diff options
| author | Peter Chubb <peterc@gelato.unsw.edu.au> | 2005-06-08 15:50:20 -0700 |
|---|---|---|
| committer | Linus Torvalds <torvalds@ppc970.osdl.org> | 2005-06-08 16:21:14 -0700 |
| commit | 05062d96a23ec0959ee5ea969f40813170c73c0e (patch) | |
| tree | 9aa09b58ff455cc43b81cae10d3c4dac9fe9f462 /arch/ia64 | |
| parent | f829fd23c87918374bac0d90404fe12f0e788d52 (diff) | |
| download | kernel_samsung_aries-05062d96a23ec0959ee5ea969f40813170c73c0e.zip kernel_samsung_aries-05062d96a23ec0959ee5ea969f40813170c73c0e.tar.gz kernel_samsung_aries-05062d96a23ec0959ee5ea969f40813170c73c0e.tar.bz2 | |
[PATCH] ia64: fix floating-point preemption problem
There've been reports of problems with CONFIG_PREEMPT=y and the high
floating point partition. This is caused by the possibility of preemption
and rescheduling on a different processor while saving or restioirng the
high partition.
The only places where the FPU state is touched are in ptrace, in
switch_to(), and where handling a floating-point exception. In switch_to()
preemption is off. So it's only in trap.c and ptrace.c that we need to
prevent preemption.
Here is a patch that adds commentary to make the conditions clear, and adds
appropriate preempt_{en,dis}able() calls to make it so. In trap.c I use
preempt_enable_no_resched(), as we're about to return to user space where
the preemption flag will be checked anyway.
Signed-off-by: Peter Chubb <peterc@gelato.unsw.edu.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch/ia64')
| -rw-r--r-- | arch/ia64/kernel/ptrace.c | 6 | ||||
| -rw-r--r-- | arch/ia64/kernel/traps.c | 11 |
2 files changed, 16 insertions, 1 deletions
diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c index 08c8a5e..575a8f6 100644 --- a/arch/ia64/kernel/ptrace.c +++ b/arch/ia64/kernel/ptrace.c @@ -635,11 +635,17 @@ ia64_flush_fph (struct task_struct *task) { struct ia64_psr *psr = ia64_psr(ia64_task_regs(task)); + /* + * Prevent migrating this task while + * we're fiddling with the FPU state + */ + preempt_disable(); if (ia64_is_local_fpu_owner(task) && psr->mfh) { psr->mfh = 0; task->thread.flags |= IA64_THREAD_FPH_VALID; ia64_save_fpu(&task->thread.fph[0]); } + preempt_enable(); } /* diff --git a/arch/ia64/kernel/traps.c b/arch/ia64/kernel/traps.c index 9bad665..1861173 100644 --- a/arch/ia64/kernel/traps.c +++ b/arch/ia64/kernel/traps.c @@ -220,13 +220,21 @@ disabled_fph_fault (struct pt_regs *regs) /* first, grant user-level access to fph partition: */ psr->dfh = 0; + + /* + * Make sure that no other task gets in on this processor + * while we're claiming the FPU + */ + preempt_disable(); #ifndef CONFIG_SMP { struct task_struct *fpu_owner = (struct task_struct *)ia64_get_kr(IA64_KR_FPU_OWNER); - if (ia64_is_local_fpu_owner(current)) + if (ia64_is_local_fpu_owner(current)) { + preempt_enable_no_resched(); return; + } if (fpu_owner) ia64_flush_fph(fpu_owner); @@ -244,6 +252,7 @@ disabled_fph_fault (struct pt_regs *regs) */ psr->mfh = 1; } + preempt_enable_no_resched(); } static inline int |