2014-02-20 09:22:45

by Jiang Liu

[permalink] [raw]
Subject: [PATCH] nouveau, ACPI: fix regression caused by b072e53

Fix regression caused by commit b072e53, which breaks loading nouveau
driver on optimus laptops.

On some platforms, ACPI _DSM method (nouveau_op_dsm_muid, function 0)
has special requirements on the fourth parameter, which is different
from ACPI specifications. So revert to the private implementation
to check availability of _DSM functions instead of using common
acpi_check_dsm() interface.

Reported-and-Tested-by: Maarten Lankhorst <[email protected]>
Signed-off-by: Jiang Liu <[email protected]>
---
drivers/gpu/drm/nouveau/nouveau_acpi.c | 26 ++++++++++++++++++++++++--
1 file changed, 24 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c
index 4ef83df..83face3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
@@ -106,6 +106,29 @@ static int nouveau_optimus_dsm(acpi_handle handle, int func, int arg, uint32_t *
return 0;
}

+/*
+ * On some platforms, _DSM(nouveau_op_dsm_muid, func0) has special
+ * requirements on the fourth parameter, so a private implementation
+ * instead of using acpi_check_dsm().
+ */
+static int nouveau_check_optimus_dsm(acpi_handle handle)
+{
+ int result;
+
+ /*
+ * Function 0 returns a Buffer containing available functions.
+ * The args parameter is ignored for function 0, so just put 0 in it
+ */
+ if (nouveau_optimus_dsm(handle, 0, 0, &result))
+ return 0;
+
+ /*
+ * ACPI Spec v4 9.14.1: if bit 0 is zero, no function is supported.
+ * If the n-th bit is enabled, function n is supported
+ */
+ return result & 1 && result & (1 << NOUVEAU_DSM_OPTIMUS_CAPS);
+}
+
static int nouveau_dsm(acpi_handle handle, int func, int arg)
{
int ret = 0;
@@ -207,8 +230,7 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev)
1 << NOUVEAU_DSM_POWER))
retval |= NOUVEAU_DSM_HAS_MUX;

- if (acpi_check_dsm(dhandle, nouveau_op_dsm_muid, 0x00000100,
- 1 << NOUVEAU_DSM_OPTIMUS_CAPS))
+ if (nouveau_check_optimus_dsm(dhandle))
retval |= NOUVEAU_DSM_HAS_OPT;

if (retval & NOUVEAU_DSM_HAS_OPT) {
--
1.7.10.4


2014-02-20 20:27:53

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: [PATCH] nouveau, ACPI: fix regression caused by b072e53

On 2/20/2014 10:23 AM, Jiang Liu wrote:
> Fix regression caused by commit b072e53, which breaks loading nouveau
> driver on optimus laptops.
>
> On some platforms, ACPI _DSM method (nouveau_op_dsm_muid, function 0)
> has special requirements on the fourth parameter, which is different
> from ACPI specifications. So revert to the private implementation
> to check availability of _DSM functions instead of using common
> acpi_check_dsm() interface.
>
> Reported-and-Tested-by: Maarten Lankhorst <[email protected]>
> Signed-off-by: Jiang Liu <[email protected]>

I'm taking this, because the commit that introduced the regression went
in through my tree.

In the future I'll appreciate CCing ACPI-related patches to linux-acpi,
however.

Thanks,
Rafael


> ---
> drivers/gpu/drm/nouveau/nouveau_acpi.c | 26 ++++++++++++++++++++++++--
> 1 file changed, 24 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c
> index 4ef83df..83face3 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
> @@ -106,6 +106,29 @@ static int nouveau_optimus_dsm(acpi_handle handle, int func, int arg, uint32_t *
> return 0;
> }
>
> +/*
> + * On some platforms, _DSM(nouveau_op_dsm_muid, func0) has special
> + * requirements on the fourth parameter, so a private implementation
> + * instead of using acpi_check_dsm().
> + */
> +static int nouveau_check_optimus_dsm(acpi_handle handle)
> +{
> + int result;
> +
> + /*
> + * Function 0 returns a Buffer containing available functions.
> + * The args parameter is ignored for function 0, so just put 0 in it
> + */
> + if (nouveau_optimus_dsm(handle, 0, 0, &result))
> + return 0;
> +
> + /*
> + * ACPI Spec v4 9.14.1: if bit 0 is zero, no function is supported.
> + * If the n-th bit is enabled, function n is supported
> + */
> + return result & 1 && result & (1 << NOUVEAU_DSM_OPTIMUS_CAPS);
> +}
> +
> static int nouveau_dsm(acpi_handle handle, int func, int arg)
> {
> int ret = 0;
> @@ -207,8 +230,7 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev)
> 1 << NOUVEAU_DSM_POWER))
> retval |= NOUVEAU_DSM_HAS_MUX;
>
> - if (acpi_check_dsm(dhandle, nouveau_op_dsm_muid, 0x00000100,
> - 1 << NOUVEAU_DSM_OPTIMUS_CAPS))
> + if (nouveau_check_optimus_dsm(dhandle))
> retval |= NOUVEAU_DSM_HAS_OPT;
>
> if (retval & NOUVEAU_DSM_HAS_OPT) {

2014-02-21 05:41:03

by Jiang Liu

[permalink] [raw]
Subject: Re: [PATCH] nouveau, ACPI: fix regression caused by b072e53

Thanks, Rafael.
Will cc ACPI maillist next time.

On 2014/2/21 4:27, Rafael J. Wysocki wrote:
> On 2/20/2014 10:23 AM, Jiang Liu wrote:
>> Fix regression caused by commit b072e53, which breaks loading nouveau
>> driver on optimus laptops.
>>
>> On some platforms, ACPI _DSM method (nouveau_op_dsm_muid, function 0)
>> has special requirements on the fourth parameter, which is different
>> from ACPI specifications. So revert to the private implementation
>> to check availability of _DSM functions instead of using common
>> acpi_check_dsm() interface.
>>
>> Reported-and-Tested-by: Maarten Lankhorst
>> <[email protected]>
>> Signed-off-by: Jiang Liu <[email protected]>
>
> I'm taking this, because the commit that introduced the regression went
> in through my tree.
>
> In the future I'll appreciate CCing ACPI-related patches to linux-acpi,
> however.
>
> Thanks,
> Rafael
>
>
>> ---
>> drivers/gpu/drm/nouveau/nouveau_acpi.c | 26
>> ++++++++++++++++++++++++--
>> 1 file changed, 24 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c
>> b/drivers/gpu/drm/nouveau/nouveau_acpi.c
>> index 4ef83df..83face3 100644
>> --- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
>> +++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
>> @@ -106,6 +106,29 @@ static int nouveau_optimus_dsm(acpi_handle
>> handle, int func, int arg, uint32_t *
>> return 0;
>> }
>> +/*
>> + * On some platforms, _DSM(nouveau_op_dsm_muid, func0) has special
>> + * requirements on the fourth parameter, so a private implementation
>> + * instead of using acpi_check_dsm().
>> + */
>> +static int nouveau_check_optimus_dsm(acpi_handle handle)
>> +{
>> + int result;
>> +
>> + /*
>> + * Function 0 returns a Buffer containing available functions.
>> + * The args parameter is ignored for function 0, so just put 0 in it
>> + */
>> + if (nouveau_optimus_dsm(handle, 0, 0, &result))
>> + return 0;
>> +
>> + /*
>> + * ACPI Spec v4 9.14.1: if bit 0 is zero, no function is supported.
>> + * If the n-th bit is enabled, function n is supported
>> + */
>> + return result & 1 && result & (1 << NOUVEAU_DSM_OPTIMUS_CAPS);
>> +}
>> +
>> static int nouveau_dsm(acpi_handle handle, int func, int arg)
>> {
>> int ret = 0;
>> @@ -207,8 +230,7 @@ static int nouveau_dsm_pci_probe(struct pci_dev
>> *pdev)
>> 1 << NOUVEAU_DSM_POWER))
>> retval |= NOUVEAU_DSM_HAS_MUX;
>> - if (acpi_check_dsm(dhandle, nouveau_op_dsm_muid, 0x00000100,
>> - 1 << NOUVEAU_DSM_OPTIMUS_CAPS))
>> + if (nouveau_check_optimus_dsm(dhandle))
>> retval |= NOUVEAU_DSM_HAS_OPT;
>> if (retval & NOUVEAU_DSM_HAS_OPT) {
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

2014-02-21 06:04:15

by Dave Airlie

[permalink] [raw]
Subject: Re: [PATCH] nouveau, ACPI: fix regression caused by b072e53

On Fri, Feb 21, 2014 at 6:27 AM, Rafael J. Wysocki
<[email protected]> wrote:
> On 2/20/2014 10:23 AM, Jiang Liu wrote:
>>
>> Fix regression caused by commit b072e53, which breaks loading nouveau
>> driver on optimus laptops.
>>
>> On some platforms, ACPI _DSM method (nouveau_op_dsm_muid, function 0)
>> has special requirements on the fourth parameter, which is different
>> from ACPI specifications. So revert to the private implementation
>> to check availability of _DSM functions instead of using common
>> acpi_check_dsm() interface.
>>
>> Reported-and-Tested-by: Maarten Lankhorst
>> <[email protected]>
>> Signed-off-by: Jiang Liu <[email protected]>
>
>
> I'm taking this, because the commit that introduced the regression went in
> through my tree.
>
> In the future I'll appreciate CCing ACPI-related patches to linux-acpi,
> however.

Thanks,

Acked-by: Dave Airlie <[email protected]>

Dave.