Re: 2.6.15-rc5: latency regression vs 2.6.14 in exit_mmap->free_pgtables

[Lee Revell <rlrevell@joe-job.com>]

On Wed, 2005-12-28 at 22:59 +0000, Hugh Dickins wrote:
> I'm to blame for that - sorry.  It didn't occur to me that I was
> moving any signficant amount of work (on mms with many vmas) into the
> section with preemption disabled.  Actually, the mm->page_table_lock
> is _not_ held there any more; but preemption is still disabled while
> using the per-cpu mmu_gathers.
> 
> I wish you'd found it at -rc1 time.  It's not something that can
> be properly corrected in a hurry.  The proper fix is to rework the
> tlb_gather_mmu stuff, so it can be done without preemption disabled.
> It's already a serious limitation in unmap_vmas, with CONFIG_PREEMPT's
> ZAP_BLOCK_SIZE spoiling throughput with far too many TLB flushes.
> 
> On my list to work on; but the TLB always needs great care, and this
> goes down into architectural divergences, with truncation of a mapped
> file adding further awkward constraints.  I imagine 2.6.16-rc1 is only
> a couple of weeks away, so it's unlikely to be fixed in 2.6.16 either.
> 

Hugh,

Is this believed to be fixed in 2.6.16-rc1?

Lee

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/