Messages in this thread | | | Date | Tue, 03 Nov 2009 14:58:10 -0500 (EST) | From | Nicolas Pitre <> | Subject | Re: [PATCH 00/11] Sort out sdio_uart - stage two |
| |
On Tue, 3 Nov 2009, Alan Cox wrote:
> With the bugs Nicolas reported hopefully now nailed.
Progress, but still not there yet.
I'm still using the same "cat /dev/ttySDIO0" test with a GPS card.
Yanking the card out does terminate the cat process properly. Inserting the card back in and restarting cat always fails with:
cat: /dev/ttySDIO0: Input/output error
The same thing occurs if I terminate cat with ^C instead of pulling the card out.
From what I can see, sdio_uart_open() and sdio_uart_activate() are called on the first cat invocation, and then sdio_uart_close() and sdio_uart_shutdown() when cat is stopped with ^C. However neither sdio_uart_open() nor sdio_uart_activate() is ever called anymore with any subsequent cat invocations until a reboot. Some upper layer must have taken upon itself to return -EIO to user space.
Nicolas
| |