From: "Maíra Canal" <mcanal@igalia.com>
To: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>,
dri-devel@lists.freedesktop.org
Cc: kernel-dev@igalia.com
Subject: Re: [PATCH 5/7] drm/syncobj: Avoid temporary allocation in drm_syncobj_timeline_signal_ioctl
Date: Mon, 24 Mar 2025 19:50:20 -0300 [thread overview]
Message-ID: <d7c59ab6-fd05-4ce6-b950-07e5d1a52ad2@igalia.com> (raw)
In-Reply-To: <20250318155424.78552-6-tvrtko.ursulin@igalia.com>
Hi Tvrtko,
On 18/03/25 12:54, Tvrtko Ursulin wrote:
> We can avoid one of the two temporary allocations if we read the userspace
> supplied timeline points as we go along.
>
> The only new complication is to unwind unused fence chains on the error
> path, but even that code was already present in the function.
>
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Best Regards,
- Maíra
> ---
> drivers/gpu/drm/drm_syncobj.c | 46 +++++++++++++++--------------------
> 1 file changed, 20 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c
> index 74d1dc0d1f8b..b4563c696056 100644
> --- a/drivers/gpu/drm/drm_syncobj.c
> +++ b/drivers/gpu/drm/drm_syncobj.c
> @@ -1557,10 +1557,10 @@ drm_syncobj_timeline_signal_ioctl(struct drm_device *dev, void *data,
> struct drm_file *file_private)
> {
> struct drm_syncobj_timeline_array *args = data;
> + uint64_t __user *points = u64_to_user_ptr(args->points);
> + uint32_t i, j, count = args->count_handles;
> struct drm_syncobj **syncobjs;
> struct dma_fence_chain **chains;
> - uint64_t *points;
> - uint32_t i, j;
> int ret;
>
> if (!drm_core_check_feature(dev, DRIVER_SYNCOBJ_TIMELINE))
> @@ -1572,33 +1572,22 @@ drm_syncobj_timeline_signal_ioctl(struct drm_device *dev, void *data,
> if (args->count_handles == 0)
> return -EINVAL;
>
> + if (!access_ok(points, count * sizeof(*points)))
> + return -EFAULT;
> +
> ret = drm_syncobj_array_find(file_private,
> u64_to_user_ptr(args->handles),
> - args->count_handles,
> + count,
> &syncobjs);
> if (ret < 0)
> return ret;
>
> - points = kmalloc_array(args->count_handles, sizeof(*points),
> - GFP_KERNEL);
> - if (!points) {
> - ret = -ENOMEM;
> - goto out;
> - }
> - if (!u64_to_user_ptr(args->points)) {
> - memset(points, 0, args->count_handles * sizeof(uint64_t));
> - } else if (copy_from_user(points, u64_to_user_ptr(args->points),
> - sizeof(uint64_t) * args->count_handles)) {
> - ret = -EFAULT;
> - goto err_points;
> - }
> -
> - chains = kmalloc_array(args->count_handles, sizeof(void *), GFP_KERNEL);
> + chains = kmalloc_array(count, sizeof(void *), GFP_KERNEL);
> if (!chains) {
> ret = -ENOMEM;
> - goto err_points;
> + goto out;
> }
> - for (i = 0; i < args->count_handles; i++) {
> + for (i = 0; i < count; i++) {
> chains[i] = dma_fence_chain_alloc();
> if (!chains[i]) {
> for (j = 0; j < i; j++)
> @@ -1608,19 +1597,24 @@ drm_syncobj_timeline_signal_ioctl(struct drm_device *dev, void *data,
> }
> }
>
> - for (i = 0; i < args->count_handles; i++) {
> + for (i = 0; i < count; i++) {
> struct dma_fence *fence = dma_fence_get_stub();
> + u64 point = 0;
>
> - drm_syncobj_add_point(syncobjs[i], chains[i],
> - fence, points[i]);
> + if (points && __get_user(point, points++)) {
> + ret = -EFAULT;
> + for (j = i; j < count; j++)
> + dma_fence_chain_free(chains[j]);
> + goto err_chains;
> + }
> +
> + drm_syncobj_add_point(syncobjs[i], chains[i], fence, point);
> dma_fence_put(fence);> }
> err_chains:
> kfree(chains);
> -err_points:
> - kfree(points);
> out:
> - drm_syncobj_array_free(syncobjs, args->count_handles);
> + drm_syncobj_array_free(syncobjs, count);
>
> return ret;
> }
next prev parent reply other threads:[~2025-03-24 22:50 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-18 15:54 [PATCH 0/7] A few drm_syncobj optimisations Tvrtko Ursulin
2025-03-18 15:54 ` [PATCH 1/7] drm/syncobj: Remove unhelpful helper Tvrtko Ursulin
2025-03-24 21:38 ` Maíra Canal
2025-03-25 9:12 ` Tvrtko Ursulin
2025-03-18 15:54 ` [PATCH 2/7] drm/syncobj: Do not allocate an array to store zeros when waiting Tvrtko Ursulin
2025-03-24 21:55 ` Maíra Canal
2025-03-18 15:54 ` [PATCH 3/7] drm/syncobj: Avoid one temporary allocation in drm_syncobj_array_find Tvrtko Ursulin
2025-03-24 22:29 ` Maíra Canal
2025-03-25 9:26 ` Tvrtko Ursulin
2025-03-18 15:54 ` [PATCH 4/7] drm/syncobj: Use put_user in drm_syncobj_query_ioctl Tvrtko Ursulin
2025-03-24 22:32 ` Maíra Canal
2025-03-18 15:54 ` [PATCH 5/7] drm/syncobj: Avoid temporary allocation in drm_syncobj_timeline_signal_ioctl Tvrtko Ursulin
2025-03-24 22:50 ` Maíra Canal [this message]
2025-03-18 15:54 ` [PATCH 6/7] drm/syncobj: Add a fast path to drm_syncobj_array_wait_timeout Tvrtko Ursulin
2025-03-24 23:00 ` Maíra Canal
2025-03-25 9:33 ` Tvrtko Ursulin
2025-03-25 19:56 ` Maíra Canal
2025-03-18 15:54 ` [PATCH 7/7] drm/syncobj: Add a fast path to drm_syncobj_array_find Tvrtko Ursulin
2025-03-24 23:06 ` Maíra Canal
2025-03-25 9:54 ` Tvrtko Ursulin
2025-03-25 20:12 ` Maíra Canal
2025-03-24 23:17 ` [PATCH 0/7] A few drm_syncobj optimisations Maíra Canal
2025-03-25 9:57 ` Tvrtko Ursulin
2025-03-25 21:10 ` Maíra Canal
-- strict thread matches above, loose matches on Subject: below --
2025-01-06 10:55 [PATCH v2 " Tvrtko Ursulin
2025-01-06 10:55 ` [PATCH 5/7] drm/syncobj: Avoid temporary allocation in drm_syncobj_timeline_signal_ioctl Tvrtko Ursulin
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=d7c59ab6-fd05-4ce6-b950-07e5d1a52ad2@igalia.com \
--to=mcanal@igalia.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=kernel-dev@igalia.com \
--cc=tvrtko.ursulin@igalia.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;
as well as URLs for NNTP newsgroup(s).