On 5/8/20 9:29 AM, Hillf Danton wrote:
> Dunno if what's missing makes grumpy.
>
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -3439,6 +3439,11 @@ static void io_close_finish(struct io_wq
> static int io_close(struct io_kiocb *req, bool force_nonblock)
> {
> int ret;
> + struct fd
On 5/7/20 8:28 PM, Jens Axboe wrote:
> On 5/7/20 5:31 PM, Al Viro wrote:
>> On Thu, May 07, 2020 at 05:03:17PM -0600, Jens Axboe wrote:
>>> On 5/7/20 4:44 PM, Al Viro wrote:
On Thu, May 07, 2020 at 04:25:24PM -0600, Jens Axboe wrote:
> static int io_close(struct io_kiocb *req, bool
On 5/7/20 5:31 PM, Al Viro wrote:
> On Thu, May 07, 2020 at 05:03:17PM -0600, Jens Axboe wrote:
>> On 5/7/20 4:44 PM, Al Viro wrote:
>>> On Thu, May 07, 2020 at 04:25:24PM -0600, Jens Axboe wrote:
>>>
static int io_close(struct io_kiocb *req, bool force_nonblock)
{
+ struct
On Thu, May 07, 2020 at 05:03:17PM -0600, Jens Axboe wrote:
> On 5/7/20 4:44 PM, Al Viro wrote:
> > On Thu, May 07, 2020 at 04:25:24PM -0600, Jens Axboe wrote:
> >
> >> static int io_close(struct io_kiocb *req, bool force_nonblock)
> >> {
> >> + struct files_struct *files = current->files;
>
On 5/7/20 4:44 PM, Al Viro wrote:
> On Thu, May 07, 2020 at 04:25:24PM -0600, Jens Axboe wrote:
>
>> static int io_close(struct io_kiocb *req, bool force_nonblock)
>> {
>> +struct files_struct *files = current->files;
>> int ret;
>>
>> req->close.put_file = NULL;
>> -ret =
On Thu, May 07, 2020 at 04:25:24PM -0600, Jens Axboe wrote:
> static int io_close(struct io_kiocb *req, bool force_nonblock)
> {
> + struct files_struct *files = current->files;
> int ret;
>
> req->close.put_file = NULL;
> - ret = __close_fd_get_file(req->close.fd,
On 5/7/20 4:06 PM, Al Viro wrote:
> On Thu, May 07, 2020 at 02:53:30PM -0600, Jens Axboe wrote:
>
>> I think the patch is correct as-is, I took a good look at how we're
>> currently handling it. None of those three ops should fiddle with
>> the fd at all, and all of them do forbid the use of
On Thu, May 07, 2020 at 02:53:30PM -0600, Jens Axboe wrote:
> I think the patch is correct as-is, I took a good look at how we're
> currently handling it. None of those three ops should fiddle with
> the fd at all, and all of them do forbid the use of fixed files (the
> descriptor table
On 5/7/20 1:29 PM, Al Viro wrote:
> On Thu, May 07, 2020 at 01:05:23PM -0600, Jens Axboe wrote:
>> On 5/7/20 1:01 PM, Al Viro wrote:
>>> On Thu, May 07, 2020 at 08:57:25PM +0200, Max Kellermann wrote:
If an operation's flag `needs_file` is set, the function
io_req_set_file() calls
On 2020/05/07 21:29, Al Viro wrote:
> Again, resolving the descriptor more than once in course of syscall
> is almost always a serious bug;
.. and that is what Linux currently does for those three operation,
and yes, it's buggy. The generic preparation code looks up the fd,
but later in the
On Thu, May 07, 2020 at 01:05:23PM -0600, Jens Axboe wrote:
> On 5/7/20 1:01 PM, Al Viro wrote:
> > On Thu, May 07, 2020 at 08:57:25PM +0200, Max Kellermann wrote:
> >> If an operation's flag `needs_file` is set, the function
> >> io_req_set_file() calls io_file_get() to obtain a `struct file*`.
>
On 2020/05/07 21:05, Jens Axboe wrote:
> On 5/7/20 1:01 PM, Al Viro wrote:
> > On Thu, May 07, 2020 at 08:57:25PM +0200, Max Kellermann wrote:
> >> If an operation's flag `needs_file` is set, the function
> >> io_req_set_file() calls io_file_get() to obtain a `struct file*`.
> >>
> >> This fails
On 5/7/20 1:01 PM, Al Viro wrote:
> On Thu, May 07, 2020 at 08:57:25PM +0200, Max Kellermann wrote:
>> If an operation's flag `needs_file` is set, the function
>> io_req_set_file() calls io_file_get() to obtain a `struct file*`.
>>
>> This fails for `O_PATH` file descriptors, because those have no
On 2020/05/07 21:01, Al Viro wrote:
> On Thu, May 07, 2020 at 08:57:25PM +0200, Max Kellermann wrote:
> > If an operation's flag `needs_file` is set, the function
> > io_req_set_file() calls io_file_get() to obtain a `struct file*`.
> >
> > This fails for `O_PATH` file descriptors, because those
On Thu, May 07, 2020 at 08:57:25PM +0200, Max Kellermann wrote:
> If an operation's flag `needs_file` is set, the function
> io_req_set_file() calls io_file_get() to obtain a `struct file*`.
>
> This fails for `O_PATH` file descriptors, because those have no
> `struct file*`
O_PATH descriptors
On 2020/05/07 20:58, Jens Axboe wrote:
> Do you happen to have a liburing test addition for this as well?
No, I'll write one tomorrow. GitHub PR or email preferred?
Max
On 5/7/20 12:57 PM, Max Kellermann wrote:
> If an operation's flag `needs_file` is set, the function
> io_req_set_file() calls io_file_get() to obtain a `struct file*`.
>
> This fails for `O_PATH` file descriptors, because those have no
> `struct file*`, causing io_req_set_file() to throw
If an operation's flag `needs_file` is set, the function
io_req_set_file() calls io_file_get() to obtain a `struct file*`.
This fails for `O_PATH` file descriptors, because those have no
`struct file*`, causing io_req_set_file() to throw `-EBADF`. This
breaks the operations `openat`, `openat2`
18 matches
Mail list logo
