[PATCH] hurd: Check return value of mach_port_mod_refs() in the dup routine of fcntl()
Zhaoming Luo
zhmingluo@163.com
Wed Mar 5 22:39:04 GMT 2025
Ignoring the return value of mach_port_mod_ref() causes the situation
that when the reference count of the io server port is full, the caller
of dup() still supposes a new file discriptor is allocated. In this
patch we check the return value of mach_port_mod_ref() and return an
error when it fails.
Signed-off-by: Zhaoming Luo <zhmingluo@163.com>
---
sysdeps/mach/hurd/fcntl.c | 38 +++++++++++++++++++++++++++-----------
1 file changed, 27 insertions(+), 11 deletions(-)
diff --git a/sysdeps/mach/hurd/fcntl.c b/sysdeps/mach/hurd/fcntl.c
index a65c190c..d97226cd 100644
--- a/sysdeps/mach/hurd/fcntl.c
+++ b/sysdeps/mach/hurd/fcntl.c
@@ -83,18 +83,34 @@ __libc_fcntl (int fd, int cmd, ...)
result = -1;
else
{
- /* Give the ports each a user ref for the new descriptor. */
- __mach_port_mod_refs (__mach_task_self (), port,
+ /* Give the ports (i.e. `ctty` and `port`) each a user ref for
+ the new descriptor. */
+ err = __mach_port_mod_refs (__mach_task_self (), port,
MACH_PORT_RIGHT_SEND, 1);
- if (ctty != MACH_PORT_NULL)
- __mach_port_mod_refs (__mach_task_self (), ctty,
- MACH_PORT_RIGHT_SEND, 1);
-
- /* Install the ports and flags in the new descriptor. */
- if (ctty != MACH_PORT_NULL)
- _hurd_port_set (&new->ctty, ctty);
- new->flags = flags;
- _hurd_port_locked_set (&new->port, port); /* Unlocks NEW. */
+
+ if (err)
+ {
+ /* When an error occurs during giving a user ref to the
+ io server port */
+ result = -1;
+
+ if (err == KERN_UREFS_OVERFLOW)
+ errno = EMFILE;
+ else
+ errno = EINVAL;
+ }
+ else
+ {
+ if (ctty != MACH_PORT_NULL)
+ __mach_port_mod_refs (__mach_task_self (), ctty,
+ MACH_PORT_RIGHT_SEND, 1);
+
+ /* Install the ports and flags in the new descriptor. */
+ if (ctty != MACH_PORT_NULL)
+ _hurd_port_set (&new->ctty, ctty);
+ new->flags = flags;
+ _hurd_port_locked_set (&new->port, port); /* Unlocks NEW. */
+ }
}
HURD_CRITICAL_END;
--
2.47.2
More information about the Libc-alpha
mailing list