At the moment we size the hashtable based on 4kB pages / 2, even on a
64kB kernel. This results in a hashtable that is much larger than it
needs to be.
Grab the real page size and size the hashtable based on that. Note: this
only works on non hypervisor machines.
Signed-off-by: Anton Blanchard
balanced:
Free memory:
Node 0: 97.53%
Node 1: 98.35%
Node 2: 98.33%
Node 3: 98.33%
For many HPC applications we are limited by the free available memory on
the smallest node, so even though the same amount of memory is used the
better balancing helps.
Signed-off-by: Anton Blanchard
---
diff --git a
Hi Ben,
> You have numbers ? :-) I'm asking mostly because I've been wondering
> whether it offsets the 16M pages vs. 4K or 64K pages in term of TLB/ERAT
> impact.
The speedup is application dependent. Things like linpack usually
improve when you throw more memmory at them.
The potential slowdo
free available memory on
the smallest node, so even though the same amount of memory is used the
better balancing helps.
Since all 64bit NUMA capable architectures should have sufficient
vmalloc space, it makes sense to enable it via CONFIG_64BIT.
Signed-off-by: Anton Blanchard
---
diff --git a
1afa9000 00:00 0 [heap]
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/include/asm/elf.h
===
--- linux-2.6.orig/arch/powerpc/include/asm/elf.h 2009-02-20
16:06:32.000
Move is_32bit_task into asm/thread_info.h, that allows us to test for
32/64bit tasks without an ugly CONFIG_PPC64 ifdef.
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/include/asm/thread_info.h
===
--- linux-2.6
h the same ~128MB hole as in
32bit. This works because we randomise the stack over a 1GB range.
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/mm/mmap.c
===
--- linux-2.6.orig/arch/powerpc/mm/mmap.c 2009-02-21
Randomise ELF_ET_DYN_BASE, which is used when loading position independent
executables.
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/include/asm/elf.h
===
--- linux-2.6.orig/arch/powerpc/include/asm/elf.h
fffd88ac000-fffd88ad000 rw-p fffd88ac000 00:00 0
fffbc62e000-fffbc62f000 rw-p fffbc62e000 00:00 0
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/mm/mmap.c
===
--- linux-2.6.orig/arch/powerpc/mm/mmap.c 2009
Rearrange mmap.c to better match the x86 version.
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/mm/mmap.c
===
--- linux-2.6.orig/arch/powerpc/mm/mmap.c 2009-02-20 13:40:26.0
+1100
+++ linux-2.6
re:
80
80
80
80
80
after:
610
490
300
6b0
d80
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/include/asm/system.h
===
--- linux-2.6.orig/arch/powerpc/include/asm/system.h2009-02-20
13:39:05.0 +1100
The following set of patches adds randomisation of mmaps, heap and
position independent executables, and increases the randomisation applied to
the stack on 64bit binaries.
--
___
Linuxppc-dev mailing list
Linuxppc-dev@ozlabs.org
https://ozlabs.org/mail
: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/mm/mmap.c
===
--- linux-2.6.orig/arch/powerpc/mm/mmap.c 2009-02-22 11:58:54.0
+1100
+++ linux-2.6/arch/powerpc/mm/mmap.c2009-02-22 12:05:01.0 +1100
@@ -46,6
.9.so
7ffc000-7fff000 rw-p 00028000 08:06 130663 /lib64/ld-2.9.so
7fff000-800 rw-p 7fff000 00:00 0
fc59000-fc6e000 rw-p ffeb000 00:00 0 [stack]
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/mm/m
000 rw-p ffeb000 00:00 0 [stack]
fffc719d000-fffc71b2000 rw-p ffeb000 00:00 0 [stack]
fffe01af000-fffe01c4000 rw-p ffeb000 00:00 0 [stack]
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/include/a
enable it via CONFIG_64BIT.
Signed-off-by: Anton Blanchard
---
diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h
index 95837bf..0c4d4b7 100644
--- a/include/linux/bootmem.h
+++ b/include/linux/bootmem.h
@@ -144,10 +144,10 @@ extern void *alloc_large_system_hash(const char
*tablena
Enable relay in pseries config, ppc64_defconfig had it enabled but pseries
did not.
Signed-off-by: Anton Blanchard
---
diff --git a/arch/powerpc/configs/pseries_defconfig
b/arch/powerpc/configs/pseries_defconfig
index 61b1008..f1889ab 100644
--- a/arch/powerpc/configs/pseries_defconfig
+++ b
We specify a 64MB RMO, but the comment says 128MB.
Signed-off-by: Anton Blanchard
---
diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index 7f1b33d..ddd9561 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -614,7 +614,7
Instead of checking for known events, pass in all 1s so we handle future
event types. We were currently missing the IO event type.
Signed-off-by: Anton Blanchard
---
diff --git a/arch/powerpc/include/asm/rtas.h b/arch/powerpc/include/asm/rtas.h
index 0aa0315..01c1233 100644
--- a/arch/powerpc
PHYP tells us how often a shared processor dispatch changed physical cpus.
This can highlight performance problems caused by the hypervisor.
Signed-off-by: Anton Blanchard
---
Index: linux-2.6/arch/powerpc/include/asm/lppaca.h
Commit 57b150cce8e004ddd36330490a68bfb59b7271e9 (irq: only update affinity if
->set_affinity() is sucessfull) broke xics irq affinity.
We need to use the cpumask passed in, instead of accessing ->affinity directly.
Signed-off-by: Anton Blanchard
---
This code needs a rework and a conv
/:00:01.0/local_cpulist
64-65
Change cpumask_of_node to check for -1 and return cpu_all_mask in this
case:
# cat /sys/devices/pci:00/:00:01.0/local_cpus
,,,
# cat /sys/devices/pci:00/:00:01.0/local_cpulist
0-127
Signed-off-by: Anton Blanchard
commit ac4c2a3bbe5db5fc570b1d0ee1e474db7cb22585 (zlib: optimize inffast when
copying direct from output) referenced include/linux/autoconf.h which
is now called include/generated/autoconf.h.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/boot/Makefile
Using perf to trace L1 dcache misses and dumping data addresses I found a few
variables taking a lot of misses. Since they are almost never written, they
should go into the __read_mostly section.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/xmon/xmon.c
8767938 2944132 1636796 13348866 cbb002 vmlinux.baseline
8767779 2944260 1505724 13217763 c9afe3 vmlinux.irq_cpustat
A saving of around 128kB.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/include/asm/hardirq.h
NR_CPUS=1024, this saves quite a lot of memory:
textdata bss dec hexfilename
8767779 2944260 1505724 13217763 c9afe3 vmlinux.irq_cpustat
8767555 2813444 1505724 13086723 c7b003 vmlinux.xics
A saving of around 128kB.
Signed-off-by: Anton Blanchard
---
Index
The cputime code has a few places that do per_cpu(, smp_processor_id()).
Replace them with __get_cpu_var().
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/include/asm/cputime.h
===
--- linux-cpumask.orig/arch
Now we use printf style alignment there is no need to manually space
these fields.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/512x/mpc5121_ads_cpld.c
===
--- linux-cpumask.orig/arch/powerpc
. On PowerPC we currently omit an interrupt completely if there is no
active handler, whereas on x86 it is printed if there is a non zero count.
The x86 code also spaces the first column correctly based on nr_irqs.
Signed-off-by: Anton Blanchard
---
This patch follows on from "powerpc: R
spurious. Since they can be a significant cause
of OS jitter, we may as well log them per cpu so we know where they are
occurring.
Signed-off-by: Anton Blanchard
---
Again, the short name matches x86 but if people are sufficiently confused by
it ("SPU"), then we can think up something el
using the perf tool. Since it's possible to take a completely
handled machine check on a System p box it also sounds like a good idea to
keep a machine check summary.
The event naming matches x86 to keep gratuitous differences to a minimum.
Signed-off-by: Anton Blanchard
---
If people really
0 0 0 0
17158:319 0 0 0 0 0
25092: 0 0 0 0 0 0
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/Kc
Hi,
> > +static inline void ack_bad_irq(unsigned int irq)
> > +{
> > + printk(KERN_CRIT "unexpected IRQ trap at vector %02x\n", irq);
> > +}
>
> Looks like some unrelated gunk slipped into this patch :-)
We need that to link :)
kernel/irq/handle.c: In function ‘handle_bad_irq’:
kernel/irq/h
Hi,
> > +#if defined(CONFIG_XICS) && defined(CONFIG_SMP)
> > + unsigned long xics_ipi;
> > +#endif
> > } cacheline_aligned irq_cpustat_t;
>
> This is still a gross abuse of irq_cpustat_t ... Can't we do
> a separate DECLARE_PER_CPU_SHARED_ALIGNED(unsigned long, xics_ipi)
> inside xics.c i
8767938 2944132 1636796 13348866 cbb002 vmlinux.baseline
8767779 2944260 1505724 13217763 c9afe3 vmlinux.irq_cpustat
A saving of around 128kB.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/include/asm/hardirq.h
using the perf tool. Since it's possible to take a completely
handled machine check on a System p box it also sounds like a good idea to
keep a machine check summary.
The event naming matches x86 to keep gratuitous differences to a minimum.
Signed-off-by: Anton Blanchard
---
If people really
dec hexfilename
8767779 2944260 1505724 13217763 c9afe3 vmlinux.irq_cpustat
8767555 2813444 1505724 13086723 c7b003 vmlinux.xics
A saving of around 128kB.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/pseries/xics.c
spurious. Since they can be a significant cause
of OS jitter, we may as well log them per cpu so we know where they are
occurring.
Signed-off-by: Anton Blanchard
---
Again, the short name matches x86 but if people are sufficiently confused by
it ("SPU"), then we can think up something el
. On PowerPC we currently omit an interrupt completely if there is no
active handler, whereas on x86 it is printed if there is a non zero count.
The x86 code also spaces the first column correctly based on nr_irqs.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/irq.c
0 0 0 0
17158:319 0 0 0 0 0
25092: 0 0 0 0 0 0
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/Kc
Now we use printf style alignment there is no need to manually space
these fields.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/512x/mpc5121_ads_cpld.c
===
--- linux-cpumask.orig/arch/powerpc
Hi,
> On recent ppc64 kernels, limiting the stack (using 'ulimit -s blah') is
> now more restrictive than it was before. On 2.6.31 with 4k pages I
> could run 'ulimit -s 16; /usr/bin/test' without a problem. Now with
> mainline, even 'ulimit -s 64; /usr/bin/test' gets killed.
>
> Using 64k pag
-by: Anton Blanchard
---
v2: Enable NUMA on ppc64_defconfig
Index: powerpc.git/arch/powerpc/configs/ppc64_defconfig
===
--- powerpc.git.orig/arch/powerpc/configs/ppc64_defconfig 2010-02-05
14:57:48.889717208 +1100
We can use the much more lightweight ida allocator since we don't
need the pointer storage idr provides.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/mm/mmu_context_hash64.c
===
--- powerpc.git.orig
The clockevent multiplier and shift is useful information, but we
only need to print it once.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/kernel/time.c
===
--- powerpc.git.orig/arch/powerpc/kernel/time.c
We use firmware_has_feature quite a lot these days, so it's worth putting
powerpc_firmware_features into __read_mostly.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/kernel/firmware.c
===
--- powerpc.git
Clean up SD_NODE_INITS so we can easily compare it to x86. Similar to the
work in 47734f89be0614b5acbd6a532390f9c72f019648 (sched: Clean up topology.h)
Signed-off-by: Anton Blanchard
---
Index: linux.trees.git/arch/powerpc/include/asm/topology.h
The cede latency stuff is relatively new and we don't need to complain about
it not working on older firmware.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/platforms/pseries/hotplug-
o 64kB) so it operates better with > 4kB pages.
But in the end its probably of little use for the default OVERCOMMIT_GUESS
setting, so the main thing is we dont terminate processes incorrectly.
Acked-by: Anton Blanchard
Anton
___
Linuxppc-dev mai
) | .vprintk() {
0) 0.488 us|._raw_spin_lock();
0) 0.572 us|.emit_log_char();
Showing the function graph of a sysrq-c crash.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/kernel/traps.c
Hi,
> Why do we need page size independent stack size? It seems to have
> compatibility breaking risk.
I don't think so. The current behaviour is clearly wrong, we dont need a
16x larger stack just because you went from a 4kB to a 64kB base page
size. The user application stack usage is the sam
Hi Wolfram,
> What about printk_once from kernel.h?
Thanks for the suggestion!
Anton
--
The clockevent multiplier and shift is useful information, but we
only need to print it once.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/kernel/tim
Hi,
> I didn't discuss which behavior is better. Michael said he want to apply
> his patch to 2.6.32 & 2.6.33. stable tree never accept the breaking
> compatibility patch.
>
> Your answer doesn't explain why can't we wait it until next merge window.
>
>
> btw, personally, I like page size inde
RTAS should never cause an exception but if it does (for example accessing
outside our RMO) then we might go a long way through the kernel before
oopsing. If we unset MSR_RI we should at least stop things on exception
exit.
Signed-off-by: Anton Blanchard
---
The setting of MSR_RI during RTAS
> The setting of MSR_RI during RTAS calls went in ages ago (2003), and I'm
> not sure why. Clearly an exception inside RTAS is fatal since we most likely
> have a bad r13.
I wrote the r13 comment without thinking :) Regardless I think we want
to catch any RTAS exception ASAP.
Anton
___
| |--36.64%-- .futex_wait_setup
which is a good test for this patch. The results (in lock/unlock operations
per second) are:
before: 1538203 ops/sec
after: 2189219 ops/sec
An improvement of 42%
Signed-off-by: Anton Blanchard
---
We've had issues in the past with chips that do the wrong thi
futex_wait_setup
Which is a good test for this patch. The results (in lock/unlock operations per
second) are:
before: 1538203 ops/sec
after: 2189219 ops/sec
An improvement of 42%
A 32 core POWER7 improves even more:
before: 1279529 ops/sec
after: 2282076 ops/sec
An improvement of 78%
Sign
Now we have real bit locks use them instead of open coding it.
Signed-off-by: Anton Blanchard
---
diff --git a/arch/powerpc/mm/hash_native_64.c b/arch/powerpc/mm/hash_native_64.c
index 056d23a..9e1aa4f 100644
--- a/arch/powerpc/mm/hash_native_64.c
+++ b/arch/powerpc/mm/hash_native_64.c
This patch implements the lwarx/ldarx hint bit for bit locks.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/include/asm/asm-compat.h
===
--- powerpc.git.orig/arch/powerpc/include/asm/asm-compat.h 2010-02
do_lwsync_fixups doesn't work on 64bit, we end up writing lwsyncs to the
wrong addresses:
0:mon> di c001000bfacc
c001000bfacc 7c2004ac lwsync
Since the lwsync section has negative offsets we need to use a signed int
pointer so we sign extend the value.
Signed-off-b
For performance reasons we are about to change ISYNC_ON_SMP to sometimes be
lwsync. Now that the macro name doesn't make sense, change it and LWSYNC_ON_SMP
to better explain what the barriers are doing.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/include/asm/ato
t the instruction. We were marking POWER3 and RS64 as lwsync
capable but in reality they treat it as a full sync (ie slow). Remove the
CPU_FTR_LWSYNC bit from these CPUs so they continue to use the faster isync
method.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/include/asm/sy
read(fd, buf, BUFSIZE, 0) != BUFSIZE) {
perror("pread");
exit(1);
}
}
unlink(tmpfile);
return 0;
}
Signed-off-by: Anton Blanchard
---
Ind
(after.tv_sec - before.tv_sec) * NSEC_PER_SEC +
(after.tv_nsec - before.tv_nsec));
return 0;
}
static void __exit copypage_exit(void)
{
}
module_init(copypage_init)
module_exit(copypage_exit)
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Anton Blanchard");
test is the private futex wake system call and that improves 5%.
This is a decent speedup on an important system call for threaded applications.
Signed-off-by: Anton Blanchard
---
If my previous patches didn't worry you then this one is sure to.
Getting this wrong will make someone&#
Hi Ben,
> Well, the main issue here is leaking kernel reservations into userspace,
> and thus the question of whether it is a big deal or not. There's also
> an issue I can see with signals.
>
> The risk with kernel reservations leaking into userspace is a problem on
> some processors that do n
Hi Nick,
> Cool. How does it go when there are significant amount of instructions
> between the lock and the unlock? A real(ish) workload, like dbench on
> ramdisk (which should hit the dcache lock).
Good question, I'll see if we can see a difference on dbench.
Anton
__
Hi Nick,
> Ah, good to see this one come back. I also tested tbench over localhost
> btw which actually did show some speedup on the G5.
>
> BTW. this was the last thing left:
> http://www.mail-archive.com/linuxppc-dev@lists.ozlabs.org/msg29738.html
>
> Don't know if you took a look at that a
igure-kernel-dump call."
This patch checks for the ibm,extended-os-term property and calls ibm,os-term
if it exists.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/rtas.c
===
--- linux-cpumask.orig/ar
ode.
The patch below sets a smaller value for RECLAIM_DISTANCE and thus enables
zone reclaim.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/include/asm/topology.h
===
--- powerpc.git.orig/arch/powerpc/include/
Hi,
> The patch below sets a smaller value for RECLAIM_DISTANCE and thus enables
> zone reclaim.
FYI even with this enabled I could trip it up pretty easily with a multi
threaded application. I tried running stream across all threads in node 0. The
machine looks like:
node 0 cpus: 0 1 2 3 4 5 6
Hi Balbir,
> A reclaim distance of 10 implies a ratio of 1, that means we'll always
> do zone_reclaim() to free page cache and slab cache before moving on
> to another node?
I want to make an effort to reclaim local pagecache before ever going
off node. As an example, a completely off node stre
Hi Mel,
> You're pretty much on the button here. Only one thread at a time enters
> zone_reclaim. The others back off and try the next zone in the zonelist
> instead. I'm not sure what the original intention was but most likely it
> was to prevent too many parallel reclaimers in the same zone po
/0xd4
I assume there are some O(n^2) issues in sysfs as we add all the memory
nodes. Bumping SECTION_SIZE_BITS to 256 MB drops the time to about 10
seconds and results in a much smaller /sys.
Signed-off-by: Anton Blanchard
---
--- linux-2.6.33/arch/powerpc/include/asm/sparsemem.h~ 2010-02-2
Hi,
> e500v1/v2 based chips will treat any reserved field being set in an
> opcode as illegal. Thus always setting the hint in the opcode is
> a bad idea.
>
> Anton should be kept away from the powerpc opcode map.
Nasty bug! I promise never to touch the opcode map again :)
Hi,
I haven't tracked down why, but e74d098c66543d0731de62eb747ccd5b636a6f4c
(hvc_console: Fix race between hvc_close and hvc_remove) is causing my PowerPC
box to hang during boot, right when init starts.
Anton
___
Linuxppc-dev mailing list
Linuxppc-de
).
Signed-off-by: Anton Blanchard
--
I assume the function prototypes for the other IPC methods would cause us
to sign or zero extend second where appropriate (avoiding any security
issues). Come to think of it, the syscall wrappers for each method should do
that for us as well.
diff --git a/ipc
Hi,
> > ipc/syscall.c:17: error: conflicting types for ‘sys_ipc’
> > include/linux/syscalls.h:691: note: previous declaration of ‘sys_ipc’ was
> > here
>
> Hmm. Right you are. Why don't I see this? (I already applied the patch)
>
> Ahh. Because this only triggers with __ARCH_WANT_SYS_IPC. But
Hi,
> And this suggests that hvc_kick() is called before hvc_task is
> initialised, ie, before hvc_init() is called.
>
> Does this help?
Looks good, tests OK on my POWER5 box. Thanks!
Tested-by: Anton Blanchard
> diff --git a/drivers/char/hvc_console.c b/drivers/char/hvc_co
;
Since the outside kref get/put balance we only do a single kref_put when
count reaches 0.
The patch below changes things to call tty_kref_put once for every
hvc_close call, and with that my machine boots fine.
Signed-off-by: Anton Blanchard
---
diff --git a/drivers/char/hvc_console.c b/driv
the client architecture, and if we see "1" in the
ibm,associativity-form property firmware supports form 1 affinity and
we should look at the first field in the ibm,associativity-reference-points
array. If not we use the second field as we always have.
Signed-off-by: Anton Blanchard
This is a first attempt at converting PowerPC across to the new cpumask
API.
I need to audit the GFP_ flags of alloc_cpumask_var. There are also a few
areas still to be done: cell, xmon, kexec, crashdump and the cputhreads code
Anton
___
Linuxppc-dev
Change &cpu_online_map to cpu_online_mask.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/pasemi/cpufreq.c
===
--- linux-cpumask.orig/arch/powerpc/platforms/pasemi/cpufreq.c 2010-04-08
1
Use cpumask_first, cpumask_next in rtasd code.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/rtasd.c
===
--- linux-cpumask.orig/arch/powerpc/kernel/rtasd.c 2010-04-08
19:46:04.419206665 +1000
Use the new cpumask_* functions and dynamically allocate the cpumask in
smp_cpus_done.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/smp.c
===
--- linux-cpumask.orig/arch/powerpc/kernel/smp.c
Use new cpumask_* functions, and dynamically allocate cpumask in fixup_irqs.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/irq.c
===
--- linux-cpumask.orig/arch/powerpc/kernel/irq.c2010-04-08
Use new cpumask functions in iseries SMP startup code.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/iseries/smp.c
===
--- linux-cpumask.orig/arch/powerpc/platforms/iseries/smp.c 2010-04-08
19
Use new cpumask functions in pseries SMP startup code.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/pseries/smp.c
===
--- linux-cpumask.orig/arch/powerpc/platforms/pseries/smp.c 2010-04-08
19
Use the new cpumask API and add some comments to clarify how get_irq_server
works.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/pseries/xics.c
===
--- linux-cpumask.orig/arch/powerpc/platforms
This separates the per cpu output from the summary output at the end of the
file, making it easier to convert to the new cpumask API in a subsequent
patch.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/setup-common.c
Use new cpumask API in /proc/cpuinfo code.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/setup-common.c
===
--- linux-cpumask.orig/arch/powerpc/kernel/setup-common.c 2010-04-22
13:03
lable (smp_prepare_cpus), similar
to x86.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/include/asm/smp.h
===
--- linux-cpumask.orig/arch/powerpc/include/asm/smp.h 2010-04-22
13:03:14.635445670 +1000
+++ linux-cpumask
Convert hotplug-cpu code to new cpumask API.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/platforms/pseries/hotplug-cpu.c
===
--- linux-cpumask.orig/arch/powerpc/platforms/pseries/hotplug-cpu.c
2010-04
Convert NUMA code to new cpumask API. We shift the node to cpumask
setup code until after we complete bootmem allocation so we can
dynamically allocate the cpumasks.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/include/asm/mmzone.h
Since the *_map cpumask variants are deprecated, change the comments to
instead refer to *_mask.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/kernel/setup-common.c
===
--- linux-cpumask.orig/arch/powerpc
to take a cpumask. This allows us
to clean up the mess in smp_mpic_message_pass.
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/sysdev/mpic.c
===
--- linux-cpumask.orig/arch/powerpc/sysdev/mpic.c 2010-04-22
Enable the DEBUG_PER_CPU_MAPS option so we can look for problems with
cpumasks .
Signed-off-by: Anton Blanchard
---
Index: linux-cpumask/arch/powerpc/Kconfig.debug
===
--- linux-cpumask.orig/arch/powerpc/Kconfig.debug 2010-04
Hi,
Alan Modra pointed out that he added an option to PowerPC gcc years ago
specifically for us to do lightweight mcount profiling.
The normal PowerPC gcc mcount stuff forces a stack spill and gets itself
tangled up in the function prolog, making it impossible to nop out easily:
# gcc -pg:
000
ode.
The patch below sets a smaller value for RECLAIM_DISTANCE and thus enables
zone reclaim.
Signed-off-by: Anton Blanchard
---
Index: powerpc.git/arch/powerpc/include/asm/topology.h
===
--- powerpc.git.orig/arch/powerpc/include/
the client architecture, and if we see "1" in the
ibm,associativity-form property firmware supports form 1 affinity and
we should look at the first field in the ibm,associativity-reference-points
array. If not we use the second field as we always have.
Signed-off-by: Anton Blanchard
801 - 900 of 1205 matches
Mail list logo
