Re: "seeing minute plus hangs during boot" - 2.6.12 and 2.6.13

From: Greg KH (greg_at_kroah.com)
Date: 07/31/05

  • Next message: Russell King: "S3 resume and PNP (was: Re: S3 resume and serial console..)"
    Date:	Sat, 30 Jul 2005 23:38:19 -0700
    To: "Francisco Figueiredo Jr." <fxjrlists@yahoo.com.br>
    
    

    On Sun, Jul 31, 2005 at 03:46:10AM -0300, Francisco Figueiredo Jr. wrote:
    > Andrew Morton wrote:
    > > "Francisco Figueiredo Jr." <fxjrlists@yahoo.com.br> wrote:
    > >
    > >>-----BEGIN PGP SIGNED MESSAGE-----
    > >>Hash: SHA1
    > >>
    > >>Andrew Morton wrote:
    > >>
    > >>>"Francisco Figueiredo Jr." <fxjrlists@yahoo.com.br> wrote:
    > >>>
    > >>>
    > >>>>udev S 00000002 0 1312 1 1224 (NOTLB)
    > >>>>c1653f4c 00000082 c1653f3c 00000002 00000001 00000040 c1653f64 c1653f0c
    > >>>> c016611b bfec96a8 c1653f0c 00000040 00000000 00000361 000241ed
    > >>>>c13fb520
    > >>>> 00000001 00001a7e 98f9769f 00000002 c146e520 df5da020 df5da148
    > >>>>c13fbf60
    > >>>>Call Trace:
    > >>>> [<c016611b>] cp_new_stat+0x15f/0x17a
    > >>>> [<c0352a74>] schedule_timeout+0x54/0xa2
    > >>>> [<c01274ce>] process_timeout+0x0/0x9
    > >>>> [<c01275c4>] sys_nanosleep+0xdd/0x18e
    > >>>> [<c0102e85>] syscall_call+0x7/0xb
    > >>>
    > >>>
    > >>>Well there's your delay: you've started running userspace and udev is
    > >>>running. Yes, it takes a long time.
    > >>>
    > >>>What makes you think this isn't normal behaviour? Do other kernels behave
    > >>>differently with the same userspace setup?
    > >>>
    > >>
    > >>
    > >>That's the point. On kernel 2.6.11 on same box I have no delay. It is
    > >>instantaneous. On 2.6.12-rc1 it was instantaneous but I didn't use it
    > >>much because I had drm problems. Later I tried 2.6.12 final and it was
    > >>hanging. I saw the "seeing a minute plugs hangs" on 2.6.13-rc1 release
    > >>notes and thought this could be the problem, but I compiled it and tried
    > >>with no luck :(
    > >>
    > >>Now, I'm thinking it could be something like the udev hang which
    > >>disapeared with udev update to 058.
    > >>
    > >>I don't know what can be happening. I think it is because of some type
    > >>of timeout.
    > >>
    > >>If you think there is something else I can do, please let me know.
    > >>
    > >
    > >
    > > Greg said in another thread: "older versions of udev (< 058) can work
    > > _really slow_ with 2.6.12. Please upgrade your version of udev and see if
    > > that solves the issue or not.".
    > >
    > > What version are you running? Looks like 058, yes?
    >
    >
    > Yeap. It is 058.

    Hm, odd. Well, as this is a udev issue, care to try 064? We've fixed a
    lot of little bugs since 058 that might have caused this.

    thanks,

    greg k-h
    -
    To 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/


  • Next message: Russell King: "S3 resume and PNP (was: Re: S3 resume and serial console..)"

    Relevant Pages

    • Re: 2.6.15-rc2-mm1
      ... > It's a userspace issue as to how udev is creating the initial device ... > nodes in Debian. ... > Odd that this only shows up in the -mm releases, ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: [PATCH] devfs: remove devfs from Kconfig preventing it from being built
      ... I'll take care of this by simply reinstalling ... >> and thus moving to udev. ... A nice addition to udev package would be an standard minimal /dev tree ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: udev and devfs - The final word
      ... udev and devfs are two different ways of providing this ... udev doesn't care squat about the major/minor that the ... a /dev entry with them, assigned to a name that it comes up with. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: Fw: signed kernel modules?
      ... On an Intel test board. ... I did a module signing patch some years ago. ... All I say is, take care. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: Q about pagecache data never written to disk
      ... The calling convention looks very very odd also; ... one of the results whenever there are multiple concurrent callers of it ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)