[PATCH v4 0/2] acpi: apei: Bug fix to enable APEI support for ARMv8

[PATCH v4 0/2] acpi: apei: Bug fix to enable APEI support for ARMv8

[Loc Ho]

v4:
* Update patch commit log for gsi.c patch
* Make change to be 80 column aligned

v3
* Change the title of the patch to reflect the patch
* Completely dropped the driver. Instead, move completely to the FW by
  accessing the GIC directly.
* Fix a bug with function acpi_gsi_to_irq
* Enable APEI multiple GHES source to share an single external IRQ

v2
* Make all code more generic naming
* Still waiting for comment from Linaro folks on APEI

---
Loc Ho (2):
  acpi: Fix proper return code for function acpi_gsi_to_irq
  acpi: apei: Enable APEI multiple GHES source to share an single
    external IRQ

 drivers/acpi/apei/ghes.c |    3 ++-
 drivers/acpi/gsi.c       |    4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)

[PATCH v4 1/2] acpi: Fix proper return code for function acpi_gsi_to_irq

[Loc Ho]

The function acpi_gsi_to_irq must return 0 on success as the caller
ghes_probe expects an 0 for success. This change also matches x86
implementation.

Signed-off-by: Tuan Phan <tphan@apm.com>
Signed-off-by: Loc Ho <lho@apm.com>
---
 drivers/acpi/gsi.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/acpi/gsi.c b/drivers/acpi/gsi.c
index ee9e0f2..7b109bd 100644
--- a/drivers/acpi/gsi.c
+++ b/drivers/acpi/gsi.c
@@ -24,7 +24,7 @@ static struct fwnode_handle *acpi_gsi_domain_id;
  *
  * irq location updated with irq value [>0 on success, 0 on failure]
  *
- * Returns: linux IRQ number on success (>0)
+ * Returns: 0 on success
  *          -EINVAL on failure
  */
 int acpi_gsi_to_irq(u32 gsi, unsigned int *irq)
@@ -37,7 +37,7 @@ int acpi_gsi_to_irq(u32 gsi, unsigned int *irq)
 	 * *irq == 0 means no mapping, that should
 	 * be reported as a failure
 	 */
-	return (*irq > 0) ? *irq : -EINVAL;
+	return (*irq > 0) ? 0 : -EINVAL;
 }
 EXPORT_SYMBOL_GPL(acpi_gsi_to_irq);
 
-- 
1.7.1

[PATCH v4 1/2] acpi: Fix proper return code for function acpi_gsi_to_irq

[Loc Ho]

Hi,

On Wed, Jan 20, 2016 at 4:17 PM, Loc Ho <lho@apm.com> wrote:
>
> The function acpi_gsi_to_irq must return 0 on success as the caller
> ghes_probe expects an 0 for success. This change also matches x86
> implementation.
>
> Signed-off-by: Tuan Phan <tphan@apm.com>
> Signed-off-by: Loc Ho <lho@apm.com>


Any issue with this version? If not, can this be pulled into 4.5?

-Loc

[PATCH v4 1/2] acpi: Fix proper return code for function acpi_gsi_to_irq

[Marc Zyngier]

Loc,

On 25/01/16 21:23, Loc Ho wrote:
> Hi,
> 
> On Wed, Jan 20, 2016 at 4:17 PM, Loc Ho <lho@apm.com> wrote:
>>
>> The function acpi_gsi_to_irq must return 0 on success as the caller
>> ghes_probe expects an 0 for success. This change also matches x86
>> implementation.
>>
>> Signed-off-by: Tuan Phan <tphan@apm.com>
>> Signed-off-by: Loc Ho <lho@apm.com>
> 
> 
> Any issue with this version? If not, can this be pulled into 4.5?

How is that 4.5 material? Does it fix an observable bug or a regression
from a previous? If not, then this is probably to be postponed until
4.6, together with the code that makes use of it.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

[PATCH v4 1/2] acpi: Fix proper return code for function acpi_gsi_to_irq

[Loc Ho]

Hi Marc,

On Tue, Jan 26, 2016 at 12:31 AM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> Loc,
>
> On 25/01/16 21:23, Loc Ho wrote:
>> Hi,
>>
>> On Wed, Jan 20, 2016 at 4:17 PM, Loc Ho <lho@apm.com> wrote:
>>>
>>> The function acpi_gsi_to_irq must return 0 on success as the caller
>>> ghes_probe expects an 0 for success. This change also matches x86
>>> implementation.
>>>
>>> Signed-off-by: Tuan Phan <tphan@apm.com>
>>> Signed-off-by: Loc Ho <lho@apm.com>
>>
>>
>> Any issue with this version? If not, can this be pulled into 4.5?
>
> How is that 4.5 material? Does it fix an observable bug or a regression
> from a previous? If not, then this is probably to be postponed until
> 4.6, together with the code that makes use of it.

Given that this and the other patch are needed for APEI to work with
X-Gene, fine with 4.6. Are we just waiting for Rafael to make the
final call then?

-Loc

[PATCH v4 2/2] acpi: apei: Enable APEI multiple GHES source to share an single external IRQ

[Loc Ho]

This patch allows APEI generic error source table with external
IRQ to share an single IRQ.

Signed-off-by: Tuan Phan <tphan@apm.com>
Signed-off-by: Loc Ho <lho@apm.com>
---
 drivers/acpi/apei/ghes.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
index 3dd9c46..76146c8 100644
--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -1011,7 +1011,8 @@ static int ghes_probe(struct platform_device *ghes_dev)
 			       generic->header.source_id);
 			goto err_edac_unreg;
 		}
-		rc = request_irq(ghes->irq, ghes_irq_func, 0, "GHES IRQ", ghes);
+		rc = request_irq(ghes->irq, ghes_irq_func, IRQF_SHARED,
+				 "GHES IRQ", ghes);
 		if (rc) {
 			pr_err(GHES_PFX "Failed to register IRQ for generic hardware error source: %d\n",
 			       generic->header.source_id);
-- 
1.7.1

[PATCH v4 2/2] acpi: apei: Enable APEI multiple GHES source to share an single external IRQ

[Loc Ho]

Hi Borislav,

On Wed, Jan 20, 2016 at 4:17 PM, Loc Ho <lho@apm.com> wrote:
> This patch allows APEI generic error source table with external
> IRQ to share an single IRQ.
>
> Signed-off-by: Tuan Phan <tphan@apm.com>
> Signed-off-by: Loc Ho <lho@apm.com>
> ---
>  drivers/acpi/apei/ghes.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
> index 3dd9c46..76146c8 100644
> --- a/drivers/acpi/apei/ghes.c
> +++ b/drivers/acpi/apei/ghes.c
> @@ -1011,7 +1011,8 @@ static int ghes_probe(struct platform_device *ghes_dev)
>                                generic->header.source_id);
>                         goto err_edac_unreg;
>                 }
> -               rc = request_irq(ghes->irq, ghes_irq_func, 0, "GHES IRQ", ghes);
> +               rc = request_irq(ghes->irq, ghes_irq_func, IRQF_SHARED,
> +                                "GHES IRQ", ghes);
>                 if (rc) {
>                         pr_err(GHES_PFX "Failed to register IRQ for generic hardware error source: %d\n",
>                                generic->header.source_id);

Can this be pull into 4.5-rcX given that it was already ack'ed week back?

-Loc

[PATCH v4 2/2] acpi: apei: Enable APEI multiple GHES source to share an single external IRQ

[Borislav Petkov]

On Mon, Jan 25, 2016 at 01:25:00PM -0800, Loc Ho wrote:
> Can this be pull into 4.5-rcX given that it was already ack'ed week back?

It is still Rafael's decision, as I've already stated.

Rafael?

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.