Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Auld <matthew.auld@intel.com>
To: "Gupta, Anshuman" <anshuman.gupta@intel.com>,
	"intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>
Cc: "Vivi, Rodrigo" <rodrigo.vivi@intel.com>
Subject: Re: [Intel-xe] [PATCH v12 04/13] drm/xe/guc_pc: add missing mem_access for freq_rpe_show
Date: Tue, 27 Jun 2023 09:20:41 +0100	[thread overview]
Message-ID: <6d1e3a45-efd0-6ed4-88eb-e2d3fc752dd8@intel.com> (raw)
In-Reply-To: <CY5PR11MB62117F609203FC0958818E3F9527A@CY5PR11MB6211.namprd11.prod.outlook.com>

Hi,

On 27/06/2023 07:53, Gupta, Anshuman wrote:
> 
> 
>> -----Original Message----- From: Auld, Matthew 
>> <matthew.auld@intel.com> Sent: Monday, June 26, 2023 4:21 PM To: 
>> intel-xe@lists.freedesktop.org Cc: Brost, Matthew 
>> <matthew.brost@intel.com>; Vivi, Rodrigo <rodrigo.vivi@intel.com>; 
>> Gupta, Anshuman <anshuman.gupta@intel.com> Subject: [PATCH v12 
>> 04/13] drm/xe/guc_pc: add missing mem_access for freq_rpe_show
>> 
>> The mem_access is meant to cover any kind of device level memory 
>> access, mmio included.
>> 
>> Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Matthew 
>> Brost <matthew.brost@intel.com> Cc: Rodrigo Vivi 
>> <rodrigo.vivi@intel.com> Cc: Anshuman Gupta 
>> <anshuman.gupta@intel.com> --- drivers/gpu/drm/xe/xe_guc_pc.c | 4 
>> ++++ 1 file changed, 4 insertions(+)
>> 
>> diff --git a/drivers/gpu/drm/xe/xe_guc_pc.c 
>> b/drivers/gpu/drm/xe/xe_guc_pc.c index 5d5cf4b0d508..e3722a805837 
>> 100644 --- a/drivers/gpu/drm/xe/xe_guc_pc.c +++ 
>> b/drivers/gpu/drm/xe/xe_guc_pc.c @@ -430,8 +430,12 @@ static 
>> ssize_t freq_rpe_show(struct device *dev, struct device_attribute 
>> *attr, char *buf)  { struct xe_guc_pc *pc = dev_to_pc(dev); + 
>> struct xe_gt *gt = pc_to_gt(pc); +	struct xe_device *xe = 
>> gt_to_xe(gt);
>> 
>> +	xe_device_mem_access_get(xe);
> What is desirable to use here , xe_pm_runtime_get() or 
> xe_device_mem_access_get() ?

Rodrigo suggested using xe_device_mem_access_get() for MMIO, since it 
can be thought of as just another type of memory access, which seems 
pretty reasonable to me. Matt Roper also had thoughts about it here:
https://patchwork.freedesktop.org/patch/542538/?series=119345&rev=1

I don't have a strong opinion either way, just so long as we are 
consistent about it and have a reliable way to 
assert_device_active/assert_mem_access.

> It is register mmio access Few places in driver using 
> xe_pm_runtime_get() for mmio access.

The one in xe_vm.c looks to be a workaround for the race in patch 1, 
which is fixed in this series. Ignoring that I only see the cases in 
compat-i915-headers/i915_drv.h, which looks to be a special case for 
display/i915, but that is also converted to use mem_access later in the 
series, since I wanted to try adding assert_mem_access from all MMIO 
calls to see what CI says (it complained about freq_rpe_show() as per 
this patch). Are there some other places I missed that are using 
xe_pm_runtime_get?

> Thanks, Anshuman Gupta.
>> pc_update_rp_values(pc); +	xe_device_mem_access_put(xe); return 
>> sysfs_emit(buf, "%d\n", pc->rpe_freq);  }  static 
>> DEVICE_ATTR_RO(freq_rpe); -- 2.41.0
> 

  reply	other threads:[~2023-06-27  8:21 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-06-26 10:50 [Intel-xe] [PATCH v12 00/13] xe_device_mem_access fixes and related bits Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 01/13] drm/xe: fix xe_device_mem_access_get() races Matthew Auld
2023-06-30 15:22   ` Gupta, Anshuman
2023-07-04 11:25     ` Matthew Auld
2023-07-04 15:29       ` Gupta, Anshuman
2023-07-04 16:00         ` Matthew Auld
2023-07-11  9:00           ` Gupta, Anshuman
2023-07-11 11:06             ` Matthew Auld
2023-07-11 17:56               ` Gupta, Anshuman
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 02/13] drm/xe/vm: tidy up xe_runtime_pm usage Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 03/13] drm/xe/debugfs: grab mem_access around forcewake Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 04/13] drm/xe/guc_pc: add missing mem_access for freq_rpe_show Matthew Auld
2023-06-27  6:53   ` Gupta, Anshuman
2023-06-27  8:20     ` Matthew Auld [this message]
2023-06-27 10:14       ` Gupta, Anshuman
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 05/13] drm/xe/mmio: grab mem_access in xe_mmio_ioctl Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 06/13] drm/xe: ensure correct access_put ordering Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 07/13] drm/xe/pci: wrap probe with mem_access Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 08/13] drm/xe/display: use mem_access underneath Matthew Auld
2023-06-28  9:51   ` Gupta, Anshuman
2023-06-29  9:19     ` Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 09/13] drm/xe/mmio: enforce xe_device_assert_mem_access Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 10/13] drm/xe: drop xe_device_mem_access_get() from guc_ct_send Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 11/13] drm/xe/ggtt: prime ggtt->lock against FS_RECLAIM Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 12/13] drm/xe: drop xe_device_mem_access_get() from invalidation_vma Matthew Auld
2023-06-26 10:50 ` [Intel-xe] [PATCH v12 13/13] drm/xe: add lockdep annotation for xe_device_mem_access_get() Matthew Auld
2023-06-26 12:55 ` [Intel-xe] ✓ CI.Patch_applied: success for xe_device_mem_access fixes and related bits (rev2) Patchwork
2023-06-26 12:56 ` [Intel-xe] ✗ CI.checkpatch: warning " Patchwork
2023-06-26 12:57 ` [Intel-xe] ✓ CI.KUnit: success " Patchwork
2023-06-26 13:01 ` [Intel-xe] ✓ CI.Build: " Patchwork
2023-06-26 13:01 ` [Intel-xe] ✓ CI.Hooks: " Patchwork
2023-06-26 13:02 ` [Intel-xe] ✓ CI.checksparse: " Patchwork
2023-06-26 13:46 ` [Intel-xe] ○ CI.BAT: info " Patchwork
2023-06-30  6:21 ` [Intel-xe] [PATCH v12 00/13] xe_device_mem_access fixes and related bits Dixit, Ashutosh
2023-06-30 11:07   ` Matthew Auld
2023-06-30 16:59     ` Dixit, Ashutosh

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=6d1e3a45-efd0-6ed4-88eb-e2d3fc752dd8@intel.com \
    --to=matthew.auld@intel.com \
    --cc=anshuman.gupta@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=rodrigo.vivi@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