public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: libc-alpha@sourceware.org, Lukasz Majewski <lukma@denx.de>,
	Carlos O'Donell <carlos@redhat.com>
Subject: [PATCH 14/18] linux: Only use 64-bit syscall if required for mq_timedsend
Date: Thu, 17 Jun 2021 08:51:00 -0300	[thread overview]
Message-ID: <20210617115104.1359598-15-adhemerval.zanella@linaro.org> (raw)
In-Reply-To: <20210617115104.1359598-1-adhemerval.zanella@linaro.org>

For !__ASSUME_TIME64_SYSCALLS there is no need to issue a 64-bit syscall
if the provided timeout fits in a 32-bit one.  The 64-bit usage should
be rare since the timeout is a relative one.

Checked on i686-linux-gnu on a 4.15 kernel and on a 5.11 kernel
(with and without --enable-kernel=5.1) and on x86_64-linux-gnu.
---
 rt/tst-mqueue10.c                      | 10 ++++++++
 sysdeps/unix/sysv/linux/mq_timedsend.c | 35 ++++++++++++++------------
 2 files changed, 29 insertions(+), 16 deletions(-)

diff --git a/rt/tst-mqueue10.c b/rt/tst-mqueue10.c
index 7fb53a1cae..18795808e0 100644
--- a/rt/tst-mqueue10.c
+++ b/rt/tst-mqueue10.c
@@ -54,6 +54,16 @@ do_test (void)
     support_delete_timer (timer);
   }
 
+  {
+    timer_t timer = support_create_timer (0, 100000000, false, NULL);
+    /* Fill the internal buffer first.  */
+    TEST_COMPARE (mq_timedsend (q, msg, sizeof (msg), 0,
+				&(struct timespec) { 0, 0 }), 0);
+    TEST_COMPARE (mq_timedsend (q, msg, sizeof (msg), 0, &ts), -1);
+    TEST_VERIFY (errno == EINTR || errno == EOVERFLOW);
+    support_delete_timer (timer);
+  }
+
   mq_unlink (name);
 
   return 0;
diff --git a/sysdeps/unix/sysv/linux/mq_timedsend.c b/sysdeps/unix/sysv/linux/mq_timedsend.c
index 5f14ecb0bf..8fa8c010ad 100644
--- a/sysdeps/unix/sysv/linux/mq_timedsend.c
+++ b/sysdeps/unix/sysv/linux/mq_timedsend.c
@@ -29,30 +29,33 @@ __mq_timedsend_time64 (mqd_t mqdes, const char *msg_ptr, size_t msg_len,
 # ifndef __NR_mq_timedsend_time64
 #  define __NR_mq_timedsend_time64 __NR_mq_timedsend
 # endif
-  int ret = SYSCALL_CANCEL (mq_timedsend_time64, mqdes, msg_ptr, msg_len,
-			    msg_prio, abs_timeout);
 
-#ifndef __ASSUME_TIME64_SYSCALLS
-  if (ret == 0 || errno != ENOSYS)
-    return ret;
+#ifdef __ASSUME_TIME64_SYSCALLS
+  return SYSCALL_CANCEL (mq_timedsend_time64, mqdes, msg_ptr, msg_len,
+			 msg_prio, abs_timeout);
+#else
+  bool is32bit = abs_timeout != NULL
+		 ? in_time_t_range (abs_timeout->tv_sec) : true;
+  if (!is32bit)
+    {
+      int r = SYSCALL_CANCEL (mq_timedsend_time64, mqdes, msg_ptr, msg_len,
+			      msg_prio, abs_timeout);
+      if (r == 0 || errno != ENOSYS)
+	return r;
+      __set_errno (EOVERFLOW);
+      return -1;
+    }
 
-  struct timespec ts32;
+  struct timespec ts32, *pts32 = NULL;
   if (abs_timeout != NULL)
     {
-      if (! in_time_t_range (abs_timeout->tv_sec))
-        {
-          __set_errno (EOVERFLOW);
-          return -1;
-        }
-
       ts32 = valid_timespec64_to_timespec (*abs_timeout);
+      pts32 = &ts32;
     }
 
-  ret = SYSCALL_CANCEL (mq_timedsend, mqdes, msg_ptr, msg_len, msg_prio,
-			abs_timeout != NULL ? &ts32 : NULL);
+  return SYSCALL_CANCEL (mq_timedsend, mqdes, msg_ptr, msg_len, msg_prio,
+			 pts32);
 #endif
-
-  return ret;
 }
 
 #if __TIMESIZE != 64
-- 
2.30.2


  parent reply	other threads:[~2021-06-17 11:51 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-17 11:50 [PATCH 00/18] More y2038 fixes Adhemerval Zanella
2021-06-17 11:50 ` [PATCH 01/18] Use 64 bit time_t stat internally Adhemerval Zanella
2021-06-21  7:42   ` Lukasz Majewski
2021-06-22 19:37   ` Florian Weimer
2021-06-22 19:51     ` Adhemerval Zanella
2021-06-17 11:50 ` [PATCH 02/18] Use LFS and 64 bit time for installed programs Adhemerval Zanella
2021-06-17 12:19   ` Andreas Schwab
2021-06-18 18:50     ` Adhemerval Zanella
2021-06-17 20:49   ` Joseph Myers
2021-06-18 18:14     ` Adhemerval Zanella
2021-06-17 11:50 ` [PATCH 03/18] support: Add support_create_timer Adhemerval Zanella
2021-06-21  7:42   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 04/18] linux: Only use 64-bit syscall if required for ppoll Adhemerval Zanella
2021-06-21  7:42   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 05/18] linux: Only use 64-bit syscall if required for pselect Adhemerval Zanella
2021-06-21  7:42   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 06/18] linux: Only use 64-bit syscall if required for select Adhemerval Zanella
2021-06-21  7:43   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 07/18] linux: Remove supports_time64 () from clock_getres Adhemerval Zanella
2021-06-21  7:43   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 08/18] linux: Remove supports_time64 () from clock_gettime Adhemerval Zanella
2021-06-21  7:43   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 09/18] linux: Remove time64-support Adhemerval Zanella
2021-06-21  7:43   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 10/18] linux: timerfd_gettime minor cleanup Adhemerval Zanella
2021-06-21  7:43   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 11/18] linux: Only use 64-bit syscall if required for semtimedop Adhemerval Zanella
2021-06-21  7:43   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 12/18] linux: Only use 64-bit syscall if required for timerfd_settime Adhemerval Zanella
2021-06-21  7:44   ` Lukasz Majewski
2021-06-17 11:50 ` [PATCH 13/18] linux: Only use 64-bit syscall if required for mq_timedreceive Adhemerval Zanella
2021-06-21  7:44   ` Lukasz Majewski
2021-06-17 11:51 ` Adhemerval Zanella [this message]
2021-06-21  7:44   ` [PATCH 14/18] linux: Only use 64-bit syscall if required for mq_timedsend Lukasz Majewski
2021-06-17 11:51 ` [PATCH 15/18] linux: Only use 64-bit syscall if required for sigtimedwait Adhemerval Zanella
2021-06-17 12:25   ` Andreas Schwab
2021-06-22 14:58     ` Adhemerval Zanella
2021-06-17 11:51 ` [PATCH 16/18] linux: Only use 64-bit syscall if required for utimensat family Adhemerval Zanella
2021-06-21  7:45   ` Lukasz Majewski
2021-06-17 11:51 ` [PATCH 17/18] linux: Only use 64-bit syscall if required for internal futex Adhemerval Zanella
2021-06-21  7:45   ` Lukasz Majewski
2021-06-17 11:51 ` [PATCH 18/18] linux: Only use 64-bit syscall if required for clock_nanosleep Adhemerval Zanella
2021-06-17 15:11   ` Lukasz Majewski
2021-06-17 17:45     ` Adhemerval Zanella
2021-06-21  7:46   ` Lukasz Majewski
2021-06-17 14:20 ` [PATCH 00/18] More y2038 fixes Lukasz Majewski
2021-06-17 17:41   ` Adhemerval Zanella
2021-06-17 20:58     ` Joseph Myers

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=20210617115104.1359598-15-adhemerval.zanella@linaro.org \
    --to=adhemerval.zanella@linaro.org \
    --cc=carlos@redhat.com \
    --cc=libc-alpha@sourceware.org \
    --cc=lukma@denx.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.
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).