Messages in this thread | | | Date | Thu, 12 Sep 2002 15:21:23 -0700 | From | Andrew Morton <> | Subject | Re: invalidate_inode_pages in 2.5.32/3 |
| |
Urban Widmark wrote: > > ... > If I understood Andrew right with the inode flag that won't purge anything > until the next userspace access. I don't think that is what smbfs wants > since the response to the oplock break should happen fairly soon ...
Well the lazy invalidation would be OK - defer that to the next userspace access, or just let the pages die a natural VM death, maybe poke `ksmbinvalidated'.
But if you want to perform writeback within the local IO daemon then hm. That's a problem which NFS seems to not have? I guess you'd have to leave i_sem held and poke `ksmbwritebackd'. Or teach pdflush about queued work items (a tq_struct would do it).
But it's the same story: the requirements of
a) non blocking local IO daemon and
b) assured pagecache takedown
are conflicting. You need at least one more thread, and locking against userspace activity. - 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/
| |