From: Michal Wajdeczko <michal.wajdeczko@intel.com>
To: Matt Roper <matthew.d.roper@intel.com>
Cc: intel-xe@lists.freedesktop.org
Subject: Re: [PATCH 10/10] drm/xe/vf: Add VF specific interrupt handler
Date: Thu, 14 Dec 2023 01:20:48 +0100 [thread overview]
Message-ID: <73bac872-4b3c-4a1f-a711-46dc27e0731e@intel.com> (raw)
In-Reply-To: <20231213005751.GQ1327160@mdroper-desk1.amr.corp.intel.com>
On 13.12.2023 01:57, Matt Roper wrote:
> On Tue, Dec 12, 2023 at 10:00:54PM +0100, Michal Wajdeczko wrote:
>> There are small differences in handling of the register based
>> interrupts on the VF driver as some registers are not accessible
>> to the VF driver. Additionally VFs must support Memory Based
>> Interrupts. Add VF specific interrupt handler for this.
>>
>> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
>> ---
>> drivers/gpu/drm/xe/xe_irq.c | 71 +++++++++++++++++++++++++++++++++++++
>> 1 file changed, 71 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_irq.c b/drivers/gpu/drm/xe/xe_irq.c
>> index d1f5ba4bb745..907c8ff0fa21 100644
>> --- a/drivers/gpu/drm/xe/xe_irq.c
>> +++ b/drivers/gpu/drm/xe/xe_irq.c
>> @@ -17,7 +17,9 @@
>> #include "xe_gt.h"
>> #include "xe_guc.h"
>> #include "xe_hw_engine.h"
>> +#include "xe_memirq.h"
>> #include "xe_mmio.h"
>> +#include "xe_sriov.h"
>>
>> /*
>> * Interrupt registers for a unit are always consecutive and ordered
>> @@ -498,6 +500,9 @@ static void xelp_irq_reset(struct xe_tile *tile)
>>
>> gt_irq_reset(tile);
>>
>> + if (IS_SRIOV_VF(tile_to_xe(tile)))
>> + return;
>> +
>> mask_and_disable(tile, PCU_IRQ_OFFSET);
>> }
>>
>> @@ -508,6 +513,9 @@ static void dg1_irq_reset(struct xe_tile *tile)
>>
>> gt_irq_reset(tile);
>>
>> + if (IS_SRIOV_VF(tile_to_xe(tile)))
>> + return;
>> +
>> mask_and_disable(tile, PCU_IRQ_OFFSET);
>> }
>>
>> @@ -518,11 +526,34 @@ static void dg1_irq_reset_mstr(struct xe_tile *tile)
>> xe_mmio_write32(mmio, GFX_MSTR_IRQ, ~0);
>> }
>>
>> +static void vf_irq_reset(struct xe_device *xe)
>> +{
>> + struct xe_tile *tile;
>> + unsigned int id;
>> +
>> + xe_assert(xe, IS_SRIOV_VF(xe));
>> +
>> + if (GRAPHICS_VERx100(xe) < 1210)
>> + xelp_intr_disable(xe);
>> + else
>> + xe_assert(xe, xe_device_has_memirq(xe));
>
> Isn't this going to fail on DG1 (version 12.10)? I don't believe it has
> memirq does it?
there is no SRIOV on DG1, so we won't be using this function there
>
> Of course I don't know how much effort we want to put into these older
> platforms that are just temporary development vehicles and will never be
> "officially" supported by the Xe driver. If we never enable SRIOV on
> those, then it doesn't really matter.
>
>
> Matt
>
>> +
>> + for_each_tile(tile, xe, id) {
>> + if (xe_device_has_memirq(xe))
>> + xe_memirq_reset(&tile->sriov.vf.memirq);
>> + else
>> + gt_irq_reset(tile);
>> + }
>> +}
>> +
>> static void xe_irq_reset(struct xe_device *xe)
>> {
>> struct xe_tile *tile;
>> u8 id;
>>
>> + if (IS_SRIOV_VF(xe))
>> + return vf_irq_reset(xe);
>> +
>> for_each_tile(tile, xe, id) {
>> if (GRAPHICS_VERx100(xe) >= 1210)
>> dg1_irq_reset(tile);
>> @@ -545,8 +576,26 @@ static void xe_irq_reset(struct xe_device *xe)
>> }
>> }
>>
>> +static void vf_irq_postinstall(struct xe_device *xe)
>> +{
>> + struct xe_tile *tile;
>> + unsigned int id;
>> +
>> + for_each_tile(tile, xe, id)
>> + if (xe_device_has_memirq(xe))
>> + xe_memirq_postinstall(&tile->sriov.vf.memirq);
>> +
>> + if (GRAPHICS_VERx100(xe) < 1210)
>> + xelp_intr_enable(xe, true);
>> + else
>> + xe_assert(xe, xe_device_has_memirq(xe));
>> +}
>> +
>> static void xe_irq_postinstall(struct xe_device *xe)
>> {
>> + if (IS_SRIOV_VF(xe))
>> + return vf_irq_postinstall(xe);
>> +
>> xe_display_irq_postinstall(xe, xe_root_mmio_gt(xe));
>>
>> /*
>> @@ -563,8 +612,30 @@ static void xe_irq_postinstall(struct xe_device *xe)
>> xelp_intr_enable(xe, true);
>> }
>>
>> +static irqreturn_t vf_mem_irq_handler(int irq, void *arg)
>> +{
>> + struct xe_device *xe = arg;
>> + struct xe_tile *tile;
>> + unsigned int id;
>> +
>> + spin_lock(&xe->irq.lock);
>> + if (!xe->irq.enabled) {
>> + spin_unlock(&xe->irq.lock);
>> + return IRQ_NONE;
>> + }
>> + spin_unlock(&xe->irq.lock);
>> +
>> + for_each_tile(tile, xe, id)
>> + xe_memirq_handler(&tile->sriov.vf.memirq);
>> +
>> + return IRQ_HANDLED;
>> +}
>> +
>> static irq_handler_t xe_irq_handler(struct xe_device *xe)
>> {
>> + if (IS_SRIOV_VF(xe) && xe_device_has_memirq(xe))
>> + return vf_mem_irq_handler;
>> +
>> if (GRAPHICS_VERx100(xe) >= 1210)
>> return dg1_irq_handler;
>> else
>> --
>> 2.25.1
>>
>
next prev parent reply other threads:[~2023-12-14 0:20 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-12 21:00 [PATCH 00/10] Introduce VF specific interrupts handler Michal Wajdeczko
2023-12-12 21:00 ` [PATCH 01/10] drm/xe: Add device flag for memory based IRQ support Michal Wajdeczko
2023-12-12 21:41 ` Matt Roper
2023-12-13 23:04 ` Michal Wajdeczko
2023-12-12 21:00 ` [PATCH 02/10] drm/xe: Add command MI_LOAD_REGISTER_MEM Michal Wajdeczko
2023-12-12 21:48 ` Matt Roper
2023-12-12 21:00 ` [PATCH 03/10] drm/xe: Define registers used by memory based irq processing Michal Wajdeczko
2023-12-12 21:55 ` Matt Roper
2023-12-12 21:00 ` [PATCH 04/10] drm/xe: Update LRC context layout definitions Michal Wajdeczko
2023-12-12 22:15 ` Matt Roper
2023-12-12 21:00 ` [PATCH 05/10] drm/xe: Update definition of GT_INTR_DW Michal Wajdeczko
2023-12-12 22:23 ` Matt Roper
2023-12-12 21:00 ` [PATCH 06/10] drm/xe: Define IRQ offsets used by HW engines Michal Wajdeczko
2023-12-12 22:43 ` Matt Roper
2023-12-12 21:00 ` [PATCH 07/10] drm/xe/vf: Introduce Memory Based Interrupts Handler Michal Wajdeczko
2023-12-13 0:15 ` Matt Roper
2023-12-14 0:18 ` Michal Wajdeczko
2023-12-12 21:00 ` [PATCH 08/10] drm/xe/vf: Update LRC with memory based interrupts data Michal Wajdeczko
2023-12-13 0:51 ` Matt Roper
2023-12-12 21:00 ` [PATCH 09/10] drm/xe/vf: Setup memory based interrupts in GuC Michal Wajdeczko
2023-12-13 0:52 ` Matt Roper
2023-12-12 21:00 ` [PATCH 10/10] drm/xe/vf: Add VF specific interrupt handler Michal Wajdeczko
2023-12-13 0:57 ` Matt Roper
2023-12-14 0:20 ` Michal Wajdeczko [this message]
2023-12-12 23:29 ` ✓ CI.Patch_applied: success for Introduce VF specific interrupts handler Patchwork
2023-12-12 23:30 ` ✗ CI.checkpatch: warning " Patchwork
2023-12-12 23:31 ` ✓ CI.KUnit: success " Patchwork
2023-12-12 23:38 ` ✓ CI.Build: " Patchwork
2023-12-12 23:39 ` ✓ CI.Hooks: " Patchwork
2023-12-12 23:40 ` ✓ CI.checksparse: " Patchwork
2023-12-13 0:14 ` ✓ CI.BAT: " Patchwork
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=73bac872-4b3c-4a1f-a711-46dc27e0731e@intel.com \
--to=michal.wajdeczko@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.d.roper@intel.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox