Re: [PATCH 1/4] drm/amdgpu: add stride to calculate oss ring offsets

["Deucher, Alexander" <Alexander.Deucher@xxxxxxx>]

[AMD Public Use]

I'm worried we'll miss a register by accident.  We went with per IP sub drivers to avoid handling complexities around IP differences if possible.  Also the scheme seems like kind of a one off compared to what we do for other IPs.  Can we structure it more like how we handle SDMA instancing since it seems to mainly affect IH RB instances? 

Alex

---

From: Kuehling, Felix <Felix.Kuehling@xxxxxxx>
Sent: Friday, March 20, 2020 10:20 AM
To: Deucher, Alexander <Alexander.Deucher@xxxxxxx>; Sierra Guiza, Alejandro (Alex) <Alex.Sierra@xxxxxxx>; amd-gfx@xxxxxxxxxxxxxxxxxxxxx <amd-gfx@xxxxxxxxxxxxxxxxxxxxx>
Subject: Re: [PATCH 1/4] drm/amdgpu: add stride to calculate oss ring offsets

 

On 2020-03-20 10:06, Deucher, Alexander wrote:

[AMD Public Use]

This seems kind of complicated and error prone.  I didn't realize the extent to the changes required.  I think it would be better to either add arcturus specific versions of these functions or just go with your original approach and add a new arcturus_ih.c.  If you go with the second route however, no need to show all your intermediate steps, just add the new files in one commit.

Hi Alex,

I suggested the approach in this patch series since to minimize code duplication and maintain readability of the code. I don't think it's very error prone. I believe this is more maintainable than a separate arcturus_ih.c. I'll have some more specific comments on Alejandro's patches.

Regards,
  Felix

Alex

 

---

From: amd-gfx <amd-gfx-bounces@xxxxxxxxxxxxxxxxxxxxx> on behalf of Alex Sierra <alex.sierra@xxxxxxx>
Sent: Thursday, March 19, 2020 8:22 PM
To: amd-gfx@xxxxxxxxxxxxxxxxxxxxx <amd-gfx@xxxxxxxxxxxxxxxxxxxxx>
Cc: Sierra Guiza, Alejandro (Alex) <Alex.Sierra@xxxxxxx>
Subject: [PATCH 1/4] drm/amdgpu: add stride to calculate oss ring offsets

 

Arcturus and vega10 share the same vega10_ih, however both
have different register offsets at the ih ring section.
This variable is used to help calculate ih ring register addresses
from the osssys, that corresponds to the current asic type.

Signed-off-by: Alex Sierra <alex.sierra@xxxxxxx>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c | 4 ++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h | 1 +
 2 files changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
index 5ed4227f304b..fa384ae9a9bc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
@@ -279,6 +279,10 @@ int amdgpu_irq_init(struct amdgpu_device *adev)
                                 amdgpu_hotplug_work_func);
         }
 
+       if (adev->asic_type == CHIP_ARCTURUS)
+               adev->irq.ring_stride = 1;
+       else
+               adev->irq.ring_stride = 0;
         INIT_WORK(&adev->irq.ih1_work, amdgpu_irq_handle_ih1);
         INIT_WORK(&adev->irq.ih2_work, amdgpu_irq_handle_ih2);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
index c718e94a55c9..1ec5b735cd9e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
@@ -97,6 +97,7 @@ struct amdgpu_irq {
         struct irq_domain               *domain; /* GPU irq controller domain */
         unsigned                        virq[AMDGPU_MAX_IRQ_SRC_ID];
         uint32_t                        srbm_soft_reset;
+       unsigned                        ring_stride;
 };
 
 void amdgpu_irq_disable_all(struct amdgpu_device *adev);
--
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://nam11.safelinks.protection.outlook.com/?url="">

_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://nam11.safelinks.protection.outlook.com/?url="">

_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx