[PATCH] drm/radeon/kms: fix fans after resume

[PATCH] drm/radeon/kms: fix fans after resume

[alexdeucher]

From: Alex Deucher <alexander.deucher@amd.com>

On pre-R600 asics, the SpeedFanControl table is not
executed as part of ASIC_Init as it is on newer asics.

Fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=29412

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/radeon/atom.c |   15 ++++++++++++++-
 drivers/gpu/drm/radeon/atom.h |    1 +
 2 files changed, 15 insertions(+), 1 deletions(-)

diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index d1bd239..5ce9bf5 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -1306,8 +1306,11 @@ struct atom_context *atom_parse(struct card_info *card, void *bios)
 
 int atom_asic_init(struct atom_context *ctx)
 {
+	struct radeon_device *rdev = ctx->card->dev->dev_private;
 	int hwi = CU16(ctx->data_table + ATOM_DATA_FWI_PTR);
 	uint32_t ps[16];
+	int ret;
+
 	memset(ps, 0, 64);
 
 	ps[0] = cpu_to_le32(CU32(hwi + ATOM_FWI_DEFSCLK_PTR));
@@ -1317,7 +1320,17 @@ int atom_asic_init(struct atom_context *ctx)
 
 	if (!CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_INIT))
 		return 1;
-	return atom_execute_table(ctx, ATOM_CMD_INIT, ps);
+	ret = atom_execute_table(ctx, ATOM_CMD_INIT, ps);
+	if (ret)
+		return ret;
+
+	memset(ps, 0, 64);
+
+	if (rdev->family < CHIP_R600) {
+		if (CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_SPDFANCNTL))
+			atom_execute_table(ctx, ATOM_CMD_SPDFANCNTL, ps);
+	}
+	return ret;
 }
 
 void atom_destroy(struct atom_context *ctx)
diff --git a/drivers/gpu/drm/radeon/atom.h b/drivers/gpu/drm/radeon/atom.h
index 93cfe20..25fea63 100644
--- a/drivers/gpu/drm/radeon/atom.h
+++ b/drivers/gpu/drm/radeon/atom.h
@@ -44,6 +44,7 @@
 #define ATOM_CMD_SETSCLK	0x0A
 #define ATOM_CMD_SETMCLK	0x0B
 #define ATOM_CMD_SETPCLK	0x0C
+#define ATOM_CMD_SPDFANCNTL	0x39
 
 #define ATOM_DATA_FWI_PTR	0xC
 #define ATOM_DATA_IIO_PTR	0x32
-- 
1.7.7.5

Re: [PATCH] drm/radeon/kms: fix fans after resume

[Michel Dänzer]

On Don, 2012-03-29 at 19:04 -0400, alexdeucher@gmail.com wrote: 
> From: Alex Deucher <alexander.deucher@amd.com>
> 
> On pre-R600 asics, the SpeedFanControl table is not
> executed as part of ASIC_Init as it is on newer asics.
> 
> Fixes:
> https://bugzilla.kernel.org/show_bug.cgi?id=29412
> 
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> Cc: stable@vger.kernel.org
> ---
>  drivers/gpu/drm/radeon/atom.c |   15 ++++++++++++++-
>  drivers/gpu/drm/radeon/atom.h |    1 +
>  2 files changed, 15 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
> index d1bd239..5ce9bf5 100644
> --- a/drivers/gpu/drm/radeon/atom.c
> +++ b/drivers/gpu/drm/radeon/atom.c
> [...]
> @@ -1317,7 +1320,17 @@ int atom_asic_init(struct atom_context *ctx)
>  
>  	if (!CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_INIT))
>  		return 1;
> -	return atom_execute_table(ctx, ATOM_CMD_INIT, ps);
> +	ret = atom_execute_table(ctx, ATOM_CMD_INIT, ps);
> +	if (ret)
> +		return ret;
> +
> +	memset(ps, 0, 64);
> +
> +	if (rdev->family < CHIP_R600) {
> +		if (CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_SPDFANCNTL))
> +			atom_execute_table(ctx, ATOM_CMD_SPDFANCNTL, ps);

Should the return value of atom_execute_table() be assigned to ret here
as well? Either way,

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>


-- 
Earthling Michel Dänzer           |                   http://www.amd.com
Libre software enthusiast         |          Debian, X and DRI developer
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH] drm/radeon/kms: fix fans after resume

[Alex Deucher]

2012/3/30 Michel Dänzer <michel@daenzer.net>:
> On Don, 2012-03-29 at 19:04 -0400, alexdeucher@gmail.com wrote:
>> From: Alex Deucher <alexander.deucher@amd.com>
>>
>> On pre-R600 asics, the SpeedFanControl table is not
>> executed as part of ASIC_Init as it is on newer asics.
>>
>> Fixes:
>> https://bugzilla.kernel.org/show_bug.cgi?id=29412
>>
>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>> Cc: stable@vger.kernel.org
>> ---
>>  drivers/gpu/drm/radeon/atom.c |   15 ++++++++++++++-
>>  drivers/gpu/drm/radeon/atom.h |    1 +
>>  2 files changed, 15 insertions(+), 1 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
>> index d1bd239..5ce9bf5 100644
>> --- a/drivers/gpu/drm/radeon/atom.c
>> +++ b/drivers/gpu/drm/radeon/atom.c
>> [...]
>> @@ -1317,7 +1320,17 @@ int atom_asic_init(struct atom_context *ctx)
>>
>>       if (!CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_INIT))
>>               return 1;
>> -     return atom_execute_table(ctx, ATOM_CMD_INIT, ps);
>> +     ret = atom_execute_table(ctx, ATOM_CMD_INIT, ps);
>> +     if (ret)
>> +             return ret;
>> +
>> +     memset(ps, 0, 64);
>> +
>> +     if (rdev->family < CHIP_R600) {
>> +             if (CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_SPDFANCNTL))
>> +                     atom_execute_table(ctx, ATOM_CMD_SPDFANCNTL, ps);
>
> Should the return value of atom_execute_table() be assigned to ret here
> as well? Either way,
>
> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
>

Yeah, I went back and forth with that.  I left it as is since
asic_init is the part we care most about in this case.

Alex

>
> --
> Earthling Michel Dänzer           |                   http://www.amd.com
> Libre software enthusiast         |          Debian, X and DRI developer