Messages in this thread | | | Date | Thu, 30 Apr 2009 18:06:06 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH] Fix dirty page accounting in redirty_page_for_writepage() |
| |
* Christoph Lameter <cl@linux.com> wrote:
> On Thu, 30 Apr 2009, Ingo Molnar wrote: > > > nice. Do these all get called with irqs off, all the time? > > In most cases yes. The VMstat patch includes an example were we > may not care too much about a event counter missing a beat once in > a while for platforms not supporting atomic per cpu ops. I know > this affects IA64. The cost of an atomic operations for an event > counter update (which would have avoided the potential of a > concurrent update) was not justifiable.
when you say "atomics", do you mean the classic meaning of atomics? Because there are no classic atomics involved. This is the before/after disassembly from Eric's commit 4e69489a0:
before:
c0436274: b8 01 00 00 00 mov $0x1,%eax c0436279: e8 42 40 df ff call c022a2c0 <add_preempt_count> c043627e: bb 20 4f 6a c0 mov $0xc06a4f20,%ebx c0436283: e8 18 ca f0 ff call c0342ca0 <debug_smp_processor_id> c0436288: 03 1c 85 60 4a 65 c0 add -0x3f9ab5a0(,%eax,4),%ebx c043628f: ff 03 incl (%ebx) c0436291: b8 01 00 00 00 mov $0x1,%eax c0436296: e8 75 3f df ff call c022a210 <sub_preempt_count> c043629b: 89 e0 mov %esp,%eax c043629d: 25 00 e0 ff ff and $0xffffe000,%eax c04362a2: f6 40 08 08 testb $0x8,0x8(%eax) c04362a6: 75 07 jne c04362af <sock_alloc+0x7f> c04362a8: 8d 46 d8 lea -0x28(%esi),%eax c04362ab: 5b pop %ebx c04362ac: 5e pop %esi c04362ad: c9 leave c04362ae: c3 ret c04362af: e8 cc 5d 09 00 call c04cc080 <preempt_schedule> c04362b4: 8d 74 26 00 lea 0x0(%esi,%eiz,1),%esi c04362b8: eb ee jmp c04362a8 <sock_alloc+0x78>
after:
c0436275: 64 83 05 20 5f 6a c0 addl $0x1,%fs:0xc06a5f20
There's no atomic instructions at all - the counters here are only accessed locally. They are local-irq-atomic, but not cacheline-atomic.
Ingo
| |