Andrew Patterson wrote:
> On Sat, 2008-11-01 at 15:55 +0900, Taku Izumi wrote:
>> Revert adf411b819adc9fa96e9b3e638c7480d5e71d270.
>>
>> The commit adf411b819adc9fa96e9b3e638c7480d5e71d270 was based on the
>> improper assumption that queried result was not updated when _OSC
>> support field was changed. But, in fact, queried result is updated
>> whenever _OSC support field was changed through __acpi_query_osc().
>> As a result, the commit adf411b819adc9fa96e9b3e638c7480d5e71d270 only
>> introduced unnecessary additional _OSC evaluation...
>>
>> Signed-off-by: Kenji Kaneshige <kaneshige.kenji@xxxxxxxxxxxxxx>
>> Signed-off-by: Taku Izumi <izumi.taku@xxxxxxxxxxxxxx>
>>
>> ---
>> drivers/pci/pci-acpi.c | 31 ++++++++++++++++++-------------
>> 1 file changed, 18 insertions(+), 13 deletions(-)
>>
>> Index: 20081031/drivers/pci/pci-acpi.c
>> ===================================================================
>> --- 20081031.orig/drivers/pci/pci-acpi.c
>> +++ 20081031/drivers/pci/pci-acpi.c
>> @@ -24,13 +24,15 @@ struct acpi_osc_data {
>> acpi_handle handle;
>> u32 support_set;
>> u32 control_set;
>> + int is_queried;
>> + u32 query_result;
>> struct list_head sibiling;
>> };
>> static LIST_HEAD(acpi_osc_data_list);
>>
>> struct acpi_osc_args {
>> u32 capbuf[3];
>> - u32 ctrl_result;
>> + u32 query_result;
>> };
>>
>> static DEFINE_MUTEX(pci_acpi_lock);
>> @@ -108,8 +110,9 @@ static acpi_status acpi_run_osc(acpi_han
>> goto out_kfree;
>> }
>> out_success:
>> - osc_args->ctrl_result =
>> - *((u32 *)(out_obj->buffer.pointer + 8));
>> + if (flags & OSC_QUERY_ENABLE)
>> + osc_args->query_result =
>> + *((u32 *)(out_obj->buffer.pointer + 8));
>> status = AE_OK;
>>
>> out_kfree:
>> @@ -117,8 +120,7 @@ out_kfree:
>> return status;
>> }
>>
>> -static acpi_status __acpi_query_osc(u32 flags, struct acpi_osc_data *osc_data,
>> - u32 *result)
>> +static acpi_status __acpi_query_osc(u32 flags, struct acpi_osc_data *osc_data)
>> {
>> acpi_status status;
>> u32 support_set;
>> @@ -133,7 +135,8 @@ static acpi_status __acpi_query_osc(u32
>> status = acpi_run_osc(osc_data->handle, &osc_args);
>> if (ACPI_SUCCESS(status)) {
>> osc_data->support_set = support_set;
>> - *result = osc_args.ctrl_result;
>> + osc_data->query_result = osc_args.query_result;
>> + osc_data->is_queried = 1;
>> }
>>
>> return status;
>> @@ -144,7 +147,7 @@ static acpi_status acpi_query_osc(acpi_h
>> {
>> acpi_status status;
>> struct acpi_osc_data *osc_data;
>> - u32 flags = (unsigned long)context, dummy;
>> + u32 flags = (unsigned long)context;
>> acpi_handle tmp;
>>
>> status = acpi_get_handle(handle, "_OSC", &tmp);
>> @@ -158,7 +161,7 @@ static acpi_status acpi_query_osc(acpi_h
>> goto out;
>> }
>>
>> - __acpi_query_osc(flags, osc_data, &dummy);
>> + __acpi_query_osc(flags, osc_data);
>> out:
>> mutex_unlock(&pci_acpi_lock);
>> return AE_OK;
>> @@ -192,7 +195,7 @@ acpi_status __pci_osc_support_set(u32 fl
>> acpi_status pci_osc_control_set(acpi_handle handle, u32 flags)
>> {
>> acpi_status status;
>> - u32 ctrlset, control_set, result;
>> + u32 ctrlset, control_set;
>> acpi_handle tmp;
>> struct acpi_osc_data *osc_data;
>> struct acpi_osc_args osc_args;
>> @@ -215,11 +218,13 @@ acpi_status pci_osc_control_set(acpi_han
>> goto out;
>> }
>>
>> - status = __acpi_query_osc(osc_data->support_set, osc_data, &result);
>> - if (ACPI_FAILURE(status))
>> - goto out;
>> + if (!osc_data->is_queried) {
>> + status = __acpi_query_osc(osc_data->support_set, osc_data);
>> + if (ACPI_FAILURE(status))
>> + goto out;
>> + }
>>
>> - if ((result & ctrlset) != ctrlset) {
>> + if ((osc_data->query_result & ctrlset) != ctrlset) {
>> status = AE_SUPPORT;
>> goto out;
>> }
>>
>
> The latest code in the pci-2.6/linux-next branch runs __acpi_query_osc()
> in acpi_pci_root_add(). This happens before any calls to
> pci_osc_control_set(), so this patch/revert should no longer be
> necessary.
>
Though I might be misunderstanding, I think you mean that we should
modify pci_osc_control_set() not to call __acpi_query_osc() in any
case rather than applying this patch (reverting the commit
adf411b819adc9fa96e9b3e638c7480d5e71d270). Correct?
I think we should apply this patch first, and then making another
patch to modify pci_osc_control_set() not to call __acpi_query_osc()
because current pci_acpi_osc_support() can fail with -ENOMEM. In
this case, pci_osc_control_set() must query control. Maybe we need to
modify some data structures (acpi_osc_data, acpi_pci_root, etc.) to
prevent -ENOMEM when we modify pci_osc_control_set() not to call
__acpi_query_osc().
Thanks,
Kenji Kaneshige
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
