From: Riana Tauro <riana.tauro@intel.com>
To: Michal Wajdeczko <michal.wajdeczko@intel.com>,
<intel-xe@lists.freedesktop.org>
Cc: <anshuman.gupta@intel.com>, <rodrigo.vivi@intel.com>,
<daniele.ceraolospurio@intel.com>, <matthew.brost@intel.com>,
<karthik.poosa@intel.com>
Subject: Re: [PATCH v4 1/2] drm/xe: modify GuC CTB enable communication
Date: Thu, 18 Jul 2024 10:47:00 +0530 [thread overview]
Message-ID: <b150c5fe-0bc8-4a46-87f8-d85286aee8eb@intel.com> (raw)
In-Reply-To: <0b92ab50-8bcd-4f1b-a095-651f6e8f955f@intel.com>
Hi Michal
Thank you for the review comments
On 7/17/2024 1:34 PM, Michal Wajdeczko wrote:
>
>
> On 17.07.2024 07:55, Riana Tauro wrote:
>> Modify GuC CTB enable communication function to only set the
>> CTB state if registration is not required.
>> This will be used while resuming from D3Hot where
>> registration of CTB is not necessary.
>>
>> Also move xe_guc_ct_set_state to header file.
>
> hmm, but this will expose internals of the guc_ct layer to the outside
> see below for alternate solution
yeah that is the reason i had split the functions in the last rev.
Will follow the alternate solution of passing the register_ctb
to the CT layer.
Thanks,
Riana
>
>>
>> v2: split patches (Rodrigo)
>>
>> Signed-off-by: Riana Tauro <riana.tauro@intel.com>
>> ---
>> drivers/gpu/drm/xe/xe_guc.c | 25 +++++++++++++++++++------
>> drivers/gpu/drm/xe/xe_guc.h | 2 +-
>> drivers/gpu/drm/xe/xe_guc_ct.c | 3 +--
>> drivers/gpu/drm/xe/xe_guc_ct.h | 2 ++
>> drivers/gpu/drm/xe/xe_uc.c | 4 ++--
>> 5 files changed, 25 insertions(+), 11 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_guc.c b/drivers/gpu/drm/xe/xe_guc.c
>> index eb655cee19f7..8131c290a4f4 100644
>> --- a/drivers/gpu/drm/xe/xe_guc.c
>> +++ b/drivers/gpu/drm/xe/xe_guc.c
>> @@ -755,7 +755,7 @@ static int vf_guc_min_load_for_hwconfig(struct xe_guc *guc)
>> if (ret)
>> return ret;
>>
>> - ret = xe_guc_enable_communication(guc);
>> + ret = xe_guc_enable_communication(guc, true);
>> if (ret)
>> return ret;
>>
>> @@ -800,7 +800,7 @@ int xe_guc_min_load_for_hwconfig(struct xe_guc *guc)
>> if (ret)
>> return ret;
>>
>> - ret = xe_guc_enable_communication(guc);
>> + ret = xe_guc_enable_communication(guc, true);
>> if (ret)
>> return ret;
>>
>> @@ -854,7 +854,16 @@ static void guc_enable_irq(struct xe_guc *guc)
>> xe_mmio_rmw32(gt, GUC_SG_INTR_MASK, events, 0);
>> }
>>
>> -int xe_guc_enable_communication(struct xe_guc *guc)
>> +/**
>> + * xe_guc_enable_communication - Enable GuC CTB communication
>> + * @guc: GuC object
>> + * @register_ctb: true to register CTB, false otherwise
>> + *
>> + * Enables GuC CTB Communication and IRQ
>> + *
>> + * Return: 0 on success, negative error code otherwise
>> + */
>> +int xe_guc_enable_communication(struct xe_guc *guc, bool register_ctb)
>> {
>> struct xe_device *xe = guc_to_xe(guc);
>> int err;
>> @@ -870,9 +879,13 @@ int xe_guc_enable_communication(struct xe_guc *guc)
>> guc_enable_irq(guc);
>> }
>>
>> - err = xe_guc_ct_enable(&guc->ct);
>> - if (err)
>> - return err;
>> + if (register_ctb) {
>> + err = xe_guc_ct_enable(&guc->ct);
>> + if (err)
>> + return err;
>> + } else {
>> + xe_guc_ct_set_state(&guc->ct, XE_GUC_CT_STATE_ENABLED);
>> + }
>
> maybe better option would be to pass the register_ctb flag to the
> xe_guc_ct_enable() so it can do either full enabling (including CTB
> registration) or just change the internal state:
>
> int xe_guc_enable_communication(struct xe_guc *guc, bool register_ctb)
> {
> err = xe_guc_ct_enable(&guc->ct, register_ctb);
> ...
> }
>
> int xe_guc_ct_enable(struct xe_guc_ct *ct, bool register_ctb)
> {
> if (register_ctb) {
> ...
> }
>
> ct_set_state(ct, XE_GUC_CT_STATE_ENABLED);
> return 0;
> }
>
>>
>> guc_handle_mmio_msg(guc);
>>
>> diff --git a/drivers/gpu/drm/xe/xe_guc.h b/drivers/gpu/drm/xe/xe_guc.h
>> index af59c9545753..b49444297eb0 100644
>> --- a/drivers/gpu/drm/xe/xe_guc.h
>> +++ b/drivers/gpu/drm/xe/xe_guc.h
>> @@ -20,7 +20,7 @@ int xe_guc_post_load_init(struct xe_guc *guc);
>> int xe_guc_reset(struct xe_guc *guc);
>> int xe_guc_upload(struct xe_guc *guc);
>> int xe_guc_min_load_for_hwconfig(struct xe_guc *guc);
>> -int xe_guc_enable_communication(struct xe_guc *guc);
>> +int xe_guc_enable_communication(struct xe_guc *guc, bool register_ctb);
>> int xe_guc_suspend(struct xe_guc *guc);
>> void xe_guc_notify(struct xe_guc *guc);
>> int xe_guc_auth_huc(struct xe_guc *guc, u32 rsa_addr);
>> diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c
>> index 7d2e937da1d8..44c44e086a98 100644
>> --- a/drivers/gpu/drm/xe/xe_guc_ct.c
>> +++ b/drivers/gpu/drm/xe/xe_guc_ct.c
>> @@ -318,8 +318,7 @@ static int guc_ct_control_toggle(struct xe_guc_ct *ct, bool enable)
>> return ret > 0 ? -EPROTO : ret;
>> }
>>
>> -static void xe_guc_ct_set_state(struct xe_guc_ct *ct,
>> - enum xe_guc_ct_state state)
>> +void xe_guc_ct_set_state(struct xe_guc_ct *ct, enum xe_guc_ct_state state)
>> {
>> mutex_lock(&ct->lock); /* Serialise dequeue_one_g2h() */
>> spin_lock_irq(&ct->fast_lock); /* Serialise CT fast-path */
>> diff --git a/drivers/gpu/drm/xe/xe_guc_ct.h b/drivers/gpu/drm/xe/xe_guc_ct.h
>> index 190202fce2d0..60d46739548c 100644
>> --- a/drivers/gpu/drm/xe/xe_guc_ct.h
>> +++ b/drivers/gpu/drm/xe/xe_guc_ct.h
>> @@ -15,6 +15,8 @@ int xe_guc_ct_enable(struct xe_guc_ct *ct);
>> void xe_guc_ct_disable(struct xe_guc_ct *ct);
>> void xe_guc_ct_stop(struct xe_guc_ct *ct);
>> void xe_guc_ct_fast_path(struct xe_guc_ct *ct);
>> +void xe_guc_ct_set_state(struct xe_guc_ct *ct,
>> + enum xe_guc_ct_state state);
>>
>> struct xe_guc_ct_snapshot *
>> xe_guc_ct_snapshot_capture(struct xe_guc_ct *ct, bool atomic);
>> diff --git a/drivers/gpu/drm/xe/xe_uc.c b/drivers/gpu/drm/xe/xe_uc.c
>> index 0f240534fb72..eb9c1443c849 100644
>> --- a/drivers/gpu/drm/xe/xe_uc.c
>> +++ b/drivers/gpu/drm/xe/xe_uc.c
>> @@ -154,7 +154,7 @@ static int vf_uc_init_hw(struct xe_uc *uc)
>> if (err)
>> return err;
>>
>> - err = xe_guc_enable_communication(&uc->guc);
>> + err = xe_guc_enable_communication(&uc->guc, true);
>> if (err)
>> return err;
>>
>> @@ -194,7 +194,7 @@ int xe_uc_init_hw(struct xe_uc *uc)
>> if (ret)
>> return ret;
>>
>> - ret = xe_guc_enable_communication(&uc->guc);
>> + ret = xe_guc_enable_communication(&uc->guc, true);
>> if (ret)
>> return ret;
>>
next prev parent reply other threads:[~2024-07-18 5:17 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-07-17 5:55 [PATCH v4 0/2] drm/xe: remove GuC reload in D3Hot path Riana Tauro
2024-07-17 5:55 ` [PATCH v4 1/2] drm/xe: modify GuC CTB enable communication Riana Tauro
2024-07-17 8:04 ` Michal Wajdeczko
2024-07-18 5:17 ` Riana Tauro [this message]
2024-07-17 5:55 ` [PATCH v4 2/2] drm/xe: remove GuC reload in D3Hot path Riana Tauro
2024-07-17 8:38 ` Michal Wajdeczko
2024-07-18 6:54 ` Riana Tauro
2024-07-18 17:56 ` Matthew Brost
2024-07-17 7:01 ` ✓ CI.Patch_applied: success for drm/xe: remove GuC reload in D3Hot path (rev4) Patchwork
2024-07-17 7:01 ` ✓ CI.checkpatch: " Patchwork
2024-07-17 7:02 ` ✓ CI.KUnit: " Patchwork
2024-07-17 7:14 ` ✓ CI.Build: " Patchwork
2024-07-17 7:16 ` ✓ CI.Hooks: " Patchwork
2024-07-17 7:18 ` ✓ CI.checksparse: " Patchwork
2024-07-17 7:41 ` ✓ CI.BAT: " Patchwork
2024-07-17 9:15 ` ✗ CI.FULL: failure " 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=b150c5fe-0bc8-4a46-87f8-d85286aee8eb@intel.com \
--to=riana.tauro@intel.com \
--cc=anshuman.gupta@intel.com \
--cc=daniele.ceraolospurio@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=karthik.poosa@intel.com \
--cc=matthew.brost@intel.com \
--cc=michal.wajdeczko@intel.com \
--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