Date:   Fri, 19 Nov 2021 14:10:18 +0000
From:   Matthew Wilcox <willy@infradead.org>
To:     NeilBrown <neilb@suse.de>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Michal Hocko <mhocko@suse.com>,
        Thierry Reding <thierry.reding@gmail.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] MM: discard __GFP_ATOMIC
Message-ID: <YZewStgII83phKtA@casper.infradead.org>
References: <163712397076.13692.4727608274002939094@noble.neil.brown.name>
 <YZUBIbALcSHn4Rub@casper.infradead.org>
 <163727727803.13692.15470049610672496362@noble.neil.brown.name>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <163727727803.13692.15470049610672496362@noble.neil.brown.name>
Precedence: bulk

On Fri, Nov 19, 2021 at 10:14:38AM +1100, NeilBrown wrote:
> On Thu, 18 Nov 2021, Matthew Wilcox wrote:
> > Surely this should be gfpflags_allow_blocking() instead of poking about
> > in the innards of gfp flags?
> 
> Possibly.  Didn't know about gfpflags_allow_blocking().  From a quick
> grep in the kernel, a whole lot of other people don't know about it
> either, though clearly some do.
> 
> Maybe we should reaname "__GFP_DIRECT_RECLAIM" to
> "__GFP_ALLOW_BLOCKING", because that is what most users seems to care
> about.

I tend towards the school of thought that the __GFP flags should make
sense to the implementation and users should use either GFP_ or functions.
When we see users adding or subtracting __GFP flags, that's a problem.

> If not, then we probably want a gfpflags_without_block() function that
> removes that flag, as lots of code wants to do that - and using the flag
> for one, and an inline for the other is not consistent.

It's not a _lot_ of code ...

block/bio.c
drivers/net/ethernet/mellanox/mlx4/icm.c
drivers/vhost/net.c
fs/btrfs/disk-io.c
fs/btrfs/volumes.c
fs/erofs/zdata.c
fs/fscache/cookie.c
fs/fscache/page.c
kernel/cgroup/cgroup.c
net/core/skbuff.c
net/core/sock.c
net/netlink/af_netlink.c

(excluding mm -- if mm wants to meddle with GFP flags, that's fine)

I think a lot of these are probably misguided, eg the filesystems should
probably be using GFP_NOFS to prevent recursion.