I'm still waiting for confirmation, but the attached patch clearly and greatly improved long-term stability of 2.5.67 for me. Without the patch, 2.5.67 would seemingly randomly get stuck occasionally (usually only when working remotely, from home). I won't claim victory just yet, but with this patch in, 2.5.67 has been running all day without any apparent problems so far. And with Arun's x86 patch, even ia32 binaries work fine again. 2.5 is starting to look quite good. --david --- From: David Mosberger <davidm@napali.hpl.hp.com> Subject: fix fs->lock deadlock Newsgroups: linux.kernel Date: Wed, 16 Apr 2003 19:50:12 +0200 Organization: linux.* mail to news gateway The patch below is needed to avoid a deadlock on fs->lock. Without the patch, if __emul_lookup_dentry() returns 0, we fail to reacquire current->fs->lock and then go ahead to read_unlock() it anyhow. Bad for your health. I believe the bug was introduced by this change set (about 9 weeks ago): http://linux.bkbits.net:8080/linux-2.5/diffs/fs/namei.c@1.63.1.2 --david ===== fs/namei.c 1.69 vs edited ===== --- 1.69/fs/namei.c Wed Apr 2 22:51:31 2003 +++ edited/fs/namei.c Wed Apr 16 10:18:40 2003 @@ -847,6 +847,7 @@ read_unlock(¤t->fs->lock); if (__emul_lookup_dentry(name,nd)) return 0; + read_lock(¤t->fs->lock); } nd->mnt = mntget(current->fs->rootmnt); nd->dentry = dget(current->fs->root); - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/Received on Wed Apr 16 17:15:33 2003
This archive was generated by hypermail 2.1.8 : 2005-08-02 09:20:13 EST