On Sun, 4 Feb 2007 11:46:09 +0100 Nick Piggin <[email protected]> wrote:
> On Sun, Feb 04, 2007 at 02:30:55AM -0800, Andrew Morton wrote:
> > On Sun, 4 Feb 2007 11:15:29 +0100 Nick Piggin <[email protected]> wrote:
> >
> > > The write path is broken. I prefer my kernels slow, than buggy.
> >
> > That won't fly.
>
> What won't fly?
I suspect the performance cost of this approach would force us to redo it
all.
> > > > What happened to the idea of doing an atomic copy into the non-uptodate
> > > > page and handling it somehow?
> > >
> > > That was my second idea.
> >
> > Coulda sworn it was mine ;) I thought you ended up deciding it wasn't
> > practical because of the games we needed to play with ->commit_write.
>
> Maybe I misunderstood what you meant, above.
The original set of half-written patches I sent you. Do an atomic copy_from_user()
inside the page lock and if that fails, zero out the remainder of the page, run
commit_write() and then redo the whole thing.
> I have an alterative fix
> where a temporary page is allocated if the write enncounters a non
> uptodate page. The usercopy then goes into that page, and from there
> into the target page after we have opened the prepare_write().
Remember that a non-uptodate page is the common case.
> My *first* idea to fix this was to do the atomic copy into a non-uptodate
> page and then calling a zero-length commit_write if it failed. I pretty
> carefully constructed all these good arguments as to why each case works
> properly, but in the end it just didn't fly because it broke lots of
> filesystems.
I forget the details now. I think we did have a workable-looking solution
based on the atomic copy_from_user() but it would have re-exposed the old
problem wherein a page would fleetingly have a bunch of zeroes in the
middle of it, if someone looked at it during the write.
If that recollection is right, I think we could afford to reintroduce that
problem, frankly. Especially as it only happens in the incredibly rare
case of that get_user()ed page getting unmapped under our feet.
> > > but you introduce the theoretical memory deadlock
> > > where a task cannot reclaim its own memory.
> >
> > Nah, that'll never happen - both pages are already allocated.
>
> Both pages? I don't get it.
>
> You set the don't-reclaim vma flag, then run get_user, which takes a
> page fault and potentially has to allocate N pages for pagetables,
> pagecache readahead, buffers and fs private data and pagecache radix
> tree nodes for all of the pages read in.
Oh, OK. Need to do the get_user() twice then. Once before taking that new
rwsem.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
