Alexander Viro wrote:
> 
> On Fri, 11 May 2001, Andrew Morton wrote:
> 
> > Okay, but where can ext3 do its mount-time truncation calls?
> >
> > >From what you're saying, it sounds like dropping the superblock
> > lock inside read_super() won't be the end of the world...
> 
> It's definitely OK as soon as you've got your initialization far enough
> to make write_super() survive.
> 

OK, thanks.  We're fine then - the internal truncate is the last
thing which happens inside ext3_read_super().

ext3_write_super() was deadlocking as well - in 2.4 it is called
inside lock_super(), and it blocks on operations by another thread,
which wants to do lock_super()....

So I made ext3_write_super asynchronous.  It'll hit disk within
a fraction of a second, but we can't block on it.  The alternative
would be to pop the superblock lock while we wait, but that's
probably worse. (ext2_write_super is async, so...)
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [EMAIL PROTECTED]

Reply via email to