Source-Changes archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: CVS commit: [vmlocking] src/sys/kern




On Sep 10, 2007, at 2:16 AM, Andrew Doran wrote:

On Sun, Sep 09, 2007 at 07:56:02PM -0700, Jason Thorpe wrote:

On Sep 9, 2007, at 4:17 PM, Andrew Doran wrote:

- Re-enable pool_cache, since it works on i386 again after today's
pmap
change. pool_cache_invalidate() no longer invalidates objects stored
in the per-CPU caches. This needs some thought.

Then how do you invalidate the per-CPU caches? That needs to be fixed.

My working idea is to use a cross call to invalidate them so that a lock isn't needed to protect the per-CPU data. That shouldn't be a problem for the pagedaemon but since they can take "a while" to run through there may be other areas where it's not a goer. I'd really like to avoid adding a lock so it's a case of whether not that's a crosscall is generally viable, I think.

Sun didn't seem to think that a lock would be too terrible in this case. Why not take their advice?

-- thorpej




Home | Main Index | Thread Index | Old Index