[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230711-legalisieren-qualvoll-c578e099c65a@brauner>
Date: Tue, 11 Jul 2023 11:39:57 +0200
From: Christian Brauner <brauner@...nel.org>
To: Wen Yang <wenyang.linux@...mail.com>
Cc: Alexander Viro <viro@...iv.linux.org.uk>,
Jens Axboe <axboe@...nel.dk>, Christoph Hellwig <hch@....de>,
Dylan Yudaken <dylany@...com>,
David Woodhouse <dwmw@...zon.co.uk>,
Matthew Wilcox <willy@...radead.org>,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] eventfd: avoid overflow to ULLONG_MAX when ctx->count is
0
On Mon, Jul 10, 2023 at 11:02:33PM +0800, Wen Yang wrote:
>
> On 2023/7/10 22:12, Christian Brauner wrote:
> > On Sun, Jul 09, 2023 at 02:54:51PM +0800, wenyang.linux@...mail.com wrote:
> > > From: Wen Yang <wenyang.linux@...mail.com>
> > >
> > > For eventfd with flag EFD_SEMAPHORE, when its ctx->count is 0, calling
> > > eventfd_ctx_do_read will cause ctx->count to overflow to ULLONG_MAX.
> > >
> > > Fixes: cb289d6244a3 ("eventfd - allow atomic read and waitqueue remove")
> > > Signed-off-by: Wen Yang <wenyang.linux@...mail.com>
> > > Cc: Alexander Viro <viro@...iv.linux.org.uk>
> > > Cc: Jens Axboe <axboe@...nel.dk>
> > > Cc: Christian Brauner <brauner@...nel.org>
> > > Cc: Christoph Hellwig <hch@....de>
> > > Cc: Dylan Yudaken <dylany@...com>
> > > Cc: David Woodhouse <dwmw@...zon.co.uk>
> > > Cc: Matthew Wilcox <willy@...radead.org>
> > > Cc: linux-fsdevel@...r.kernel.org
> > > Cc: linux-kernel@...r.kernel.org
> > > ---
> > So this looks ok but I would like to see an analysis how the overflow
> > can happen. I'm looking at the callers and it seems that once ctx->count
> > hits 0 eventfd_read() won't call eventfd_ctx_do_read() anymore. So is
> > there a caller that can call directly or indirectly
> > eventfd_ctx_do_read() on a ctx->count == 0?
> eventfd_read() ensures that ctx->count is not 0 before calling
> eventfd_ctx_do_read() and it is correct.
>
> But it is not appropriate for eventfd_ctx_remove_wait_queue() to call
> eventfd_ctx_do_read() unconditionally,
>
> as it may not only causes ctx->count to overflow, but also unnecessarily
> calls wake_up_locked_poll().
>
>
> I am sorry for just adding the following string in the patch:
> Fixes: cb289d6244a3 ("eventfd - allow atomic read and waitqueue remove")
>
>
> Looking forward to your suggestions.
>
> --
>
> Best wishes,
>
> Wen
>
>
> > I'm just slightly skeptical about patches that fix issues without an
> > analysis how this can happen.
> >
> > > fs/eventfd.c | 4 +++-
> > > 1 file changed, 3 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/fs/eventfd.c b/fs/eventfd.c
> > > index 8aa36cd37351..10a101df19cd 100644
> > > --- a/fs/eventfd.c
> > > +++ b/fs/eventfd.c
> > > @@ -189,7 +189,7 @@ void eventfd_ctx_do_read(struct eventfd_ctx *ctx, __u64 *cnt)
> > > {
> > > lockdep_assert_held(&ctx->wqh.lock);
> > > - *cnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
> > > + *cnt = ((ctx->flags & EFD_SEMAPHORE) && ctx->count) ? 1 : ctx->count;
> > > ctx->count -= *cnt;
> > > }
> > > EXPORT_SYMBOL_GPL(eventfd_ctx_do_read);
> > > @@ -269,6 +269,8 @@ static ssize_t eventfd_write(struct file *file, const char __user *buf, size_t c
> > > return -EFAULT;
> > > if (ucnt == ULLONG_MAX)
> > > return -EINVAL;
> > > + if ((ctx->flags & EFD_SEMAPHORE) && !ucnt)
> > > + return -EINVAL;
Hm, why is bit necessary though? What's wrong with specifying ucnt == 0
with EFD_SEMAPHORE? This also looks like a (very low potential) uapi
break.
Powered by blists - more mailing lists