AGP and PAT (induced?) problem (on AMD family 6) - Kernel

This is a discussion on AGP and PAT (induced?) problem (on AMD family 6) - Kernel ; On 11-09-08 10:17, Ingo Molnar wrote: > * Rene Herman wrote: > >> On 22-08-08 22:02, Rene Herman wrote: >> >>> On 22-08-08 06:15, Ingo Molnar wrote: >>>> The frontside cache itself is in x86/urgent: >>>> >>>> # 80c5e73: x86: ...

+ Reply to Thread
Page 3 of 3 FirstFirst 1 2 3
Results 41 to 46 of 46

Thread: AGP and PAT (induced?) problem (on AMD family 6)

  1. Re: AGP PAT issue.

    On 11-09-08 10:17, Ingo Molnar wrote:

    > * Rene Herman wrote:
    >
    >> On 22-08-08 22:02, Rene Herman wrote:
    >>
    >>> On 22-08-08 06:15, Ingo Molnar wrote:
    >>>> The frontside cache itself is in x86/urgent:
    >>>>
    >>>> # 80c5e73: x86: fix Xorg startup/shutdown slowdown with PAT
    >>>>
    >>>> ... and should at least solve the symptom that you've hit in practice
    >>>> (the slowdown), without changing the underlying PAT machinery. (which
    >>>> would be way too dangerous for v2.6.27)
    >>> Well, please note that that specific commit only fixes X startup -- it
    >>> doesn't do anything for shutdown. With only that one, I'm still at 14
    >>> seconds for X shutdown (first time after boot that is, 5 seconds
    >>> subsequent shutdowns) versus 1 (or sub 1, feels immediate) normally.
    >>>
    >>> It's also a black-screen "hang", so we'll probably be getting a lot of
    >>> "long hang at shutdown" reports without something additionally for .27.
    >>>
    >>> Venki?

    >> Haven't been subcribed to any lists recently and someone was talking
    >> about "the other thread" before but just noticed that an -rc6 was cut.
    >>
    >> Please note that the shutdown issue remains unfixed in it (I'm doing
    >> my coalescing changes locally).

    >
    > here's what is pending in tip/x86/pat for v2.6.28:


    Only talking about .27 and just making sure again the issue is known.

    The above mentioned subject for the entry cache one ("fix Xorg
    startup/shutdown slowdown with PAT") after all mistakingly says it does
    something for shutdown.

    Rene.
    --
    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/

  2. RE: AGP PAT issue.



    >-----Original Message-----
    >From: Rene Herman [mailto:rene.herman@keyaccess.nl]
    >Sent: Thursday, September 11, 2008 1:31 AM
    >To: Ingo Molnar
    >Cc: Pallipadi, Venkatesh; Dave Airlie; Li, Shaohua; Yinghai
    >Lu; Andreas Herrmann; Arjan van de Ven; Linux Kernel; Siddha,
    >Suresh B; Thomas Gleixner; H. Peter Anvin; Dave Jones
    >Subject: Re: AGP PAT issue.
    >
    >On 11-09-08 10:17, Ingo Molnar wrote:
    >
    >> * Rene Herman wrote:
    >>
    >>> On 22-08-08 22:02, Rene Herman wrote:
    >>>
    >>>> On 22-08-08 06:15, Ingo Molnar wrote:
    >>>>> The frontside cache itself is in x86/urgent:
    >>>>>
    >>>>> # 80c5e73: x86: fix Xorg startup/shutdown slowdown with PAT
    >>>>>
    >>>>> ... and should at least solve the symptom that you've hit

    >in practice
    >>>>> (the slowdown), without changing the underlying PAT

    >machinery. (which
    >>>>> would be way too dangerous for v2.6.27)
    >>>> Well, please note that that specific commit only fixes X

    >startup -- it
    >>>> doesn't do anything for shutdown. With only that one, I'm

    >still at 14
    >>>> seconds for X shutdown (first time after boot that is, 5 seconds
    >>>> subsequent shutdowns) versus 1 (or sub 1, feels immediate)

    >normally.
    >>>>
    >>>> It's also a black-screen "hang", so we'll probably be

    >getting a lot of
    >>>> "long hang at shutdown" reports without something

    >additionally for .27.
    >>>>
    >>>> Venki?
    >>> Haven't been subcribed to any lists recently and someone was talking
    >>> about "the other thread" before but just noticed that an

    >-rc6 was cut.
    >>>
    >>> Please note that the shutdown issue remains unfixed in it (I'm doing
    >>> my coalescing changes locally).

    >>
    >> here's what is pending in tip/x86/pat for v2.6.28:

    >
    >Only talking about .27 and just making sure again the issue is known.
    >
    >The above mentioned subject for the entry cache one ("fix Xorg
    >startup/shutdown slowdown with PAT") after all mistakingly says it does
    >something for shutdown.
    >


    Can you try the patch here
    http://www.ussg.iu.edu/hypermail/lin...09.1/2074.html

    That should resolve both reserve and free issues..

    Thanks,
    Venki
    --
    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/

  3. Re: AGP PAT issue.

    On 13-09-08 02:26, Pallipadi, Venkatesh wrote:

    >>>> Haven't been subcribed to any lists recently and someone was
    >>>> talking about "the other thread" before but just noticed that
    >>>> an-rc6 was cut.
    >>>>
    >>>> Please note that the shutdown issue remains unfixed in it (I'm
    >>>> doing my coalescing changes locally).
    >>>>
    >>> here's what is pending in tip/x86/pat for v2.6.28:
    >>>

    >> Only talking about .27 and just making sure again the issue is
    >> known.
    >>
    >> The above mentioned subject for the entry cache one ("fix Xorg
    >> startup/shutdown slowdown with PAT") after all mistakingly says it
    >> does something for shutdown.
    >>

    > Can you try the patch here
    > http://www.ussg.iu.edu/hypermail/lin...09.1/2074.html
    >
    > That should resolve both reserve and free issues..


    Not for .27 though.

    Rene.
    --
    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/

  4. Re: AGP PAT issue.

    Rene Herman wrote:
    > On 13-09-08 02:26, Pallipadi, Venkatesh wrote:
    >
    >>>>> Haven't been subcribed to any lists recently and someone was
    >>>>> talking about "the other thread" before but just noticed that
    >>>>> an-rc6 was cut.
    >>>>>
    >>>>> Please note that the shutdown issue remains unfixed in it (I'm
    >>>>> doing my coalescing changes locally).
    >>>>>
    >>>> here's what is pending in tip/x86/pat for v2.6.28:
    >>>>
    >>> Only talking about .27 and just making sure again the issue is known.
    >>>
    >>> The above mentioned subject for the entry cache one ("fix Xorg
    >>> startup/shutdown slowdown with PAT") after all mistakingly says it
    >>> does something for shutdown.
    >>>

    >> Can you try the patch here
    >> http://www.ussg.iu.edu/hypermail/lin...09.1/2074.html
    >>
    >> That should resolve both reserve and free issues..

    >
    > Not for .27 though.
    >
    > Rene.


    Hi!
    What's the status on this?

    There are graphics devices we're working with that require large
    (non-AGP) write-combined memory buffers at any time.
    We can solve the tlb- and cache flush latencies by using pools of these
    pages to allocate from and free to, but sooner or later we'll probably
    end up with _huge_ memtype lists. In the (very unlikely) worst case I
    guess they'd contain every other lowmem page in they system.

    If I understand things correctly the patch above will fix this issue?
    Will it be considered for future kernel inclusion.

    Any enlightenment on this subject would be greatly appreciated.
    Thanks,
    /Thomas



    --
    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/

  5. RE: AGP PAT issue.



    >-----Original Message-----
    >From: Thomas Hellstrom [mailto:thomas@tungstengraphics.com]
    >Sent: Thursday, October 09, 2008 8:53 AM
    >To: Rene Herman
    >Cc: Pallipadi, Venkatesh; Ingo Molnar; Dave Airlie; Li,
    >Shaohua; Yinghai Lu; Andreas Herrmann; Arjan van de Ven; Linux
    >Kernel; Siddha, Suresh B; Thomas Gleixner; H. Peter Anvin;
    >Dave Jones; Alan Hourihane
    >Subject: Re: AGP PAT issue.
    >
    >Rene Herman wrote:
    >> On 13-09-08 02:26, Pallipadi, Venkatesh wrote:
    >>
    >>>>>> Haven't been subcribed to any lists recently and someone was
    >>>>>> talking about "the other thread" before but just noticed that
    >>>>>> an-rc6 was cut.
    >>>>>>
    >>>>>> Please note that the shutdown issue remains unfixed in it (I'm
    >>>>>> doing my coalescing changes locally).
    >>>>>>
    >>>>> here's what is pending in tip/x86/pat for v2.6.28:
    >>>>>
    >>>> Only talking about .27 and just making sure again the

    >issue is known.
    >>>>
    >>>> The above mentioned subject for the entry cache one ("fix Xorg
    >>>> startup/shutdown slowdown with PAT") after all mistakingly says it
    >>>> does something for shutdown.
    >>>>
    >>> Can you try the patch here
    >>> http://www.ussg.iu.edu/hypermail/lin...09.1/2074.html
    >>>
    >>> That should resolve both reserve and free issues..

    >>
    >> Not for .27 though.
    >>
    >> Rene.

    >
    >Hi!
    >What's the status on this?


    This patch is now in upstream git and well on its way to .28

    >There are graphics devices we're working with that require large
    >(non-AGP) write-combined memory buffers at any time.
    >We can solve the tlb- and cache flush latencies by using pools of these
    >pages to allocate from and free to, but sooner or later we'll probably
    >end up with _huge_ memtype lists. In the (very unlikely) worst case I
    >guess they'd contain every other lowmem page in they system.
    >
    >If I understand things correctly the patch above will fix this issue?
    >Will it be considered for future kernel inclusion.


    Yes this patch will help the above case.

    Thanks,
    Venki
    --
    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/

  6. Re: AGP PAT issue.

    Pallipadi, Venkatesh wrote:
    >
    >> -----Original Message-----
    >> From: Thomas Hellstrom [mailto:thomas@tungstengraphics.com]
    >> Sent: Thursday, October 09, 2008 8:53 AM
    >> To: Rene Herman
    >> Cc: Pallipadi, Venkatesh; Ingo Molnar; Dave Airlie; Li,
    >> Shaohua; Yinghai Lu; Andreas Herrmann; Arjan van de Ven; Linux
    >> Kernel; Siddha, Suresh B; Thomas Gleixner; H. Peter Anvin;
    >> Dave Jones; Alan Hourihane
    >> Subject: Re: AGP PAT issue.
    >>
    >> Rene Herman wrote:
    >>
    >>> On 13-09-08 02:26, Pallipadi, Venkatesh wrote:
    >>>
    >>>
    >>>>>>> Haven't been subcribed to any lists recently and someone was
    >>>>>>> talking about "the other thread" before but just noticed that
    >>>>>>> an-rc6 was cut.
    >>>>>>>
    >>>>>>> Please note that the shutdown issue remains unfixed in it (I'm
    >>>>>>> doing my coalescing changes locally).
    >>>>>>>
    >>>>>>>
    >>>>>> here's what is pending in tip/x86/pat for v2.6.28:
    >>>>>>
    >>>>>>
    >>>>> Only talking about .27 and just making sure again the
    >>>>>

    >> issue is known.
    >>
    >>>>> The above mentioned subject for the entry cache one ("fix Xorg
    >>>>> startup/shutdown slowdown with PAT") after all mistakingly says it
    >>>>> does something for shutdown.
    >>>>>
    >>>>>
    >>>> Can you try the patch here
    >>>> http://www.ussg.iu.edu/hypermail/lin...09.1/2074.html
    >>>>
    >>>> That should resolve both reserve and free issues..
    >>>>
    >>> Not for .27 though.
    >>>
    >>> Rene.
    >>>

    >> Hi!
    >> What's the status on this?
    >>

    >
    > This patch is now in upstream git and well on its way to .28
    >
    >
    >> There are graphics devices we're working with that require large
    >> (non-AGP) write-combined memory buffers at any time.
    >> We can solve the tlb- and cache flush latencies by using pools of these
    >> pages to allocate from and free to, but sooner or later we'll probably
    >> end up with _huge_ memtype lists. In the (very unlikely) worst case I
    >> guess they'd contain every other lowmem page in they system.
    >>
    >> If I understand things correctly the patch above will fix this issue?
    >> Will it be considered for future kernel inclusion.
    >>

    >
    > Yes this patch will help the above case.
    >
    > Thanks,
    > Venki
    >

    Thanks, Venki.
    Much appreciated.

    /Thomas



    --
    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/

+ Reply to Thread
Page 3 of 3 FirstFirst 1 2 3