From: Jani Nikula <jani.nikula@linux.intel.com>
To: "Murthy, Arun R" <arun.r.murthy@intel.com>,
Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
Maxime Ripard <mripard@kernel.org>,
Thomas Zimmermann <tzimmermann@suse.de>,
David Airlie <airlied@gmail.com>, Simona Vetter <simona@ffwll.ch>,
Rodrigo Vivi <rodrigo.vivi@intel.com>,
Joonas Lahtinen <joonas.lahtinen@linux.intel.com>,
Tvrtko Ursulin <tursulin@ursulin.net>,
xaver.hugl@kde.org, harry.wentland@amd.com,
uma.shankar@intel.com
Cc: dri-devel@lists.freedesktop.org, intel-gfx@lists.freedesktop.org,
intel-xe@lists.freedesktop.org
Subject: Re: [PATCH v4 3/4] drm/atomic: Return user readable error in atomic_ioctl
Date: Wed, 03 Sep 2025 10:20:12 +0300 [thread overview]
Message-ID: <bc6dfb6765b153901c3af5b1c008c8d82f99bdc2@intel.com> (raw)
In-Reply-To: <d8513636-f616-4219-8e46-24376dca7b97@intel.com>
On Wed, 03 Sep 2025, "Murthy, Arun R" <arun.r.murthy@intel.com> wrote:
> On 02-09-2025 15:46, Jani Nikula wrote:
>> On Tue, 02 Sep 2025, Arun R Murthy <arun.r.murthy@intel.com> wrote:
>>> Add user readable error codes for failure cases in drm_atomic_ioctl() so
>>> that user can decode the error code and take corrective measurements.
>>>
>>> Signed-off-by: Arun R Murthy <arun.r.murthy@intel.com>
>>> ---
>>> drivers/gpu/drm/drm_atomic.c | 23 +++++++----
>>> drivers/gpu/drm/drm_atomic_uapi.c | 83 ++++++++++++++++++++++++++++++---------
>>> include/drm/drm_atomic.h | 19 +++++++++
>>> 3 files changed, 100 insertions(+), 25 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
>>> index cd15cf52f0c9144711da5879da57884674aea9e4..4f6c7e659b362f0887ffcc85dade1122fd30df3d 100644
>>> --- a/drivers/gpu/drm/drm_atomic.c
>>> +++ b/drivers/gpu/drm/drm_atomic.c
>>> @@ -1463,6 +1463,7 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
>>> unsigned int requested_crtc = 0;
>>> unsigned int affected_crtc = 0;
>>> int i, ret = 0;
>>> + char *err_string;
>>>
>>> drm_dbg_atomic(dev, "checking %p\n", state);
>>>
>>> @@ -1511,8 +1512,13 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
>>> if (!state->allow_modeset) {
>>> for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
>>> if (drm_atomic_crtc_needs_modeset(new_crtc_state)) {
>>> - drm_dbg_atomic(dev, "[CRTC:%d:%s] requires full modeset\n",
>>> - crtc->base.id, crtc->name);
>>> + err_string = "requires full modeset";
>> This points a non-const pointer to const data.
>>
>> Anyway none of this should require or start a precendent of using local
>> variables all over the place for the error strings. It's ugly. Ditto
>> below for all the places.
>>
>> If all the places calling drm_mode_atomic_add_error_msg() do debug
>> logging, maybe that's where the debug logging should be?
>>
>>> + drm_dbg_atomic(dev, "[CRTC:%d:%s] %s\n",
>>> + crtc->base.id, crtc->name, err_string);
>>> + drm_mode_atomic_add_error_msg(state->error_code,
>>> + DRM_MODE_ATOMIC_CRTC_NEED_FULL_MODESET,
>>> + err_string);
>> I guess you didn't want to follow what Maarten suggested.
> Yes, felt so, as adding a new function doesnt make much sense just for
> filling the struct drm_mode_atomic_err_code.
>
> Will remove this function and fill the struct drm_mode_atomic_err_code
> in place, thereby the below comment also should be handled.
Nobody suggested that! Keep the function.
>
> Thanks and Regards,
> Arun R Murthy
> -------------------
>
>>
>>> +
>>> return -EINVAL;
>> Adding the return is a functional change, and I mentioned it
>> earlier. Ignoring review is one of the certain ways to stop receiving
>> review in the future.
> Will take it out of this series.
>>> }
>>> }
>>> @@ -1534,11 +1540,14 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
>>> * so compositors know what's going on.
>>> */
>>> if (affected_crtc != requested_crtc) {
>>> - drm_dbg_atomic(dev,
>>> - "driver added CRTC to commit: requested 0x%x, affected 0x%0x\n",
>>> - requested_crtc, affected_crtc);
>>> - WARN(!state->allow_modeset, "adding CRTC not allowed without modesets: requested 0x%x, affected 0x%0x\n",
>>> - requested_crtc, affected_crtc);
>>> + err_string = "adding CRTC not allowed without modesets";
>>> + drm_dbg_atomic(dev, "%s: requested 0x%x, affected 0x%0x\n",
>>> + err_string, requested_crtc, affected_crtc);
>>> + drm_mode_atomic_add_error_msg(state->error_code,
>>> + DRM_MODE_ATOMIC_NEED_FULL_MODESET,
>>> + err_string);
>>> +
>>> + return -EINVAL;
>> Adding the return is a functional change, and I mentioned it earlier.
> Will take it out of this series.
>>
>>> }
>>>
>>> return 0;
>>> diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
>>> index 85dbdaa4a2e25878c953b9b41539c8566d55c6d9..60f1b8baebce0db1ce984c8cda56039261b519e8 100644
>>> --- a/drivers/gpu/drm/drm_atomic_uapi.c
>>> +++ b/drivers/gpu/drm/drm_atomic_uapi.c
>>> @@ -1017,6 +1017,7 @@ int drm_atomic_set_property(struct drm_atomic_state *state,
>>> struct drm_mode_object *ref;
>>> u64 old_val;
>>> int ret;
>>> + char *err_string;
>>>
>>> if (!drm_property_change_valid_get(prop, prop_value, &ref))
>>> return -EINVAL;
>>> @@ -1058,6 +1059,12 @@ int drm_atomic_set_property(struct drm_atomic_state *state,
>>> ret = drm_atomic_crtc_get_property(crtc, crtc_state,
>>> prop, &old_val);
>>> ret = drm_atomic_check_prop_changes(ret, old_val, prop_value, prop);
>>> + if (ret) {
>>> + err_string = "property change not allowed in async flip";
>>> + drm_mode_atomic_add_error_msg(state->error_code,
>>> + DRM_MODE_ATOMIC_ASYNC_PROP_CHANGED,
>>> + err_string);
>>> + }
>>> break;
>>> }
>>>
>>> @@ -1096,9 +1103,14 @@ int drm_atomic_set_property(struct drm_atomic_state *state,
>>> ret = plane_funcs->atomic_async_check(plane, state, true);
>>>
>>> if (ret) {
>>> + err_string = "plane does not support async flip";
>>> + drm_mode_atomic_add_error_msg(state->error_code,
>>> + DRM_MODE_ATOMIC_ASYNC_NOT_SUPP_PLANE,
>>> + err_string);
>>> drm_dbg_atomic(prop->dev,
>>> - "[PLANE:%d:%s] does not support async flips\n",
>>> - obj->id, plane->name);
>>> + "[PLANE:%d:%s] %s\n",
>>> + obj->id, plane->name,
>>> + err_string);
>>> break;
>>> }
>>> }
>>> @@ -1390,42 +1402,63 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
>>> uint32_t __user *props_ptr = (uint32_t __user *)(unsigned long)(arg->props_ptr);
>>> uint64_t __user *prop_values_ptr = (uint64_t __user *)(unsigned long)(arg->prop_values_ptr);
>>> unsigned int copied_objs, copied_props;
>>> - struct drm_atomic_state *state;
>>> + struct drm_atomic_state *state = NULL;
>> Unrelated change.
> Noted!
>>
>>> struct drm_modeset_acquire_ctx ctx;
>>> struct drm_out_fence_state *fence_state;
>>> + struct drm_mode_atomic_err_code error_code;
>>> + struct drm_mode_atomic_err_code __user *error_code_ptr;
>>> int ret = 0;
>>> unsigned int i, j, num_fences;
>>> bool async_flip = false;
>>> + char *err_string;
>>>
>>> /* disallow for drivers not supporting atomic: */
>>> if (!drm_core_check_feature(dev, DRIVER_ATOMIC))
>>> return -EOPNOTSUPP;
>>>
>>> + if (!arg->reserved)
>>> + drm_err(dev, "memory not allocated for drm_atomic error reporting\n");
>>> + else
>>> + /* update the error code if any error to allow user handling it */
>>> + error_code_ptr = (struct drm_mode_atomic_err_code __user *)
>>> + (unsigned long)arg->reserved;
>>> +
>>> + memset(&error_code, 0, sizeof(struct drm_mode_atomic_err_code));
>>> +
>>> /* disallow for userspace that has not enabled atomic cap (even
>>> * though this may be a bit overkill, since legacy userspace
>>> * wouldn't know how to call this ioctl)
>>> */
>>> if (!file_priv->atomic) {
>>> + err_string = "DRM_ATOMIC capability not enabled";
>> Yeah, just repeating, don't start using this err_string local variable
>> idea.
>>
>>> drm_dbg_atomic(dev,
>>> - "commit failed: atomic cap not enabled\n");
>>> - return -EINVAL;
>>> + "commit failed: %s\n", err_string);
>>> + drm_mode_atomic_add_error_msg(&error_code,
>>> + DRM_MODE_ATOMIC_INVALID_API_USAGE,
>>> + err_string);
>>> + ret = -EINVAL;
>>> + goto out;
>>> }
>>>
>>> if (arg->flags & ~DRM_MODE_ATOMIC_FLAGS) {
>>> - drm_dbg_atomic(dev, "commit failed: invalid flag\n");
>>> - return -EINVAL;
>>> - }
>>> -
>>> - if (arg->reserved) {
>>> - drm_dbg_atomic(dev, "commit failed: reserved field set\n");
>>> - return -EINVAL;
>>> + err_string = "invalid flag";
>>> + drm_dbg_atomic(dev, "commit failed: %s\n", err_string);
>>> + drm_mode_atomic_add_error_msg(&error_code,
>>> + DRM_MODE_ATOMIC_INVALID_API_USAGE,
>>> + err_string);
>>> + ret = -EINVAL;
>>> + goto out;
>>> }
>>>
>>> if (arg->flags & DRM_MODE_PAGE_FLIP_ASYNC) {
>>> if (!dev->mode_config.async_page_flip) {
>>> - drm_dbg_atomic(dev,
>>> - "commit failed: DRM_MODE_PAGE_FLIP_ASYNC not supported\n");
>>> - return -EINVAL;
>>> + err_string = "DRM_MODE_PAGE_FLIP_ASYNC not supported with ATOMIC ioctl";
>>> + drm_dbg_atomic(dev, "commit failed: %s\n", err_string);
>>> + drm_mode_atomic_add_error_msg(&error_code,
>>> + DRM_MODE_ATOMIC_INVALID_API_USAGE,
>>> + err_string);
>>> + ret = -EINVAL;
>>> + goto out;
>>> }
>>>
>>> async_flip = true;
>>> @@ -1434,9 +1467,13 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
>>> /* can't test and expect an event at the same time. */
>>> if ((arg->flags & DRM_MODE_ATOMIC_TEST_ONLY) &&
>>> (arg->flags & DRM_MODE_PAGE_FLIP_EVENT)) {
>>> - drm_dbg_atomic(dev,
>>> - "commit failed: page-flip event requested with test-only commit\n");
>>> - return -EINVAL;
>>> + err_string = "page-flip event requested with test-only commit";
>>> + drm_dbg_atomic(dev, "commit failed: %s\n", err_string);
>>> + drm_mode_atomic_add_error_msg(&error_code,
>>> + DRM_MODE_ATOMIC_INVALID_API_USAGE,
>>> + err_string);
>>> + ret = -EINVAL;
>>> + goto out;
>>> }
>>>
>>> state = drm_atomic_state_alloc(dev);
>>> @@ -1447,6 +1484,8 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
>>> state->acquire_ctx = &ctx;
>>> state->allow_modeset = !!(arg->flags & DRM_MODE_ATOMIC_ALLOW_MODESET);
>>>
>>> + state->error_code = &error_code;
>>> +
>>> retry:
>>> copied_objs = 0;
>>> copied_props = 0;
>>> @@ -1543,6 +1582,14 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
>>> }
>>>
>>> out:
>>> + /* update the error code if any error to allow user handling it */
>>> + if (ret < 0 && arg->reserved)
>>> + if (copy_to_user(error_code_ptr, &error_code, sizeof(error_code)))
>>> + return -EFAULT;
>>> +
>>> + if (!state)
>>> + return ret;
>>> +
>>> complete_signaling(dev, state, fence_state, num_fences, !ret);
>>>
>>> if (ret == -EDEADLK) {
>>> diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
>>> index d380001b24b4223baa54dae6c3c43e19dfb1958d..8956b175795128422eefc2bc047b320b80aedc3f 100644
>>> --- a/include/drm/drm_atomic.h
>>> +++ b/include/drm/drm_atomic.h
>>> @@ -1255,4 +1255,23 @@ struct drm_bridge_state *
>>> drm_atomic_get_new_bridge_state(const struct drm_atomic_state *state,
>>> struct drm_bridge *bridge);
>>>
>>> +/**
>>> + * drm_mode_atomic_add_error_msg - function to add error code and error string
>>> + *
>>> + * @err_code: pointer to struct drm_mode_atomic_err_code that stores the failure
>>> + * reason
>>> + * @failure_code: failure code in enum drm_mode_atomic_failure_codes
>>> + * @failure_string: failure reason string message
>>> + *
>>> + * Returns: void
>>> + */
>>> +static inline void drm_mode_atomic_add_error_msg(struct drm_mode_atomic_err_code *err_code,
>>> + __u64 failure_code,
>>> + char *failure_string)
>> So this should be "const char *format, ..." with printf format, with
>> printf format annotation, to help callers pass in other stuff than just
>> fixed strings.
>>
>> This should be a proper function instead of static inline.
>>
>>> +{
>>> + err_code->failure_code = failure_code;
>>> + strscpy_pad(err_code->failure_string, failure_string,
>>> + strlen(err_code->failure_string));
>> If the size is fixed, why do we just silently truncate the string? If
>> the message is for the user, it'll be ugly to emit truncated strings.
>>
>>> +}
>>> +
>>> #endif /* DRM_ATOMIC_H_ */
--
Jani Nikula, Intel
next prev parent reply other threads:[~2025-09-03 7:20 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-02 7:47 [PATCH v4 0/4] User readable error codes on atomic_ioctl failure Arun R Murthy
2025-09-02 7:47 ` [PATCH v4 1/4] drm: Define user readable error codes for atomic ioctl Arun R Murthy
2025-09-02 10:06 ` Jani Nikula
2025-09-03 6:42 ` Murthy, Arun R
2025-09-02 7:47 ` [PATCH v4 2/4] drm/atomic: Add error_code element in atomic_state Arun R Murthy
2025-09-02 7:47 ` [PATCH v4 3/4] drm/atomic: Return user readable error in atomic_ioctl Arun R Murthy
2025-09-02 10:16 ` Jani Nikula
2025-09-03 6:52 ` Murthy, Arun R
2025-09-03 7:20 ` Jani Nikula [this message]
2025-09-02 7:47 ` [PATCH v4 4/4] drm/i915/display: Error codes for async flip failures Arun R Murthy
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=bc6dfb6765b153901c3af5b1c008c8d82f99bdc2@intel.com \
--to=jani.nikula@linux.intel.com \
--cc=airlied@gmail.com \
--cc=arun.r.murthy@intel.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=harry.wentland@amd.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=joonas.lahtinen@linux.intel.com \
--cc=maarten.lankhorst@linux.intel.com \
--cc=mripard@kernel.org \
--cc=rodrigo.vivi@intel.com \
--cc=simona@ffwll.ch \
--cc=tursulin@ursulin.net \
--cc=tzimmermann@suse.de \
--cc=uma.shankar@intel.com \
--cc=xaver.hugl@kde.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).