Hi Takashi,
>>> The write handler allocates skbs and queues them into data->readq.
>>> Read side should read them, if there is any. If there is none, skbs
>>> should be dropped by hdev->flush. But this happens only if the device
>>> is HCI_UP, i.e. hdev->power_on work was triggered already. When it was
>>> not, skbs stay allocated in the queue when /dev/vhci is closed. So
>>> purge the queue in ->release.
>>>
>>> Program to reproduce:
>>> #include <err.h>
>>> #include <fcntl.h>
>>> #include <stdio.h>
>>> #include <unistd.h>
>>>
>>> #include <sys/stat.h>
>>> #include <sys/types.h>
>>> #include <sys/uio.h>
>>>
>>> int main()
>>> {
>>> char buf[] = { 0xff, 0 };
>>> struct iovec iov = {
>>> .iov_base = buf,
>>> .iov_len = sizeof(buf),
>>> };
>>> int fd;
>>>
>>> while (1) {
>>> fd = open("/dev/vhci", O_RDWR);
>>> if (fd < 0)
>>> err(1, "open");
>>>
>>> usleep(50);
>>>
>>> if (writev(fd, &iov, 1) < 0)
>>> err(1, "writev");
>>>
>>> usleep(50);
>>>
>>> close(fd);
>>> }
>>>
>>> return 0;
>>> }
>>>
>>> Result:
>>> kmemleak: 4609 new suspected memory leaks
>>> unreferenced object 0xffff88059f4d5440 (size 232):
>>> comm "vhci", pid 1084, jiffies 4294912542 (age 37569.296s)
>>> hex dump (first 32 bytes):
>>> 20 f0 23 87 05 88 ff ff 20 f0 23 87 05 88 ff ff .#..... .#.....
>>> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
>>> backtrace:
>>> ...
>>> [<ffffffff81ece010>] __alloc_skb+0x0/0x5a0
>>> [<ffffffffa021886c>] vhci_create_device+0x5c/0x580 [hci_vhci]
>>> [<ffffffffa0219436>] vhci_write+0x306/0x4c8 [hci_vhci]
>>>
>>> Fixes: 23424c0d31 (Bluetooth: Add support creating virtual AMP controllers)
>>> Signed-off-by: Jiri Slaby <[email protected]>
>>> Cc: Marcel Holtmann <[email protected]>
>>> Cc: Gustavo Padovan <[email protected]>
>>> Cc: Johan Hedberg <[email protected]>
>>> Cc: <[email protected]>
>>> Cc: stable 3.13+ <[email protected]>
>>> ---
>>> drivers/bluetooth/hci_vhci.c | 1 +
>>> 1 file changed, 1 insertion(+)
>>
>> patch has been applied to bluetooth-next tree.
>
> Could you check another race of vhci mentioned in the thread?
> https://www.mail-archive.com/[email protected]/msg1106134.html
>
> If the suggested fix is OK, I'll submit a properly cooked patch.
lets get the patch send to the mailing list and then I can have a look at it. I
might need to test the final patch so that it works with legacy userspace.
Regards
Marcel
