thu May 8, 12:37 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Thu May 8, 10:58 am)

Re: AIM7 40% regression with 2.6.26-rc1, Andrew Morton, (Tue May 6, 1:21 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Wed May 7, 12:38 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 12:55 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 1:08 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 1:22 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 1:31 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 1:49 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 2:02 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 2:17 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 2:27 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 2:43 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 3:01 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Wed May 7, 3:24 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 3:44 pm)

Oi. NFS people. Read this., Matthew Wilcox, (Wed May 7, 4:00 pm)

Re: Oi. NFS people. Read this., Trond Myklebust, (Wed May 7, 6:10 pm)

Re: Oi. NFS people. Read this., J. Bruce Fields, (Thu May 8, 9:43 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 3:09 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 1:47 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Wed May 7, 1:25 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Andrew Morton, (Wed May 7, 1:16 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Wed May 7, 1:27 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Tue May 6, 1:45 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Tue May 6, 1:39 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Zhang, Yanmin, (Wed May 7, 2:49 am)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Tue May 6, 1:31 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Ingo Molnar, (Tue May 6, 1:49 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Andrew Morton, (Tue May 6, 2:07 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Sun May 11, 7:11 am)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Tue May 6, 12:36 pm)

Re: AIM7 40% regression with 2.6.26-rc1, J. Bruce Fields, (Tue May 6, 12:44 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Tue May 6, 12:42 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Tue May 6, 4:28 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Alan Cox, (Tue May 6, 12:39 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Tue May 6, 12:51 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Linus Torvalds, (Tue May 6, 1:42 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Alan Cox, (Tue May 6, 12:45 pm)

Re: AIM7 40% regression with 2.6.26-rc1, Matthew Wilcox, (Tue May 6, 7:18 am)

Navigation

Popular discussions


linux-kernel:
Greg KH	[GIT PATCH] driver core patches against 2.6.24
Joerg Roedel	[PATCH 03/34] AMD IOMMU: add defines and structures for ACPI scanning code
Ingo Molnar	[patch 03/13] syslets: generic kernel bits
Greg KH	Re: [malware-list] [RFC 0/5] [TALPA] Intro to a linux interface for on access scan...
git:

linux-netdev:

Gerrit Renker	[PATCH 27/37] dccp: Integration of dynamic feature activation - part 2 (server side)
David Miller	[GIT]: Networking
Jarek Poplawski	[PATCH] pkt_sched: Destroy gen estimators under rtnl_lock().
openbsd-misc:

Colocation donated by:

Online users

Syndicate