From: Jeff King <peff@peff.net>
To: "René Scharfe" <l.s.r@web.de>
Cc: Junio C Hamano <gitster@pobox.com>,
git@vger.kernel.org,
Johannes Schindelin <Johannes.Schindelin@gmx.de>
Subject: Re: [PATCH] mingw: handle writes to non-blocking pipe
Date: Wed, 17 Aug 2022 01:39:32 -0400 [thread overview]
Message-ID: <Yvx/FNlTx+DCiCOj@coredump.intra.peff.net> (raw)
In-Reply-To: <c7c6524c-4f02-10f6-1a58-738cef5aecf2@web.de>
On Sun, Aug 14, 2022 at 05:37:08PM +0200, René Scharfe wrote:
> Turns out that's not the case on Windows: 94f4d01932 (mingw: workaround
> for hangs when sending STDIN, 2020-02-17) changed the compatibility
> implementation to 'Make `poll()` always reply "writable" for write end
> of the pipe.'.
Ah, good find. That kind of explains everything, then, I think. ;)
> > I'm not sure what "small enough" we can rely on, though. Really it is
> > the interplay between poll() and write() that we care about here. We
> > would like to know at what point poll() will tell us it's OK to write().
> > But we don't know what the OS thinks of that.
>
> Based on the output above I think Linux' poll() won't consider a pipe
> writable that has less than PIPE_BUF (4096) available bytes.
Right, that makes sense. It would have to in order to meet the atomicity
requirement for write(), but still always make forward progress for each
write().
> Perhaps we should take the advice about PIPE_NOWAIT in the docs serious
> and use overlapping (asynchronous) writes on Windows instead. This
> would mean reimplementing the whole pipe_command() with Windows API
> commands, I imagine.
I wouldn't be opposed to that, in the sense that it's supposed to be a
black box to the caller, and it's relatively small in size. But I think
we're pretty close to having something usable without that, so I'd like
to pursue a smaller fix in the interim.
> Avoiding xwrite() in pump_io_round() on top lets the test suite
> finish successfully.
That makes sense. We end up busy-looping between poll() and write()
while we wait for our read descriptor to become available. But if poll()
doesn't block, that's the best we can do.
-Peff
next prev parent reply other threads:[~2022-08-17 5:39 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-08-02 4:13 [RFC/PATCH] pipe_command(): mark stdin descriptor as non-blocking Jeff King
2022-08-02 15:04 ` Junio C Hamano
2022-08-02 15:39 ` Jeff King
2022-08-02 16:16 ` Junio C Hamano
2022-08-03 3:53 ` [PATCH v2] " Jeff King
2022-08-03 16:45 ` René Scharfe
2022-08-03 17:20 ` Jeff King
2022-08-03 21:56 ` René Scharfe
2022-08-05 15:36 ` Jeff King
2022-08-05 21:13 ` René Scharfe
2022-08-07 10:15 ` René Scharfe
2022-08-07 17:41 ` Jeff King
2022-08-10 5:39 ` René Scharfe
2022-08-10 19:53 ` Jeff King
2022-08-10 22:35 ` René Scharfe
2022-08-11 8:52 ` Jeff King
2022-08-10 5:39 ` [PATCH] mingw: handle writes to non-blocking pipe René Scharfe
2022-08-10 9:07 ` Johannes Schindelin
2022-08-10 20:02 ` Jeff King
2022-08-10 22:34 ` René Scharfe
2022-08-11 8:47 ` Jeff King
2022-08-11 17:35 ` René Scharfe
2022-08-11 18:20 ` Jeff King
2022-08-14 15:37 ` René Scharfe
2022-08-17 5:39 ` Jeff King [this message]
2022-08-17 6:04 ` [PATCH v2 0/6] fix pipe_command() deadlock Jeff King
2022-08-17 6:04 ` [PATCH v2 1/6] compat: add function to enable nonblocking pipes Jeff King
2022-08-17 20:23 ` Junio C Hamano
2022-08-18 5:41 ` Jeff King
2022-08-17 6:05 ` [PATCH v2 2/6] nonblock: support Windows Jeff King
2022-08-17 6:06 ` [PATCH v2 3/6] git-compat-util: make MAX_IO_SIZE define globally available Jeff King
2022-08-17 6:08 ` [PATCH v2 4/6] pipe_command(): avoid xwrite() for writing to pipe Jeff King
2022-08-17 6:09 ` [PATCH v2 5/6] pipe_command(): handle ENOSPC when writing to a pipe Jeff King
2022-08-17 18:57 ` Junio C Hamano
2022-08-18 5:38 ` Jeff King
2022-08-17 6:10 ` [PATCH v2 6/6] pipe_command(): mark stdin descriptor as non-blocking Jeff King
2022-08-17 6:20 ` [PATCH v2 0/6] fix pipe_command() deadlock Jeff King
2022-08-19 21:19 ` René Scharfe
2022-08-20 7:04 ` Jeff King
2022-08-07 10:14 ` [PATCH v2] pipe_command(): mark stdin descriptor as non-blocking René Scharfe
2022-08-08 12:55 ` Johannes Schindelin
2022-08-08 12:59 ` Johannes Schindelin
2022-08-09 13:04 ` Jeff King
2022-08-09 22:10 ` Johannes Schindelin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Yvx/FNlTx+DCiCOj@coredump.intra.peff.net \
--to=peff@peff.net \
--cc=Johannes.Schindelin@gmx.de \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=l.s.r@web.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).