All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Mosberger <davidm@hpl.hp.com>
To: linux-ia64@vger.kernel.org
Subject: [Linux-ia64] fix for 2.4.7 UP hang
Date: Thu, 26 Jul 2001 07:06:48 +0000	[thread overview]
Message-ID: <marc-linux-ia64-105590693005921@msgid-missing> (raw)

The attached patch fixes the bug that causes 2.4.7 to hang on UP
systems.  It turns out that this is a very old bug (since 2.3.11!)
that happens to get triggered reliably in 2.4.7 on UP machines.  The
problem is that in the UP case, we failed to call schedule_tail()
after a fork/clone.  This had the effect that if a task yielded the
CPU and the next task to be execute was a newly forked/cloned task,
then the yielding task would never get scheduled again.  In other
words, on a UP system it would have been possible for some tasks to
get "stuck", even though they were marked runnable.  Can't say I have
seen this in practice, but I'm not using UP machines very often
anymore.

	--david

--- lia64-kdb/arch/ia64/kernel/entry.S~	Sat Jul 21 20:20:10 2001
+++ lia64-kdb/arch/ia64/kernel/entry.S	Wed Jul 25 23:49:38 2001
@@ -492,15 +492,13 @@
 
 GLOBAL_ENTRY(ia64_ret_from_clone)
 	PT_REGS_UNWIND_INFO(0)
-#ifdef CONFIG_SMP
 	/*
-	 * In SMP mode, we need to call invoke_schedule_tail to complete the scheduling process.
+	 * We need to call schedule_tail() to complete the scheduling process.
 	 * Called by ia64_switch_to after do_fork()->copy_thread().  r8 contains the
 	 * address of the previously executing task.
 	 */
 	br.call.sptk.few rp=invoke_schedule_tail
 .ret8:
-#endif
 	adds r2=IA64_TASK_PTRACE_OFFSET,r13
 	;;
 	ld8 r2=[r2]
@@ -554,7 +552,7 @@
 #ifdef CONFIG_PERFMON
 (p9)	br.call.spnt.many b7=pfm_overflow_notify
 #endif
-# if __GNUC__ < 3
+#if __GNUC__ < 3
 (p7)	br.call.spnt.many b7=invoke_schedule
 #else
 (p7)	br.call.spnt.many b7=schedule
@@ -766,7 +764,6 @@
 	br.cond.sptk.many ia64_leave_kernel
 END(handle_syscall_error)
 
-# ifdef CONFIG_SMP
 	/*
 	 * Invoke schedule_tail(task) while preserving in0-in7, which may be needed
 	 * in case a system call gets restarted.
@@ -783,7 +780,7 @@
 	br.ret.sptk.many rp
 END(invoke_schedule_tail)
 
-# endif /* CONFIG_SMP */
+#if __GNUC__ < 3
 
 	/*
 	 * Invoke schedule() while preserving in0-in7, which may be needed
@@ -805,6 +802,8 @@
 	mov rp=loc0
 	br.ret.sptk.many rp
 END(invoke_schedule)
+
+#endif /* __GNUC__ < 3 */
 
 	/*
 	 * Setup stack and call ia64_do_signal.  Note that pSys and pNonSys need to


                 reply	other threads:[~2001-07-26  7:06 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=marc-linux-ia64-105590693005921@msgid-missing \
    --to=davidm@hpl.hp.com \
    --cc=linux-ia64@vger.kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.