[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZKinHejv+xBq+gti@casper.infradead.org>
Date: Sat, 8 Jul 2023 01:00:29 +0100
From: Matthew Wilcox <willy@...radead.org>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Ahelenia ZiemiaĆska
<nabijaczleweli@...ijaczleweli.xyz>,
Christian Brauner <brauner@...nel.org>,
Jens Axboe <axboe@...nel.dk>,
David Howells <dhowells@...hat.com>,
Alexander Viro <viro@...iv.linux.org.uk>,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: Pending splice(file -> FIFO) excludes all other FIFO operations
forever (was: ... always blocks read(FIFO), regardless of O_NONBLOCK on read
side?)
On Thu, Jul 06, 2023 at 02:56:45PM -0700, Linus Torvalds wrote:
> +static int busy_pipe_buf_confirm(struct pipe_inode_info *pipe,
> + struct pipe_buffer *buf)
> +{
> + struct page *page = buf->page;
> +
> + if (folio_wait_bit_interruptible(page_folio(page), PG_locked))
> + return -EINTR;
Do we really want interruptible here rather than killable? That is,
do we want SIGWINCH or SIGALRM to result in a short read? I assume
it's OK to return a short read because userspace has explicitly asked
for O_NONBLOCK and can therefore be expected to actually check the
return value from read().
Powered by blists - more mailing lists