Re: [Resend PATCH 3/3] scsi:stex.c Add S3/S4 support

2016-02-21 Thread Charles Chiou 

Hi Oliver,
sure, we'll fix it at the next patch version. Thank you.
Charles

On 02/04/2016 07:37 PM, Oliver Neukum wrote:

On Thu, 2016-02-04 at 19:22 +0800, Charles Chiou wrote:

+static int stex_choice_sleep_mic(pm_message_t state)
+{
+switch (state.event) {
+case PM_EVENT_SUSPEND:
+return ST_S3;
+case PM_EVENT_FREEZE:


Why do you react to PM_EVENT_FREEZE at all?
That is too early. You will get a HIBERNATE
event anyway. If the write out fails you are
in trouble if you already reacted to
PM_EVENT_FREEZE

Regards
Oliver

Re: [Resend PATCH 3/3] scsi:stex.c Add S3/S4 support

2016-02-21 Thread Charles Chiou 

Hi Oliver,
sure, we'll fix it at the next patch version. Thank you.
Charles

On 02/04/2016 07:37 PM, Oliver Neukum wrote:

On Thu, 2016-02-04 at 19:22 +0800, Charles Chiou wrote:

+static int stex_choice_sleep_mic(pm_message_t state)
+{
+switch (state.event) {
+case PM_EVENT_SUSPEND:
+return ST_S3;
+case PM_EVENT_FREEZE:


Why do you react to PM_EVENT_FREEZE at all?
That is too early. You will get a HIBERNATE
event anyway. If the write out fails you are
in trouble if you already reacted to
PM_EVENT_FREEZE

Regards
Oliver

Re: [Resend PATCH 3/3] scsi:stex.c Add S3/S4 support

2016-02-04 Thread Oliver Neukum 
On Thu, 2016-02-04 at 19:22 +0800, Charles Chiou wrote:
> +static int stex_choice_sleep_mic(pm_message_t state)
> +{
> +switch (state.event) {
> +case PM_EVENT_SUSPEND:
> +return ST_S3;
> +case PM_EVENT_FREEZE:

Why do you react to PM_EVENT_FREEZE at all?
That is too early. You will get a HIBERNATE
event anyway. If the write out fails you are
in trouble if you already reacted to 
PM_EVENT_FREEZE

Regards
Oliver

Re: [Resend PATCH 3/3] scsi:stex.c Add S3/S4 support

2016-02-04 Thread Oliver Neukum 
On Thu, 2016-02-04 at 19:22 +0800, Charles Chiou wrote:
> +static int stex_choice_sleep_mic(pm_message_t state)
> +{
> +switch (state.event) {
> +case PM_EVENT_SUSPEND:
> +return ST_S3;
> +case PM_EVENT_FREEZE:

Why do you react to PM_EVENT_FREEZE at all?
That is too early. You will get a HIBERNATE
event anyway. If the write out fails you are
in trouble if you already reacted to 
PM_EVENT_FREEZE

Regards
Oliver