Re: [uml-devel] Occasional hang starting up.

[Jeff Dike <jdike@addtoit.com>]

[-- Attachment #1: Type: text/plain, Size: 765 bytes --]

On Fri, Mar 24, 2006 at 01:37:54AM +0100, Blaisorblade wrote:
> Yep, it seems it looks good. But 1 bug still and two suggestions.
> See also attached patch (it misses one of the suggestions because it's late 
> and I'm going to sleep).

I merged that patch.

> If possible (there should be such array constructors) this shouldn't be 
> redundant / duplicated with the initial declarations.
> 
> Like write_sigio_fds = (int[]) {-1,-1} and having a common macro like 
> 
> #define SIGIO_FDS_INIT {-1,-1}
> 
> static int write_sigio_fds[2] = SIGIO_FDS_INIT;
> ...
> write_sigio_fds = (int[]) SIGIO_FDS_INIT

I tried this - I couldn't get anything like that to compile.  I always
got an "incompatible types in assignment" error.

Below is the latest version.

				Jeff

[-- Attachment #2: sigio-race --]
[-- Type: text/plain, Size: 2349 bytes --]

# This fixes a race in the starting of write_sigio_thread.
# Previously, some of the data needed by the thread was initialized
# after the clone.  If the thread ean immediately, it would see the
# uninitialized data, including an empty pollfds, which would cause it
# to hang.
# We move the data initialization to before the clone, and adjust the
# error paths and cleanup accordingly.
#
# Signed-off-by: Jeff Dike <jdike@addtoit.com>

Index: linux-2.6.16/arch/um/os-Linux/sigio.c
===================================================================
--- linux-2.6.16.orig/arch/um/os-Linux/sigio.c	2006-03-23 17:30:27.000000000 -0500
+++ linux-2.6.16/arch/um/os-Linux/sigio.c	2006-03-23 17:44:11.000000000 -0500
@@ -270,42 +270,40 @@ void write_sigio_workaround(void)
 	/* Did we race? Don't try to optimize this, please, it's not so likely
 	 * to happen, and no more than once at the boot. */
 	if(write_sigio_pid != -1)
-		goto out_unlock;
+		goto out_free;
 
-	write_sigio_pid = run_helper_thread(write_sigio_thread, NULL,
-					    CLONE_FILES | CLONE_VM, &stack, 0);
-
-	if (write_sigio_pid < 0)
-		goto out_clear;
+	current_poll = ((struct pollfds) { .poll 	= p,
+					   .used 	= 1,
+					   .size 	= 1 });
 
 	if (write_sigio_irq(l_write_sigio_fds[0]))
-		goto out_kill;
+		goto out_free;
 
-	/* Success, finally. */
 	memcpy(write_sigio_fds, l_write_sigio_fds, sizeof(l_write_sigio_fds));
 	memcpy(sigio_private, l_sigio_private, sizeof(l_sigio_private));
 
-	current_poll = ((struct pollfds) { .poll 	= p,
-					   .used 	= 1,
-					   .size 	= 1 });
+	write_sigio_pid = run_helper_thread(write_sigio_thread, NULL,
+					    CLONE_FILES | CLONE_VM, &stack, 0);
 
-	sigio_unlock();
-	return;
+	if (write_sigio_pid < 0)
+		goto out_clear;
 
- out_kill:
-	l_write_sigio_pid = write_sigio_pid;
-	write_sigio_pid = -1;
 	sigio_unlock();
-	/* Going to call waitpid, avoid holding the lock. */
-	os_kill_process(l_write_sigio_pid, 1);
-	goto out_free;
+	return;
 
  out_clear:
 	write_sigio_pid = -1;
- out_unlock:
-	sigio_unlock();
+	current_poll = ((struct pollfds) { .poll	= NULL,
+					   .size	= 0,
+					   .used	= 0 });
+	write_sigio_fds[0] = -1;
+	write_sigio_fds[1] = -1;
+
+	sigio_private[0] = -1;
+	sigio_private[1] = -1;
  out_free:
 	kfree(p);
+	sigio_unlock();
  out_close2:
 	close(l_sigio_private[0]);
 	close(l_sigio_private[1]);