From: Christian Couder <christian.couder@gmail.com>
To: Taylor Blau <me@ttaylorr.com>
Cc: git <git@vger.kernel.org>, "Junio C Hamano" <gitster@pobox.com>,
"Jeff King" <peff@peff.net>,
"SZEDER Gábor" <szeder.dev@gmail.com>,
"Derrick Stolee" <dstolee@microsoft.com>,
"Christian Couder" <chriscool@tuxfamily.org>
Subject: Re: [RFC PATCH] fetch-pack: try harder to read an ERR packet
Date: Tue, 28 Apr 2020 09:39:46 +0200 [thread overview]
Message-ID: <CAP8UFD2DdziVH0esBr9h_T17LfA73QgosViEpedSZ6NU+X=JsQ@mail.gmail.com> (raw)
In-Reply-To: <20200422232744.GA19100@syl.local>
On Thu, Apr 23, 2020 at 1:27 AM Taylor Blau <me@ttaylorr.com> wrote:
>
> On Wed, Apr 22, 2020 at 06:33:57PM +0200, Christian Couder wrote:
>
> It looks like this may be missing a:
>
> From: SZEDER Gábor <szeder.dev@gmail.com>
>
> header.
Yeah, I wanted to have him as the author but forgot. The next version
will have it.
[...]
> > ---
> > This just formats the following patch from SZEDER Gábor:
> >
> > https://lore.kernel.org/git/20190830121005.GI8571@szeder.dev/
> >
> > I haven't tried to implement some suggestions discussed later
> > in the above thread like:
> >
> > - renaming send_request()
>
> Agreed that this is probably something we should do. Maybe
> 'send_request_retry' or something? That name is kind of terrible.
Not sure 'send_request_retry' is better as we are not really retrying
to send the request. My take would be something like
'send_request_read_err'. For now I have left it as is though.
> > - covering more code pathes
>
> I see where Peff raised this point originally, but I think that this is
> a good step forward as-is. No need to hold this up looking for complete
> coverage, since this is obviously improving the situation.
Ok.
> > - avoid blocking indefinitely by looking for an ERR packet
> > only if the write() resulted in ECONNRESET or EPIPE
>
> I think that I may have addressed this point below.
>
> > - first printing an error message with error_errno() before
> > going on to look for an ERR packet
>
> I'm not sure what I think about this one. I'd certainly welcome all
> opinions on this matter.
>
> > - implementing a timeout
>
> A timeout may be a good thing to do. See what you think about my
> suggestion below, first, though.
Ok, thanks for your suggestion.
> > diff --git a/fetch-pack.c b/fetch-pack.c
> > index 1734a573b0..63e8925e2b 100644
> > --- a/fetch-pack.c
> > +++ b/fetch-pack.c
> > @@ -185,14 +185,27 @@ static enum ack_type get_ack(struct packet_reader *reader,
> > }
> >
> > static void send_request(struct fetch_pack_args *args,
> > - int fd, struct strbuf *buf)
> > + int fd, struct strbuf *buf,
> > + struct packet_reader *reader)
> > {
> > if (args->stateless_rpc) {
> > send_sideband(fd, -1, buf->buf, buf->len, LARGE_PACKET_MAX);
> > packet_flush(fd);
> > } else {
> > - if (write_in_full(fd, buf->buf, buf->len) < 0)
> > + if (write_in_full(fd, buf->buf, buf->len) < 0) {
>
> This makes sense; if 'write_in_full' fails, we want to go into the
> error-handling case below.
>
> But, I wonder if we could do better than re-using 'write_in_full' here.
> We definitely do want to write 'buf->len' bytes overall, but what
> happens when a 'write()' fails? I think it's at _that_ point we want to
> try and read a packet or two off from the remote.
Yeah, good idea.
> What if instead this looked something like:
>
> const char *p = buf;
> ssize_t total = 0;
>
> while (count > 0) {
> ssize_t written = xwrite(fd, p, count);
> if (written < 0)
> return -1;
> /* note the change on this line */
> if (!written && packet_reader_read(reader) == PACKET_READ_EOF) {
> errno = ENOSPC;
> return -1;
> }
> count -= written;
> p += written;
> total += written;
> }
>
> return total;
>
> That is basically the definition of 'write_in_full', but when we didn't
> get a chance to write anything, then we try to read one packet.
Yeah, your code above looks correct. I have added a new function doing
the above in the new version I will send soon.
> This way, we only read exactly as many packets as we need to when we hit
> this case. I'm not sure that it matters in practice, though.
I am not sure I understand what you think doesn't matter in practice.
Thanks,
Christian.
next prev parent reply other threads:[~2020-04-28 7:40 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-22 16:33 [RFC PATCH] fetch-pack: try harder to read an ERR packet Christian Couder
2020-04-22 23:27 ` Taylor Blau
2020-04-28 7:39 ` Christian Couder [this message]
2020-04-28 8:33 ` Jeff King
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='CAP8UFD2DdziVH0esBr9h_T17LfA73QgosViEpedSZ6NU+X=JsQ@mail.gmail.com' \
--to=christian.couder@gmail.com \
--cc=chriscool@tuxfamily.org \
--cc=dstolee@microsoft.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=me@ttaylorr.com \
--cc=peff@peff.net \
--cc=szeder.dev@gmail.com \
/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).