All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jonathan Marek <jonathan@marek.ca>
To: Akhil P Oommen <akhilpo@codeaurora.org>, freedreno@lists.freedesktop.org
Cc: Rob Clark <robdclark@gmail.com>, Sean Paul <sean@poorly.run>,
	David Airlie <airlied@linux.ie>, Daniel Vetter <daniel@ffwll.ch>,
	Jordan Crouse <jordan@cosmicpenguin.net>,
	Eric Anholt <eric@anholt.net>,
	Sharat Masetty <smasetty@codeaurora.org>,
	"open list:DRM DRIVER FOR MSM ADRENO GPU" 
	<linux-arm-msm@vger.kernel.org>,
	"open list:DRM DRIVER FOR MSM ADRENO GPU" 
	<dri-devel@lists.freedesktop.org>,
	open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2 2/8] drm/msm/a6xx: use AOP-initialized PDC for a650
Date: Tue, 8 Jun 2021 11:54:44 -0400	[thread overview]
Message-ID: <b16de5ab-1485-a814-0885-c266a8706f0c@marek.ca> (raw)
In-Reply-To: <bd5dc0e5-2c49-31fe-a290-0d8e75b45c94@codeaurora.org>

On 5/31/21 3:24 AM, Akhil P Oommen wrote:
> On 5/13/2021 10:43 PM, Jonathan Marek wrote:
>> SM8250 AOP firmware already sets up PDC registers for us, and it only 
>> needs
>> to be enabled. This path will be used for other newer GPUs.
>>
>> Signed-off-by: Jonathan Marek <jonathan@marek.ca>
>> ---
>>   drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 21 ++++++++++++++++-----
>>   1 file changed, 16 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c 
>> b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
>> index 3d55e153fa9c..c1ee02d6371d 100644
>> --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
>> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
>> @@ -512,19 +512,26 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu 
>> *gmu)
>>       struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
>>       struct platform_device *pdev = to_platform_device(gmu->dev);
>>       void __iomem *pdcptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc");
>> -    void __iomem *seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq");
>> +    void __iomem *seqptr;
>>       uint32_t pdc_address_offset;
>> +    bool pdc_in_aop = false;
>> -    if (!pdcptr || !seqptr)
>> +    if (!pdcptr)
>>           goto err;
>> -    if (adreno_is_a618(adreno_gpu) || adreno_is_a640(adreno_gpu))
>> +    if (adreno_is_a650(adreno_gpu))
>> +        pdc_in_aop = true;
>> +    else if (adreno_is_a618(adreno_gpu) || adreno_is_a640(adreno_gpu))
>>           pdc_address_offset = 0x30090;
>> -    else if (adreno_is_a650(adreno_gpu))
>> -        pdc_address_offset = 0x300a0;
>>       else
>>           pdc_address_offset = 0x30080;
>> +    if (!pdc_in_aop) {
>> +        seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq");
>> +        if (!seqptr)
>> +            goto err;
>> +    }
>> +
>>       /* Disable SDE clock gating */
>>       gmu_write_rscc(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0, BIT(24));
>> @@ -556,6 +563,9 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu)
>>           gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 4, 
>> 0x0020e8a8);
>>       }
>> +    if (pdc_in_aop)
>> +        goto setup_pdc;
>> +
>>       /* Load PDC sequencer uCode for power up and power down sequence */
>>       pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0, 0xfebea1e1);
>>       pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 1, 0xa5a4a3a2);
>> @@ -596,6 +606,7 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu)
>>       pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3);
>>       /* Setup GPU PDC */
>> +setup_pdc:
>>       pdc_write(pdcptr, REG_A6XX_PDC_GPU_SEQ_START_ADDR, 0);
>>       pdc_write(pdcptr, REG_A6XX_PDC_GPU_ENABLE_PDC, 0x80000001);
>>
> 
> We can simply swap the order of PDC and rsc programming here and skip 
> pdc sequence to jump to the rscc programming for a650. This is the order 
> followed in the downstream driver anyway.
> 
> -Akhil.

The order is the same as the msm-4.19 kernel (msm-4.19 is what a650 
hardware are using).

Looks like the order was swapped for the msm-5.4 kernel, but if the 
order isn't important, I think it is preferable to keep the current 
order (to avoid a large diff mainly).

WARNING: multiple messages have this Message-ID (diff)
From: Jonathan Marek <jonathan@marek.ca>
To: Akhil P Oommen <akhilpo@codeaurora.org>, freedreno@lists.freedesktop.org
Cc: David Airlie <airlied@linux.ie>,
	"open list:DRM DRIVER FOR MSM ADRENO GPU"
	<linux-arm-msm@vger.kernel.org>,
	Sharat Masetty <smasetty@codeaurora.org>,
	open list <linux-kernel@vger.kernel.org>,
	"open list:DRM DRIVER FOR MSM ADRENO GPU"
	<dri-devel@lists.freedesktop.org>,
	Jordan Crouse <jordan@cosmicpenguin.net>,
	Sean Paul <sean@poorly.run>
Subject: Re: [PATCH v2 2/8] drm/msm/a6xx: use AOP-initialized PDC for a650
Date: Tue, 8 Jun 2021 11:54:44 -0400	[thread overview]
Message-ID: <b16de5ab-1485-a814-0885-c266a8706f0c@marek.ca> (raw)
In-Reply-To: <bd5dc0e5-2c49-31fe-a290-0d8e75b45c94@codeaurora.org>

On 5/31/21 3:24 AM, Akhil P Oommen wrote:
> On 5/13/2021 10:43 PM, Jonathan Marek wrote:
>> SM8250 AOP firmware already sets up PDC registers for us, and it only 
>> needs
>> to be enabled. This path will be used for other newer GPUs.
>>
>> Signed-off-by: Jonathan Marek <jonathan@marek.ca>
>> ---
>>   drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 21 ++++++++++++++++-----
>>   1 file changed, 16 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c 
>> b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
>> index 3d55e153fa9c..c1ee02d6371d 100644
>> --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
>> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c
>> @@ -512,19 +512,26 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu 
>> *gmu)
>>       struct adreno_gpu *adreno_gpu = &a6xx_gpu->base;
>>       struct platform_device *pdev = to_platform_device(gmu->dev);
>>       void __iomem *pdcptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc");
>> -    void __iomem *seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq");
>> +    void __iomem *seqptr;
>>       uint32_t pdc_address_offset;
>> +    bool pdc_in_aop = false;
>> -    if (!pdcptr || !seqptr)
>> +    if (!pdcptr)
>>           goto err;
>> -    if (adreno_is_a618(adreno_gpu) || adreno_is_a640(adreno_gpu))
>> +    if (adreno_is_a650(adreno_gpu))
>> +        pdc_in_aop = true;
>> +    else if (adreno_is_a618(adreno_gpu) || adreno_is_a640(adreno_gpu))
>>           pdc_address_offset = 0x30090;
>> -    else if (adreno_is_a650(adreno_gpu))
>> -        pdc_address_offset = 0x300a0;
>>       else
>>           pdc_address_offset = 0x30080;
>> +    if (!pdc_in_aop) {
>> +        seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq");
>> +        if (!seqptr)
>> +            goto err;
>> +    }
>> +
>>       /* Disable SDE clock gating */
>>       gmu_write_rscc(gmu, REG_A6XX_GPU_RSCC_RSC_STATUS0_DRV0, BIT(24));
>> @@ -556,6 +563,9 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu)
>>           gmu_write_rscc(gmu, REG_A6XX_RSCC_SEQ_MEM_0_DRV0 + 4, 
>> 0x0020e8a8);
>>       }
>> +    if (pdc_in_aop)
>> +        goto setup_pdc;
>> +
>>       /* Load PDC sequencer uCode for power up and power down sequence */
>>       pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0, 0xfebea1e1);
>>       pdc_write(seqptr, REG_A6XX_PDC_GPU_SEQ_MEM_0 + 1, 0xa5a4a3a2);
>> @@ -596,6 +606,7 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu)
>>       pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3);
>>       /* Setup GPU PDC */
>> +setup_pdc:
>>       pdc_write(pdcptr, REG_A6XX_PDC_GPU_SEQ_START_ADDR, 0);
>>       pdc_write(pdcptr, REG_A6XX_PDC_GPU_ENABLE_PDC, 0x80000001);
>>
> 
> We can simply swap the order of PDC and rsc programming here and skip 
> pdc sequence to jump to the rscc programming for a650. This is the order 
> followed in the downstream driver anyway.
> 
> -Akhil.

The order is the same as the msm-4.19 kernel (msm-4.19 is what a650 
hardware are using).

Looks like the order was swapped for the msm-5.4 kernel, but if the 
order isn't important, I think it is preferable to keep the current 
order (to avoid a large diff mainly).

  reply	other threads:[~2021-06-08 15:57 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-13 17:13 [PATCH v2 0/8] drm/msm/a6xx: add support for Adreno 660 GPU Jonathan Marek
2021-05-13 17:13 ` Jonathan Marek
2021-05-13 17:13 ` [PATCH v2 1/8] drm/msm: remove unused icc_path/ocmem_icc_path Jonathan Marek
2021-05-13 17:13   ` Jonathan Marek
2021-05-31  7:26   ` Akhil P Oommen
2021-05-31  7:26     ` Akhil P Oommen
2021-05-13 17:13 ` [PATCH v2 2/8] drm/msm/a6xx: use AOP-initialized PDC for a650 Jonathan Marek
2021-05-13 17:13   ` Jonathan Marek
2021-05-31  7:24   ` Akhil P Oommen
2021-05-31  7:24     ` Akhil P Oommen
2021-06-08 15:54     ` Jonathan Marek [this message]
2021-06-08 15:54       ` Jonathan Marek
2021-05-13 17:13 ` [PATCH v2 3/8] drm/msm/a6xx: fix incorrectly set uavflagprd_inv field for A650 Jonathan Marek
2021-05-13 17:13   ` Jonathan Marek
2021-05-31  7:33   ` Akhil P Oommen
2021-05-31  7:33     ` Akhil P Oommen
2021-05-13 17:13 ` [PATCH v2 4/8] drm/msm/a6xx: update/fix CP_PROTECT initialization Jonathan Marek
2021-05-13 17:13   ` Jonathan Marek
2021-05-31  8:09   ` Akhil P Oommen
2021-05-31  8:09     ` Akhil P Oommen
2021-05-13 17:14 ` [PATCH v2 5/8] drm/msm/a6xx: avoid shadow NULL reference in failure path Jonathan Marek
2021-05-13 17:14   ` Jonathan Marek
2021-05-31  9:25   ` Akhil P Oommen
2021-05-31  9:25     ` Akhil P Oommen
2021-05-13 17:14 ` [PATCH v2 6/8] drm/msm/a6xx: add support for Adreno 660 GPU Jonathan Marek
2021-05-13 17:14   ` Jonathan Marek
2021-05-31 15:05   ` Akhil P Oommen
2021-05-31 15:05     ` Akhil P Oommen
2021-06-08 16:53     ` Jonathan Marek
2021-06-08 16:53       ` Jonathan Marek
2021-05-13 17:14 ` [PATCH v2 7/8] drm/msm/a6xx: update a6xx_ucode_check_version for a660 Jonathan Marek
2021-05-13 17:14   ` Jonathan Marek
2021-05-31 15:06   ` Akhil P Oommen
2021-05-31 15:06     ` Akhil P Oommen
2021-05-13 17:14 ` [PATCH v2 8/8] drm/msm/a6xx: add a660 hwcg table Jonathan Marek
2021-05-13 17:14   ` Jonathan Marek
2021-05-31 15:10   ` Akhil P Oommen
2021-05-31 15:10     ` Akhil P Oommen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b16de5ab-1485-a814-0885-c266a8706f0c@marek.ca \
    --to=jonathan@marek.ca \
    --cc=airlied@linux.ie \
    --cc=akhilpo@codeaurora.org \
    --cc=daniel@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=eric@anholt.net \
    --cc=freedreno@lists.freedesktop.org \
    --cc=jordan@cosmicpenguin.net \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=robdclark@gmail.com \
    --cc=sean@poorly.run \
    --cc=smasetty@codeaurora.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.