Re: [PATCH v1 4/6] dma-buf: Acquire wait-wound context on attachment

From: Dmitry Osipenko <dmitry.osipenko@collabora.com>
To: "Christian König" <christian.koenig@amd.com>,
	"David Airlie" <airlied@linux.ie>,
	"Gerd Hoffmann" <kraxel@redhat.com>,
	"Gurchetan Singh" <gurchetansingh@chromium.org>,
	"Chia-I Wu" <olvaffe@gmail.com>,
	"Daniel Vetter" <daniel@ffwll.ch>,
	"Daniel Almeida" <daniel.almeida@collabora.com>,
	"Gert Wollny" <gert.wollny@collabora.com>,
	"Gustavo Padovan" <gustavo.padovan@collabora.com>,
	"Daniel Stone" <daniel@fooishbar.org>,
	"Tomeu Vizoso" <tomeu.vizoso@collabora.com>,
	"Maarten Lankhorst" <maarten.lankhorst@linux.intel.com>,
	"Maxime Ripard" <mripard@kernel.org>,
	"Thomas Zimmermann" <tzimmermann@suse.de>,
	"Rob Clark" <robdclark@gmail.com>,
	"Sumit Semwal" <sumit.semwal@linaro.org>,
	"Pan, Xinhui" <Xinhui.Pan@amd.com>,
	"Thierry Reding" <thierry.reding@gmail.com>,
	"Tomasz Figa" <tfiga@chromium.org>,
	"Marek Szyprowski" <m.szyprowski@samsung.com>,
	"Mauro Carvalho Chehab" <mchehab@kernel.org>,
	"Alex Deucher" <alexander.deucher@amd.com>,
	"Jani Nikula" <jani.nikula@linux.intel.com>,
	"Joonas Lahtinen" <joonas.lahtinen@linux.intel.com>,
	"Rodrigo Vivi" <rodrigo.vivi@intel.com>,
	"Tvrtko Ursulin" <tvrtko.ursulin@linux.intel.com>,
	"Thomas Hellström" <thomas_os@shipmail.org>
Cc: dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org,
	Dmitry Osipenko <digetx@gmail.com>,
	linux-media@vger.kernel.org, linaro-mm-sig@lists.linaro.org,
	amd-gfx@lists.freedesktop.org, intel-gfx@lists.freedesktop.org,
	kernel@collabora.com, virtualization@lists.linux-foundation.org,
	spice-devel@lists.freedesktop.org, linux-rdma@vger.kernel.org,
	linux-arm-msm@vger.kernel.org
Subject: Re: [PATCH v1 4/6] dma-buf: Acquire wait-wound context on attachment
Date: Fri, 15 Jul 2022 09:59:47 +0300	[thread overview]
Message-ID: <1ce233a2-36c9-3698-59f0-c4ff902bec60@collabora.com> (raw)
In-Reply-To: <5ec9313e-8498-2838-0320-331c347ce905@amd.com>

On 7/15/22 09:50, Christian König wrote:
> Am 15.07.22 um 02:52 schrieb Dmitry Osipenko:
>> Intel i915 GPU driver uses wait-wound mutex to lock multiple GEMs on the
>> attachment to the i915 dma-buf. In order to let all drivers utilize
>> shared
>> wait-wound context during attachment in a general way, make dma-buf
>> core to
>> acquire the ww context internally for the attachment operation and update
>> i915 driver to use the importer's ww context instead of the internal one.
>>
>>  From now on all dma-buf exporters shall use the importer's ww context
>> for
>> the attachment operation.
>>
>> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
>> ---
>>   drivers/dma-buf/dma-buf.c                     |  8 +++++-
>>   drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c    |  2 +-
>>   .../gpu/drm/i915/gem/i915_gem_execbuffer.c    |  2 +-
>>   drivers/gpu/drm/i915/gem/i915_gem_object.h    |  6 ++---
>>   drivers/gpu/drm/i915/i915_gem_evict.c         |  2 +-
>>   drivers/gpu/drm/i915/i915_gem_ww.c            | 26 +++++++++++++++----
>>   drivers/gpu/drm/i915/i915_gem_ww.h            | 15 +++++++++--
>>   7 files changed, 47 insertions(+), 14 deletions(-)
>>
>> diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
>> index 0ee588276534..37545ecb845a 100644
>> --- a/drivers/dma-buf/dma-buf.c
>> +++ b/drivers/dma-buf/dma-buf.c
>> @@ -807,6 +807,8 @@ static struct sg_table * __map_dma_buf(struct
>> dma_buf_attachment *attach,
>>    * Optionally this calls &dma_buf_ops.attach to allow
>> device-specific attach
>>    * functionality.
>>    *
>> + * Exporters shall use ww_ctx acquired by this function.
>> + *
>>    * Returns:
>>    *
>>    * A pointer to newly created &dma_buf_attachment on success, or a
>> negative
>> @@ -822,6 +824,7 @@ dma_buf_dynamic_attach_unlocked(struct dma_buf
>> *dmabuf, struct device *dev,
>>                   void *importer_priv)
>>   {
>>       struct dma_buf_attachment *attach;
>> +    struct ww_acquire_ctx ww_ctx;
>>       int ret;
>>         if (WARN_ON(!dmabuf || !dev))
>> @@ -841,7 +844,8 @@ dma_buf_dynamic_attach_unlocked(struct dma_buf
>> *dmabuf, struct device *dev,
>>       attach->importer_ops = importer_ops;
>>       attach->importer_priv = importer_priv;
>>   -    dma_resv_lock(dmabuf->resv, NULL);
>> +    ww_acquire_init(&ww_ctx, &reservation_ww_class);
>> +    dma_resv_lock(dmabuf->resv, &ww_ctx);
> 
> That won't work like this. The core property of a WW context is that you
> need to unwind all the locks and re-quire them with the contended one
> first.
> 
> When you statically lock the imported one here you can't do that any more.

You're right. I felt that something is missing here, but couldn't
notice. I'll think more about this and enable
CONFIG_DEBUG_WW_MUTEX_SLOWPATH. Thank you!

-- 
Best regards,
Dmitry