Skip to content
  1. Sep 02, 2010
    • Anton Blanchard's avatar
      powerpc: Feature nop out reservation clear when stcx checks address · f89451fb
      Anton Blanchard authored 
      The POWER architecture does not require stcx to check that it is operating
on the same address as the larx. This means it is possible for an
an exception handler to execute a larx, get a reservation, decide
not to do the stcx and then return back with an active reservation. If the
interrupted code was in the middle of a larx/stcx sequence the stcx could
incorrectly succeed.

All recent POWER CPUs check the address before letting the stcx succeed
so we can create a CPU feature and nop it out. As Ben suggested, we can
only do this in our syscall path because there is a remote possibility
some kernel code gets interrupted by an exception that ends up operating
on the same cacheline.

Thanks to Paul Mackerras and Derek Williams for the idea.

To test this I used a very simple null syscall (actually getppid) testcase
at http://ozlabs.org/~anton/junkcode/null_syscall.c



I tested against 2.6.35-git10 with the following changes against the
pseries_defconfig:

CONFIG_VIRT_CPU_ACCOUNTING=n
CONFIG_AUDIT=n
CONFIG_PPC_4K_PAGES=n
CONFIG_PPC_64K_PAGES=y
CONFIG_FORCE_MAX_ZONEORDER=9
CONFIG_PPC_SUBPAGE_PROT=n
CONFIG_FUNCTION_TRACER=n
CONFIG_FUNCTION_GRAPH_TRACER=n
CONFIG_IRQSOFF_TRACER=n
CONFIG_STACK_TRACER=n

to remove the overhead of virtual CPU accounting, syscall auditing and
the ftrace mcount tracers. 64kB pages were enabled to minimise TLB misses.

POWER6: +8.2%
POWER7: +7.0%

Another suggestion was to use a larx to something in the L1 instead of a stcx.
This was almost as fast as removing the larx on POWER6, but only 3.5% faster
on POWER7. We can use this to speed up the reservation clear in our
exception exit code.

Signed-off-by: default avatarAnton Blanchard <anton@samba.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      f89451fb
    • Anton Blanchard's avatar
      powerpc: Add 64bit csum_and_copy_to_user · 8c773914
      Anton Blanchard authored 
      

This adds the equivalent of csum_and_copy_from_user for the receive side so we
can copy and checksum in one pass. It is modelled on the generic checksum
routine.

Signed-off-by: default avatarAnton Blanchard <anton@samba.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      8c773914
    • Anton Blanchard's avatar
      powerpc: Optimise 64bit csum_partial_copy_generic and add csum_and_copy_from_user · fdd374b6
      Anton Blanchard authored 
      

We use the same core loop as the new csum_partial, adding in the
stores and exception handling code. To keep things simple we do all the
exception fixup in csum_and_copy_from_user. This wrapper function is
modelled on the generic checksum code and is careful to always calculate
a complete checksum even if we only copied part of the data to userspace.

To test this I forced checksumming on over loopback and ran socklib (a
simple TCP benchmark). On a POWER6 575 throughput improved by 19% with
this patch. If I forced both the sender and receiver onto the same cpu
(with the hope of shifting the benchmark from being cache bandwidth limited
to cpu limited), adding this patch improved performance by 55%

Signed-off-by: default avatarAnton Blanchard <anton@samba.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      fdd374b6
    • Anton Blanchard's avatar
      powerpc: Optimise 64bit csum_partial · 9b83ecb0
      Anton Blanchard authored 
      

The main loop of csum_partial runs very slowly on recent POWER CPUs. After some
analysis on both POWER6 and POWER7 I came up with routine below. First we get
the source aligned to a double word, ignoring any odd alignment to keep things
simple. Then we do 64 bytes at a time, with an entry and exit limb of a further
64 bytes. On both POWER6 and POWER7 this should be as fast as we can go since
we are limited by the latency of the adde instructions.

To test this I forced checksumming on over loopback and ran socklib (a
simple TCP benchmark). On a POWER6 575 throughput improved by 11% with
this patch.

Signed-off-by: default avatarAnton Blanchard <anton@samba.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      9b83ecb0
    • Nathan Fontenot's avatar
      powerpc/pseries: Correct rtas_data_buf locking in dlpar code · 93f68f1e
      Nathan Fontenot authored 
      

The dlpar code can cause a deadlock to occur when making the RTAS
configure-connector call.  This occurs because we make kmalloc calls,
which can block, while parsing the rtas_data_buf and holding the
rtas_data_buf_lock.  This an cause issues if someone else attempts
to grab the rtas_data_bug_lock.

This patch alleviates this issue by copying the contents of the rtas_data_buf
to a local buffer before parsing.  This allows us to only hold the
rtas_data_buf_lock around the RTAS configure-connector calls.

Signed-off-by: default avatarNathan Fontenot <nfont@austin.ibm.com>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      93f68f1e
  3. Sep 01, 2010
  5. Aug 31, 2010
    • Michael Neuling's avatar
      powerpc: Don't use kernel stack with translation off · 54a83404
      Michael Neuling authored 
      In f761622e

 we changed
early_setup_secondary so it's called using the proper kernel stack
rather than the emergency one.

Unfortunately, this stack pointer can't be used when translation is off
on PHYP as this stack pointer might be outside the RMO.  This results in
the following on all non zero cpus:
  cpu 0x1: Vector: 300 (Data Access) at [c00000001639fd10]
      pc: 000000000001c50c
      lr: 000000000000821c
      sp: c00000001639ff90
     msr: 8000000000001000
     dar: c00000001639ffa0
   dsisr: 42000000
    current = 0xc000000016393540
    paca    = 0xc000000006e00200
      pid   = 0, comm = swapper

The original patch was only tested on bare metal system, so it never
caught this problem.

This changes __secondary_start so that we calculate the new stack
pointer but only start using it after we've called early_setup_secondary.

With this patch, the above problem goes away.

Signed-off-by: default avatarMichael Neuling <mikey@neuling.org>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      54a83404
    • Paul Mackerras's avatar
      powerpc/perf_event: Reduce latency of calling perf_event_do_pending · b0d278b7
      Paul Mackerras authored 
      Commit 0fe1ac48

 ("powerpc/perf_event: Fix oops due to
perf_event_do_pending call") moved the call to perf_event_do_pending
in timer_interrupt() down so that it was after the irq_enter() call.
Unfortunately this moved it after the code that checks whether it
is time for the next decrementer clock event.  The result is that
the call to perf_event_do_pending() won't happen until the next
decrementer clock event is due.  This was pointed out by Milton
Miller.

This fixes it by moving the check for whether it's time for the
next decrementer clock event down to the point where we're about
to call the event handler, after we've called perf_event_do_pending.

This has the side effect that on old pre-Core99 Powermacs where we
use the ppc_n_lost_interrupts mechanism to replay interrupts, a
replayed interrupt will incur a little more latency since it will
now do the code from the irq_enter down to the irq_exit, that it
used to skip.  However, these machines are now old and rare enough
that this doesn't matter.  To make it clear that ppc_n_lost_interrupts
is only used on Powermacs, and to speed up the code slightly on
non-Powermac ppc32 machines, the code that tests ppc_n_lost_interrupts
is now conditional on CONFIG_PMAC as well as CONFIG_PPC32.

Signed-off-by: default avatarPaul Mackerras <paulus@samba.org>
Cc: stable@kernel.org
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      b0d278b7
    • Matthew McClintock's avatar
      powerpc/kexec: Adds correct calling convention for kexec purgatory · 4562c986
      Matthew McClintock authored 
      

Call kexec purgatory code correctly. We were getting lucky before.
If you examine the powerpc 32bit kexec "purgatory" code you will
see it expects the following:

>From kexec-tools: purgatory/arch/ppc/v2wrap_32.S
-> calling convention:
->   r3 = physical number of this cpu (all cpus)
->   r4 = address of this chunk (master only)

As such, we need to set r3 to the current core, r4 happens to be
unused by purgatory at the moment but we go ahead and set it
here as well

Signed-off-by: default avatarMatthew McClintock <msm@freescale.com>
Signed-off-by: default avatarBenjamin Herrenschmidt <benh@kernel.crashing.org>
      4562c986
  7. Aug 29, 2010
  9. Aug 28, 2010
    • Dan Carpenter's avatar
      ALSA: pcm: add more format names · 7a28826a
      Dan Carpenter authored 
      There were some new formats added in commit 15c0cee6

 "ALSA: pcm:
Define G723 3-bit and 5-bit formats".  That commit increased
SNDRV_PCM_FORMAT_LAST as well.  My concern is that there are a couple
places which do:

        for (i = 0; i < SNDRV_PCM_FORMAT_LAST; i++) {
                if (dummy->pcm_hw.formats & (1ULL << i))
                        snd_iprintf(buffer, " %s", snd_pcm_format_name(i));
        }

I haven't tested these but it looks like if "i" were equal to
SNDRV_PCM_FORMAT_G723_24 or higher then we might read past the end of
the array.

Signed-off-by: default avatarDan Carpenter <error27@gmail.com>
Signed-off-by: default avatarTakashi Iwai <tiwai@suse.de>
      7a28826a