From: "Michel Dänzer" <michel@daenzer.net>
To: Jason Ekstrand <jason@jlekstrand.net>
Cc: Maling list - DRI developers <dri-devel@lists.freedesktop.org>
Subject: Re: [PATCH 3/3] dma-buf: Add an API for exporting sync files (v6)
Date: Tue, 16 Mar 2021 09:51:04 +0100 [thread overview]
Message-ID: <82b40177-74c2-a692-499f-54b64c523ce7@daenzer.net> (raw)
In-Reply-To: <CAOFGe96CDjucaqppf8a2pnw8N2-deTnHEQ_oyq5irRL-7=2NLw@mail.gmail.com>
On 2021-03-16 12:10 a.m., Jason Ekstrand wrote:
> On Mon, Mar 15, 2021 at 4:05 PM Jason Ekstrand <jason@jlekstrand.net> wrote:
>>
>> Modern userspace APIs like Vulkan are built on an explicit
>> synchronization model. This doesn't always play nicely with the
>> implicit synchronization used in the kernel and assumed by X11 and
>> Wayland. The client -> compositor half of the synchronization isn't too
>> bad, at least on intel, because we can control whether or not i915
>> synchronizes on the buffer and whether or not it's considered written.
>>
>> The harder part is the compositor -> client synchronization when we get
>> the buffer back from the compositor. We're required to be able to
>> provide the client with a VkSemaphore and VkFence representing the point
>> in time where the window system (compositor and/or display) finished
>> using the buffer. With current APIs, it's very hard to do this in such
>> a way that we don't get confused by the Vulkan driver's access of the
>> buffer. In particular, once we tell the kernel that we're rendering to
>> the buffer again, any CPU waits on the buffer or GPU dependencies will
>> wait on some of the client rendering and not just the compositor.
>>
>> This new IOCTL solves this problem by allowing us to get a snapshot of
>> the implicit synchronization state of a given dma-buf in the form of a
>> sync file. It's effectively the same as a poll() or I915_GEM_WAIT only,
>> instead of CPU waiting directly, it encapsulates the wait operation, at
>> the current moment in time, in a sync_file so we can check/wait on it
>> later. As long as the Vulkan driver does the sync_file export from the
>> dma-buf before we re-introduce it for rendering, it will only contain
>> fences from the compositor or display. This allows to accurately turn
>> it into a VkFence or VkSemaphore without any over- synchronization.
>>
>> v2 (Jason Ekstrand):
>> - Use a wrapper dma_fence_array of all fences including the new one
>> when importing an exclusive fence.
>>
>> v3 (Jason Ekstrand):
>> - Lock around setting shared fences as well as exclusive
>> - Mark SIGNAL_SYNC_FILE as a read-write ioctl.
>> - Initialize ret to 0 in dma_buf_wait_sync_file
>>
>> v4 (Jason Ekstrand):
>> - Use the new dma_resv_get_singleton helper
>>
>> v5 (Jason Ekstrand):
>> - Rename the IOCTLs to import/export rather than wait/signal
>> - Drop the WRITE flag and always get/set the exclusive fence
>>
>> v6 (Jason Ekstrand):
>> - Drop the sync_file import as it was all-around sketchy and not nearly
>> as useful as import.
>> - Re-introduce READ/WRITE flag support for export
>> - Rework the commit message
>>
>> Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
>> ---
>> drivers/dma-buf/dma-buf.c | 55 ++++++++++++++++++++++++++++++++++++
>> include/uapi/linux/dma-buf.h | 6 ++++
>> 2 files changed, 61 insertions(+)
>>
>> diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
>> index f264b70c383eb..e7f9dd62c19a9 100644
>> --- a/drivers/dma-buf/dma-buf.c
>> +++ b/drivers/dma-buf/dma-buf.c
[...]
>> @@ -362,6 +363,57 @@ static long dma_buf_set_name(struct dma_buf *dmabuf, const char __user *buf)
>> return ret;
>> }
>>
>> +static long dma_buf_export_sync_file(struct dma_buf *dmabuf,
>> + void __user *user_data)
>> +{
>> + struct dma_buf_sync_file arg;
>> + struct dma_fence *fence = NULL;
>> + struct sync_file *sync_file;
>> + int fd, ret;
>> +
>> + if (copy_from_user(&arg, user_data, sizeof(arg)))
>> + return -EFAULT;
>> +
>> + if (arg.flags & ~DMA_BUF_SYNC_RW)
>> + return -EINVAL;
>> +
>> + fd = get_unused_fd_flags(O_CLOEXEC);
>> + if (fd < 0)
>> + return fd;
>> +
>> + if (arg.flags & DMA_BUF_SYNC_WRITE) {
>> + ret = dma_resv_get_singleton(dmabuf->resv, NULL, &fence);
>> + if (ret)
>> + goto err_put_fd;
>> + } else if (arg.flags & DMA_BUF_SYNC_READ) {
>> + fence = dma_resv_get_excl(dmabuf->resv);
>> + }
>> +
>> + if (!fence)
>> + fence = dma_fence_get_stub();
>> +
>> + sync_file = sync_file_create(fence);
>> +
>> + dma_fence_put(fence);
>> +
>> + if (!sync_file) {
>> + ret = -EINVAL;
>
> Should this be -EINVAL or -ENOMEM?
The latter makes more sense to me, since sync_file_create returning NULL is not related to invalid ioctl parameters.
--
Earthling Michel Dänzer | https://redhat.com
Libre software enthusiast | Mesa and X developer
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2021-03-16 8:51 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-02-25 23:58 [PATCH] RFC: dma-buf: Add an API for importing and exporting sync files Jason Ekstrand
2020-02-26 9:16 ` Christian König
2020-02-26 10:05 ` Daniel Vetter
2020-02-26 15:28 ` Jason Ekstrand
2020-02-26 16:46 ` Bas Nieuwenhuizen
2020-02-27 8:28 ` Christian König
2020-03-03 19:10 ` Jason Ekstrand
2020-03-04 8:34 ` Christian König
2020-03-04 16:27 ` Jason Ekstrand
2020-03-04 16:41 ` Jason Ekstrand
2020-03-05 13:06 ` Christian König
2020-03-05 15:54 ` Jason Ekstrand
2020-03-09 16:21 ` Christian König
2020-03-11 3:43 ` Jason Ekstrand
2020-02-26 18:09 ` [PATCH] RFC: dma-buf: Add an API for importing and exporting sync files (v2) Jason Ekstrand
2020-03-03 19:03 ` [PATCH] RFC: dma-buf: Add an API for importing and exporting sync files (v3) Jason Ekstrand
2020-03-03 19:05 ` Jason Ekstrand
2020-03-11 3:43 ` [PATCH 1/3] dma-buf: add dma_fence_array_for_each (v2) Jason Ekstrand
2020-03-11 3:43 ` [PATCH 2/3] dma-buf: add dma_resv_get_singleton (v2) Jason Ekstrand
2020-03-11 3:43 ` [PATCH 3/3] RFC: dma-buf: Add an API for importing and exporting sync files (v4) Jason Ekstrand
2020-03-11 13:18 ` Christian König
2020-03-12 15:57 ` Jason Ekstrand
2020-03-13 10:33 ` Christian König
2020-03-17 21:21 ` [PATCH 3/3] RFC: dma-buf: Add an API for importing and exporting sync files (v5) Jason Ekstrand
2020-09-30 9:39 ` Michel Dänzer
2020-09-30 9:55 ` Daniel Vetter
2021-03-15 21:11 ` Jason Ekstrand
2021-03-15 21:30 ` Daniel Vetter
2021-03-15 21:04 ` [PATCH 0/3] dma-buf: Add an API for exporting sync files (v6) Jason Ekstrand
2021-03-15 21:04 ` [PATCH 1/3] dma-buf: add dma_fence_array_for_each (v2) Jason Ekstrand
2021-03-15 21:04 ` [PATCH 2/3] dma-buf: add dma_resv_get_singleton (v2) Jason Ekstrand
2021-03-15 21:04 ` [PATCH 3/3] dma-buf: Add an API for exporting sync files (v6) Jason Ekstrand
2021-03-15 23:10 ` Jason Ekstrand
2021-03-16 8:51 ` Michel Dänzer [this message]
2021-03-16 14:35 ` Jason Ekstrand
2021-03-16 0:10 ` kernel test robot
2021-03-16 2:37 ` kernel test robot
2021-03-16 0:15 ` [PATCH 0/3] " Jason Ekstrand
2021-03-16 4:53 ` [PATCH 0/3] dma-buf: Add an API for exporting sync files (v7) Jason Ekstrand
2021-03-16 4:53 ` [PATCH 1/3] dma-buf: add dma_fence_array_for_each (v2) Jason Ekstrand
2021-03-16 4:53 ` [PATCH 2/3] dma-buf: add dma_resv_get_singleton_rcu (v3) Jason Ekstrand
2021-03-16 4:53 ` [PATCH 3/3] dma-buf: Add an API for exporting sync files (v7) Jason Ekstrand
2021-03-16 8:06 ` kernel test robot
2021-03-17 22:19 ` [PATCH 0/3] dma-buf: Add an API for exporting sync files (v8) Jason Ekstrand
2021-03-17 22:19 ` [PATCH 1/3] dma-buf: add dma_fence_array_for_each (v2) Jason Ekstrand
2021-03-18 9:38 ` Christian König
2021-03-18 13:13 ` Daniel Vetter
2021-03-19 1:40 ` Jason Ekstrand
2021-03-17 22:19 ` [PATCH 2/3] dma-buf: add dma_resv_get_singleton_rcu (v3) Jason Ekstrand
2021-03-17 22:19 ` [PATCH 3/3] dma-buf: Add an API for exporting sync files (v8) Jason Ekstrand
2021-03-23 17:57 ` Jason Ekstrand
2021-03-23 19:06 ` Simon Ser
2021-03-23 19:34 ` Jason Ekstrand
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=82b40177-74c2-a692-499f-54b64c523ce7@daenzer.net \
--to=michel@daenzer.net \
--cc=dri-devel@lists.freedesktop.org \
--cc=jason@jlekstrand.net \
/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).