linux-stable.git/arch/powerpc/kernel/entry_64.S, branch linux-3.5.y

powerpc: Restore correct DSCR in context switch

2012-09-14T16:59:41+00:00

commit 714332858bfd40dcf8f741498336d93875c23aa7 upstream.

During a context switch we always restore the per thread DSCR value.
If we aren't doing explicit DSCR management
(ie thread.dscr_inherit == 0) and the default DSCR changed while
the process has been sleeping we end up with the wrong value.

Check thread.dscr_inherit and select the default DSCR or per thread
DSCR as required.

This was found with the following test case, when running with
more threads than CPUs (ie forcing context switching):

http://ozlabs.org/~anton/junkcode/dscr_default_test.c

With the four patches applied I can run a combination of all
test cases successfully at the same time:

http://ozlabs.org/~anton/junkcode/dscr_default_test.c
http://ozlabs.org/~anton/junkcode/dscr_explicit_test.c
http://ozlabs.org/~anton/junkcode/dscr_inherit_test.c

Signed-off-by: Anton Blanchard 
Signed-off-by: Benjamin Herrenschmidt 
Signed-off-by: Greg Kroah-Hartman

ppc64: fix missing to check all bits of _TIF_USER_WORK_MASK in preempt

2012-06-29T04:35:35+00:00

In entry_64.S version of ret_from_except_lite, you'll notice that
in the !preempt case, after we've checked MSR_PR we test for any
TIF flag in _TIF_USER_WORK_MASK to decide whether to go to do_work
or not. However, in the preempt case, we do a convoluted trick to
test SIGPENDING only if PR was set and always test NEED_RESCHED ...
but we forget to test any other bit of _TIF_USER_WORK_MASK !!! So
that means that with preempt, we completely fail to test for things
like single step, syscall tracing, etc...

This should be fixed as the following path:

 - Test PR. If not set, go to resume_kernel, else continue.

 - If go resume_kernel, to do that original do_work.

 - If else, then always test for _TIF_USER_WORK_MASK to decide to do
that original user_work, else restore directly.

Signed-off-by: Tiejun Chen 
Signed-off-by: Benjamin Herrenschmidt

Merge branch 'merge' into next

2012-05-14T00:19:22+00:00

We want the irq fixes from the "merge" branch.

powerpc/irq: Fix another case of lazy IRQ state getting out of sync

2012-05-11T23:40:41+00:00

So we have another case of paca->irq_happened getting out of
sync with the HW irq state. This can happen when a perfmon
interrupt occurs while soft disabled, as it will return to a
soft disabled but hard enabled context while leaving a stale
PACA_IRQ_HARD_DIS flag set.

This patch fixes it, and also adds a test for the condition
of those flags being out of sync in arch_local_irq_restore()
when CONFIG_TRACE_IRQFLAGS is enabled.

This helps catching those gremlins faster (and so far I
can't seem see any anymore, so that's good news).

Signed-off-by: Benjamin Herrenschmidt

Merge branch 'merge' into next

2012-05-09T00:57:57+00:00

powerpc/irq: Fix bug with new lazy IRQ handling code

2012-05-08T23:42:21+00:00

We had a case where we could turn on hard interrupts while
leaving the PACA_IRQ_HARD_DIS bit set in the PACA. This can
in turn cause a BUG_ON() to hit in __check_irq_replay() due
to interrupt state getting out of sync.

The assembly code was also way too convoluted. Instead, we
now leave it to the C code to do the right thing which ends
up being smaller and more readable.

Signed-off-by: Benjamin Herrenschmidt

powerpc: Better scheduling of CR save code in system call path

2012-04-30T05:37:14+00:00

At the moment system call entry looks like:

crclr	so
...
mfcr	r9
...
std	r9,_CCR(r1)

commit bd19c8994a82 ([POWERPC] system call micro optimisation) put
some space between the crclr and mfcr in order to avoid a stall.

There is still a stall seen between the mfcr and std. We can avoid
the crclr by doing it in a GPR with rlwinm which gives us more room
to better schedule the sequence.

Signed-off-by: Anton Blanchard 
Signed-off-by: Benjamin Herrenschmidt

powerpc: No need to preserve count register across system call

2012-04-30T05:35:10+00:00

The count register is volatile so we don't need to preserve it.
Store zero to the entry in the exception frame.

Signed-off-by: Anton Blanchard 
Signed-off-by: Benjamin Herrenschmidt

powerpc: No need to save XER in a system call

2012-04-30T05:34:44+00:00

The XER is a volatile register so there is no need to save and restore
it over a system call - zero it out in the exception stack frame
instead.

This should fix a 5 cycle stall of the mfxer/std seen on POWER7.

Signed-off-by: Anton Blanchard 
Signed-off-by: Benjamin Herrenschmidt

powerpc: Hide some system call labels from profile tools

2012-04-30T05:34:43+00:00

syscall_dotrace_cont and syscall_error_cont tend to complicate perf
output so make them local.

Signed-off-by: Anton Blanchard 
Signed-off-by: Benjamin Herrenschmidt