Paul Jackson wrote:
At the suggestion of Nick Piggin and Dinakar, totally disable
the facility to allow cpu_exclusive cpusets to define dynamic
sched domains in Linux 2.6.13, in order to avoid problems
first reported by John Hawkes (corrupt sched data structures
and kernel oops).
This has been b
Hi Tony,
Here's an even more tolerant version of palinfo.c.
I don't much like gotos if there's a reasonable alternative, so I just
used the else branches to avoid printing guff that wasn't valid.
Signed-off-by: Peter Chubb <[EMAIL PROTECTED]>
diff --git a/arch/ia64/kernel/palinfo.c b/arc
On Thu, 25 Aug 2005 13:15:33 -0500,
Smarduch Mario-CMS063 <[EMAIL PROTECTED]> wrote:
>
>We've been noticing some TLB issues on 2.4 and 2.6.10 kernels running
>on IA64 NUMA architecture. To this end I have some questions regarding
>TLB purging, please bare through with the initial introduction.
>
>
> "Tony" == Tony Luck writes:
>> - if (ia64_pal_mem_attrib(&attrib) != 0) - return 0; + if
>> (ia64_pal_mem_attrib(&attrib) != 0){ + *p++ = '\n'; + return
>> p-page; + }
Tony> Could we be even more tolerant with a series (gasp) goto
Tony> statements instead of the "return"? Like this:
Tony
>- if (ia64_pal_mem_attrib(&attrib) != 0)
>- return 0;
>+ if (ia64_pal_mem_attrib(&attrib) != 0){
>+ *p++ = '\n';
>+ return p-page;
>+ }
Could we be even more tolerant with a series (gasp) goto
statements instead of the "return"? Like this:
>+1:cmp.eq p6,p7=PAL_MEM_ATTRIB,r28
>+(p7) br.cond.sptk.few 1f
>+ mov r8=0
>+ mov r9=0x80|0x01/* WB, NatPage */
Applied ... but why so stingy with the Attrib bits?
According to /proc/pal/cpu0/vm_info a real cpu allows
all of WB, UC, UCE, WC, NaTPage.
-Tony
-
To
ia64_do_page_fault is a path critical for system performance. The code to call
notify_die() should not be compiled into that critical path if the system
is not configured to use KPROBES.
Signed-off-by: Christoph Lameter <[EMAIL PROTECTED]>
Index: linux-2.6.13-rc7/arch/ia64/mm/fault.c
At the suggestion of Nick Piggin and Dinakar, totally disable
the facility to allow cpu_exclusive cpusets to define dynamic
sched domains in Linux 2.6.13, in order to avoid problems
first reported by John Hawkes (corrupt sched data structures
and kernel oops).
This has been built for ppc64, i386,
The partial disabling of Dinakar's new facility to allow
cpu_exclusive cpusets to define dynamic sched domains
doesn't go far enough. At the suggestion of Nick Piggin
and Dinakar, let us instead totally disable this facility
for 2.6.13, in order to avoid problems first reported
by John Hawkes (cor
Altix patch to abstract irq_affinity down to the pci provider level since
different SGI hardware implements this in different ways.
Patch generated against:
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6.git#test
Signed-off-by: Mark Maule <[EMAIL PROTECTED]>
Index: affinity/ar
Dinakar wrote:
> I'll ack this for now until I fix the problems that I am seeing
> on ppc64
Thanks, Dinakar.
Linus - do *NOT* actually apply the literal patch that Dinakar ack'd.
1) It's logic is backwards - arrgh.
2) It doesn't undo the other attempt to partially disable this.
3) It's not a
Tony,
The ia64 allnoconfig build fails for some ACPI vs PCI reasons I've fixed
in my tree, but this small IA64 specific fix is needed also.
It stems from allnoconfig choosing CONFIG_PROC_FS=n
and sn_forc_interrupt_flag being defined within
CONFIG_PROC_FS, but used outside it.
cheers,
-Len
Index:
On Wed, Aug 24, 2005 at 01:31:07PM -0700, Paul Jackson wrote:
> ==
>
> The safest, mind numbingly simple thing to do that would avoid the oops
> that Hawkes reported is to simply not have the cpuset code call the
> code to setup a dynamic sched domain. This is choice (2) above, and
> coul
On Thu, Aug 18, 2005 at 08:14:15AM -0500, Mark Maule wrote:
>
> I'm no git/cogito expert, but the following worked for me to get a test
> branch snapshot:
>
> cg-clone
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6.git#test test
>
> This gets the correct source, but the bra
On Wed, Aug 24, 2005 at 02:09:55PM -0600, Bjorn Helgaas wrote:
> On Wednesday 24 August 2005 10:53 am, Kumar Gala wrote:
> > Removed IA64 architecture specific users of asm/segment.h and
> > asm-ia64/segment.h itself
>
> I posted a similar patch a month ago, but I only removed the
> arch/ia64 inc
15 matches
Mail list logo