Re: [PATCH 2/3 v2] xhci: Fix race related to abort operation

2016-11-16 Thread OGAWA Hirofumi
Mathias Nyman writes: > The retry was a workaround for a case triggered on v4.1.4 by > Vincent Pelletier (added to CC) > >http://marc.info/?l=linux-usb=144031185222108=2 > > The race could very well explain that issue. >From logs of that thread, it would has

Re: [PATCH 2/3 v2] xhci: Fix race related to abort operation

2016-11-16 Thread Mathias Nyman
On 16.11.2016 11:50, OGAWA Hirofumi wrote: Current abort operation has race. xhci_handle_command_timeout() xhci_abort_cmd_ring() xhci_write_64(CMD_RING_ABORT) xhci_handshake(5s) do { check CMD_RING_RUNNING udelay(1)

[PATCH 2/3 v2] xhci: Fix race related to abort operation

2016-11-16 Thread OGAWA Hirofumi
Current abort operation has race. xhci_handle_command_timeout() xhci_abort_cmd_ring() xhci_write_64(CMD_RING_ABORT) xhci_handshake(5s) do { check CMD_RING_RUNNING udelay(1) ...