From: Mark Geisert <mark@maxrnd.com>
To: cygwin-patches@cygwin.com
Subject: Re: [PATCH] Cygwin: Have tmpfile(3) use O_TMPFILE
Date: Wed, 10 Feb 2021 01:07:30 -0800 [thread overview]
Message-ID: <69e6c815-eafe-1b9e-948c-fd64c977ae88@maxrnd.com> (raw)
In-Reply-To: <20210209152510.GV4251@calimero.vinschen.de>
Hi Corinna,
Corinna Vinschen via Cygwin-patches wrote:
> Hi Mark,
>
> On Feb 9 02:50, Mark Geisert wrote:
>> Per discussion on cygwin-developers, a Cygwin tmpfile(3) implementation
>> has been added to syscalls.cc. This overrides the one supplied by
>> newlib. Then the open(2) flag O_TMPFILE was added to the open call that
>> tmpfile internally makes.
>> ---
>> winsup/cygwin/release/3.2.0 | 4 ++++
>> winsup/cygwin/syscalls.cc | 20 ++++++++++++++++++++
>> 2 files changed, 24 insertions(+)
>>
>> diff --git a/winsup/cygwin/release/3.2.0 b/winsup/cygwin/release/3.2.0
>> index f748a9bc8..d02d16863 100644
>> --- a/winsup/cygwin/release/3.2.0
>> +++ b/winsup/cygwin/release/3.2.0
>> @@ -19,6 +19,10 @@ What changed:
>>
>> - A few FAQ updates.
>>
>> +- Have tmpfile(3) make use of Win32 FILE_ATTRIBUTE_TEMPORARY via open(2)
>> + flag O_TMPFILE.
>> + Addresses: https://cygwin.com/pipermail/cygwin/2021-January/247304.html
>> +
>>
>> Bug Fixes
>> ---------
>> diff --git a/winsup/cygwin/syscalls.cc b/winsup/cygwin/syscalls.cc
>> index 52a020f07..b79c1c7cd 100644
>> --- a/winsup/cygwin/syscalls.cc
>> +++ b/winsup/cygwin/syscalls.cc
>> @@ -5225,3 +5225,23 @@ pipe2 (int filedes[2], int mode)
>> syscall_printf ("%R = pipe2([%d, %d], %y)", res, read, write, mode);
>> return res;
>> }
>> +
>> +extern "C" FILE *
>> +tmpfile (void)
>> +{
>> + char *dir = getenv ("TMPDIR");
>
> This isn't what Linux tmpfile does. Per the man page, it tries to
> create the file in P_tmpdir first, and if that fails, it tries
> "/tmp".
Oops, I was following newlib's code here. I'll adjust this.
>> + if (!dir)
>> + dir = P_tmpdir;
>> + int fd = open (dir, O_RDWR | O_CREAT | O_BINARY | O_TMPFILE,
>
> You have to specify O_EXCL here. The idea is that this file cannot be
> made permanent, and missing the O_EXCL flag allows exactly that. See
> https://man7.org/linux/man-pages/man2/open.2.html, the lengthy
> description in terms of O_TMPFILE.
I started out with O_EXCL as you suggested, but found syscalls.cc:1504 reporting
EEXIST. Is there some clash there between fh->exists() and O_TMPFILE? Hmm.
..mark
next prev parent reply other threads:[~2021-02-10 9:07 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-09 10:50 Mark Geisert
2021-02-09 15:25 ` Corinna Vinschen
2021-02-10 9:07 ` Mark Geisert [this message]
2021-02-10 10:12 ` Corinna Vinschen
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
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=69e6c815-eafe-1b9e-948c-fd64c977ae88@maxrnd.com \
--to=mark@maxrnd.com \
--cc=cygwin-patches@cygwin.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.
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).