luc14n0/glib
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/glib.git synced 2025-05-31 09:50:05 +02:00
Code Issues Packages Projects Releases Wiki Activity

gstdio: make g_close() async-signal-safe under certain conditions

Browse Source 
g_close() does something useful. It is not trivial to get EINTR handling of
close() right, in a portable manner. g_close() abstracts this.

We should allow glib users to use the function even in async-signal-safe
contexts, at least if the user heeds the caveat about GError and take care
not to fail assertions.

Backport 2.74: Modified to drop documentation changes to g_close() which
document its new async-signal-safe guarantees. They are not public
guarantees until 2.76. Also modified to include moving the code to
ignore `EINTR` from commit d5dc7d266f2b8d0f7d.
This commit is contained in:
Thomas Haller 2022-10-18 09:07:35 +02:00 committed by Philip Withnall
parent f0bcb7f79c
commit 0cfc5b054a
1 changed files with 31 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
glib/gstdio.c
View File

@ -1749,8 +1749,9 @@ g_utime (const gchar *filename,
* @fd: A file descriptor
* @error: a #GError
*
* This wraps the close() call; in case of error, %errno will be
* This wraps the close() call. In case of error, %errno will be
* preserved, but the error will also be stored as a #GError in @error.
* In case of success, %errno is undefined.
*
* Besides using #GError, there is another major reason to prefer this
* function over the call provided by the system; on Unix, it will
@ -1766,24 +1767,38 @@ g_close (gint fd,
GError **error)
{
int res;
res = close (fd);
/* Just ignore EINTR for now; a retry loop is the wrong thing to do
* on Linux at least. Anyone who wants to add a conditional check
* for e.g. HP-UX is welcome to do so later...
*
* http://lkml.indiana.edu/hypermail/linux/kernel/0509.1/0877.html
* https://bugzilla.gnome.org/show_bug.cgi?id=682819
* http://utcc.utoronto.ca/~cks/space/blog/unix/CloseEINTR
* https://sites.google.com/site/michaelsafyan/software-engineering/checkforeintrwheninvokingclosethinkagain
/* Important: if @error is NULL, we must not do anything that is
* not async-signal-safe.
*/
if (G_UNLIKELY (res == -1 && errno == EINTR))
return TRUE;
else if (res == -1)
res = close (fd);
if (res == -1)
{
int errsv = errno;
g_set_error_literal (error, G_FILE_ERROR,
g_file_error_from_errno (errsv),
g_strerror (errsv));
if (errsv == EINTR)
{
/* Just ignore EINTR for now; a retry loop is the wrong thing to do
* on Linux at least. Anyone who wants to add a conditional check
* for e.g. HP-UX is welcome to do so later...
*
* https://lwn.net/Articles/576478/
* http://lkml.indiana.edu/hypermail/linux/kernel/0509.1/0877.html
* https://bugzilla.gnome.org/show_bug.cgi?id=682819
* http://utcc.utoronto.ca/~cks/space/blog/unix/CloseEINTR
* https://sites.google.com/site/michaelsafyan/software-engineering/checkforeintrwheninvokingclosethinkagain
*/
return TRUE;
}
if (error)
{
g_set_error_literal (error, G_FILE_ERROR,
g_file_error_from_errno (errsv),
g_strerror (errsv));
}
errno = errsv;
return FALSE;
}