Re: [PATCH 1/8] writeback: get rid of generic_sync_sb_inodes()export

From: Jan Kara
Date: Wed Sep 02 2009 - 10:01:44 EST

On Wed 02-09-09 12:26:26, Jens Axboe wrote:
> On Wed, Sep 02 2009, Jan Kara wrote:
> > On Wed 02-09-09 10:42:40, Jens Axboe wrote:
> > > This adds two new exported functions:
> > >
> > > - sync_inodes_sb(), which writes out dirty inodes on a super_block, and
> > > - sync_inodes_sb_wait(), which does the same but also waits for IO
> > > completion.
> > This is a nice cleanup. I only find the name sync_inodes_sb() slightly
> > misleading and the comment by that function as well. The name should rather
> > be something like writeback_inodes_sb() (and sync_inodes_sb_wait() could
> > stay just sync_inodes_sb()) - the writeback it does does not really
> > guarantee anything. For example it can skip inodes or pages it does not
> > like for some reason. What that function really does is - try to write some
> > dirty pages on that superblock and don't try too hard.
> > I don't insist on the renaming of the function but I really thing the
> > comment should be improved.
> I don't disagree, I was a bit torn on the naming as well. I will make
> that change, thanks for the feedback!
OK, thanks.

> I'd really like your feedback on the pin_sb_for_writeback() stuff too,
> since that is the contentious bit. And, this goes for others as well,
> I'd appreciate any reviewed-by and/or acked-by on patches.
I'll have a look at it probably tomorrow.

Jan Kara <jack@xxxxxxx>
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at