Re: 6.0-current panic: loading radeon module

From: Warner Losh (imp_at_bsdimp.com)
Date: 06/17/05

  • Next message: Brooks Davis: "Re: HEADSUP: OpenBSD dhclient incoming" 
    Date: Fri, 17 Jun 2005 11:22:03 -0600 (MDT)
To: cracauer@cons.org

    From: Martin Cracauer <cracauer@cons.org>
    Subject: Re: 6.0-current panic: loading radeon module
    Date: Fri, 17 Jun 2005 13:15:08 -0400

    > Warner Losh wrote on Fri, Jun 17, 2005 at 11:08:47AM -0600:
    > > From: Martin Cracauer <cracauer@cons.org>
    > > Subject: Re: 6.0-current panic: loading radeon module
    > > Date: Fri, 17 Jun 2005 11:46:07 -0400
    > >
    > > > M. Warner Losh wrote on Thu, Jun 16, 2005 at 09:05:29PM -0600:
    > > > > In message: <20050616182421.A87199@cons.org>
    > > > > Martin Cracauer <cracauer@cons.org> writes:
    > > > > : Loading the radeon module segfaults the kernel, 6.0-current of
    > > > > : yesterday.
    > > > > :
    > > > > : I have a Thinkpad with Radeon mobility 7500. 5-stable works on the
    > > > > : same hardware.
    > > > > :
    > > > > : ddb trace shows:
    > > > > :
    > > > > : bus_generic_activate_resource+0x64
    > > > > : bus_generic_activate_resource+0x64
    > > > > : bus_generic_activate_resource+0x64
    > > > > : pci_alloc_resource+0x226
    > > > > : pci_alloc_resource+0x7x
    > > > > : drm_get_resource_len+0x2f
    > > > > : radeon_preinit+0xa2
    > > > > : drm_attach+0x2d
    > > > > :
    > > > > : Not sure what is involved to use gdb at this time. Can I do serial
    > > > > : gdb with a 6.0-current AMD64 machine running the display? Can I use a
    > > > > : 5.x box?
    > > > > :
    > > > > : It is 100% reproducable so I can gather more info as needed.
    > > > >
    > > > > Is this a module you compiled, or a module that's vendor supplied?
    > > >
    > > > It is the FreeBSD module built with the kernel. I used the same
    > > > procedure as with 5-stable where it worked fine (as in 3D graphics
    > > > actually work, not just it didn't panic).
    > >
    > > OK. There were some changes to the number of pages we reserve for uma
    > > allocation of malloc types. Can you try those?
    >
    > Yes, the box is ready for scratch testing. (Famous last words :-)
    >
    > Where do I find the changes?

    It was committed yesterday, I believe.

    Warner
    _______________________________________________
    freebsd-current@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-current
    To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"

  • Next message: Brooks Davis: "Re: HEADSUP: OpenBSD dhclient incoming"

    Relevant Pages

    • Re: [PATCH 00/10] foundations for reserve-based allocation
      ... page reserve limit, and then need to provide a reserve of pages. ... And we need to ensure that once we hit the reserve, the slab ... That is, a regular allocation ... This patch set is _way_ less intimidating than its predecessor. ...
      (Linux-Kernel)
    • Re: [patch 3/9] mempool - Make mempools NUMA aware
      ... I still think some sort of reserve pool ... under both memory pressure and network load. ... Any "real" reserve system will suffer from that problem. ... for trivially reclaimable allocation while not in active use. ...
      (Linux-Kernel)
    • Re: [PATCH 03/33] mm: slub: add knowledge of reserve pages
      ... contexts that are entitled to it. ... This is done to ensure reserve pages don't leak out and get consumed. ... I think this is generally a good idea (to prevent slab allocators ... what was the problem with failing the slab allocation ...
      (Linux-Kernel)
    • Re: Possible ways of dealing with OOM conditions.
      ... There is _one_ place for allocation of any kind of object. ... You do not have enough reserve to support whole traffic. ... Ah, but there you violate my requirement, any network allocation can ... claim the last bit of memory. ...
      (Linux-Kernel)
    • Re: [PATCH 2/5] cpuset memory spread page cache implementation and hooks
      ... NUMA too. ... This patch consolidates the UMA and NUMA memory allocation paths in the ... #define cache_alloc_debugcheck_after(a,b,objp,d) ...
      (Linux-Kernel)