Messages in this thread | | | From | Daniel Phillips <> | Subject | Re: [Tux3] Tux3 report: Tux3 Git tree available | Date | Sat, 14 Mar 2009 20:58:57 -0700 |
| |
On Thursday 12 March 2009, OGAWA Hirofumi wrote: > Andrew Morton <akpm@linux-foundation.org> writes: > > > On Fri, 13 Mar 2009 01:18:29 +0900 > > OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> wrote: > > > >> Andrew Morton <akpm@linux-foundation.org> writes: > >> > >> > - When 'bh' is known to be non-NULL, use put_bh() rather than brelse(). > >> > >> It sounds strange. Almost all bh is non-NULL. This means we are going to > >> replace almost all brelse() by put_bh()? > >> > > > > Well.. you can make up your own mind about this. If you see benefit > > in the NULL-checking and extra debugging which brelse() provides then > > continue to use brelse(). > > I thought someone started to convert it. Ok, personally, I think > NULL-check is just not needed always, and if it is needed, check it > explicitly.
I checked in a patch last week to convert all the brelses to "blockput" (note how close the spelling is to Nick's block_put, an unsurprising accident). This should mean a little less work when the time comes to try out the block handles idea, and it surely doesn't hurt anything.
Regards,
Daniel
| |