Re: Bug: devfs is sure to have the bug.

From: Kohji Okuno <okuno.kohji_at_jp.panasonic.com>
Date: Thu, 04 Aug 2011 19:09:32 +0900 (JST)
Hello Kostik,

>> > On Thu, Aug 04, 2011 at 11:41:39AM +0900, Kohji Okuno wrote:
>> >> But, now I'm using 8.1-RELEASE. May I have advice about 8.X ?
>> > Do you mean a patch for the stable/8 ? I believe it is enough to
>> > apply rev. 211628 to stable/8, then the patch I posted yesterday
>> > should be compilable.
>> 
>> I'm sorry.
>> I need a patch for 8.1-RELEASE. Could you propose patch?
> Did you tried to apply the 211628 and the patch I mailed, to 8.1 ?
> I am not very interested in porting this stuff for such old system.
> On the other hand, I am unaware of large changes in devfs between
> 8.1 and latest stable.

No. Because the difference was large as you were poingint out, I did
not your patch. Now, I'm trying the following patch.

 devfs_populate(struct devfs_mount *dm)
 {
 
+#if 1
+	unsigned gen;
+	sx_assert(&dm->dm_lock, SX_XLOCKED);
+	gen = devfs_generation;
+	if (dm->dm_generation == gen)
+		return;
+	while (devfs_populate_loop(dm, 0))
+		continue;
+	dm->dm_generation = gen;
+#else
 	sx_assert(&dm->dm_lock, SX_XLOCKED);
 	if (dm->dm_generation == devfs_generation)
 		return;
 	while (devfs_populate_loop(dm, 0))
 		continue;
 	dm->dm_generation = devfs_generation;
+#endif
 }
 
 /*
Received on Thu Aug 04 2011 - 08:09:35 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:16 UTC