Hello Gedare,
thanks for the review.
On 17/07/2020 07:43, Gedare Bloom wrote:
> On Thu, Jul 9, 2020 at 2:21 AM Christian Mauderer
> <christian.maude...@embedded-brains.de> wrote:
>>
>> This adds the possibility to request a retry in the media-listener if an
>> operation failed. Usefull for example if you want to automatically
>> reformat a disk if it wasn't possible to mount it.
>> ---
>> cpukit/include/rtems/media.h | 3 +++
>> cpukit/libblock/src/media.c | 29 ++++++++++++++++-------------
>> 2 files changed, 19 insertions(+), 13 deletions(-)
>>
>> diff --git a/cpukit/include/rtems/media.h b/cpukit/include/rtems/media.h
>> index b2a3e2dc91..db13664975 100644
>> --- a/cpukit/include/rtems/media.h
>> +++ b/cpukit/include/rtems/media.h
>> @@ -281,6 +281,9 @@ typedef enum {
>> *
>> * @retval RTEMS_SUCCESSFUL Successful operation.
>> * @retval RTEMS_IO_ERROR In the inquiry state this will abort the action.
>> + * @retval RTEMS_INCORRECT_STATE In the failed state this will cause a
>> retry.
>> + * Make sure to have a retry counter or similar to avoid endless loops if
>> you
>> + * use this this value.
> delete one 'this'
>
> Glad you have this comment, that was my first thought when I read the
> commit msg.
>
>> */
>> typedef rtems_status_code (*rtems_media_listener)(
>> rtems_media_event event,
>> diff --git a/cpukit/libblock/src/media.c b/cpukit/libblock/src/media.c
>> index 5b2b06b5b2..c00825587c 100644
>> --- a/cpukit/libblock/src/media.c
>> +++ b/cpukit/libblock/src/media.c
>> @@ -420,26 +420,29 @@ static rtems_status_code process_event(
>> )
>> {
>> rtems_status_code sc = RTEMS_SUCCESSFUL;
>> + rtems_status_code sc_retry = RTEMS_SUCCESSFUL;
>> rtems_media_state state = RTEMS_MEDIA_STATE_FAILED;
>> char *dest = NULL;
>>
>> - sc = notify(event, RTEMS_MEDIA_STATE_INQUIRY, src, NULL);
>> - if (sc == RTEMS_SUCCESSFUL) {
>> - state = RTEMS_MEDIA_STATE_READY;
>> - } else {
>> - state = RTEMS_MEDIA_STATE_ABORTED;
>> - }
>> -
>> - sc = (*worker)(state, src, &dest, worker_arg);
>> - if (state == RTEMS_MEDIA_STATE_READY) {
>> + do {
>> + sc = notify(event, RTEMS_MEDIA_STATE_INQUIRY, src, NULL);
>> if (sc == RTEMS_SUCCESSFUL) {
>> - state = RTEMS_MEDIA_STATE_SUCCESS;
>> + state = RTEMS_MEDIA_STATE_READY;
>> } else {
>> - state = RTEMS_MEDIA_STATE_FAILED;
>> + state = RTEMS_MEDIA_STATE_ABORTED;
>> + }
>> +
>> + sc = (*worker)(state, src, &dest, worker_arg);
>> + if (state == RTEMS_MEDIA_STATE_READY) {
>> + if (sc == RTEMS_SUCCESSFUL) {
>> + state = RTEMS_MEDIA_STATE_SUCCESS;
>> + } else {
>> + state = RTEMS_MEDIA_STATE_FAILED;
>> + }
>> }
>> - }
>>
>> - notify(event, state, src, dest);
>> + sc_retry = notify(event, state, src, dest);
>> + } while (state == RTEMS_MEDIA_STATE_FAILED && sc_retry ==
>> RTEMS_INCORRECT_STATE);
> line length
>
> How do you use a retry counter? It's not obvious to me, but this is
> not an area I work with much.
>
The retry is only done if explicitly requested by the media listener.
The media listener is a software provided by the user. If the user
requests a retry he should also decide how often he wants to retry the
operation.
A possibility to add a retry counter is to use the user specifiable
argument of the media listener. The user can provide a pointer to some
own data structure where the retries are counted.
I think the retries in general are a quite exotic feature. Like my
comment suggests I used it to try a mount for an on-board memory, if it
doesn't work I format the disk and retry the mount.
Best regards
Christian
>> remember_event(event, state, src, dest);
>>
>> if (state == RTEMS_MEDIA_STATE_SUCCESS) {
>> --
>> 2.26.2
>>
>> _______________________________________________
>> devel mailing list
>> devel@rtems.org
>> http://lists.rtems.org/mailman/listinfo/devel
--
--------------------------------------------
embedded brains GmbH
Herr Christian Mauderer
Dornierstr. 4
D-82178 Puchheim
Germany
email: christian.maude...@embedded-brains.de
Phone: +49-89-18 94 741 - 18
Fax: +49-89-18 94 741 - 08
PGP: Public key available on request.
Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.
_______________________________________________
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel
