From: Thomas Zimmermann <tzimmermann@suse.de>
To: maarten.lankhorst@linux.intel.com, mripard@kernel.org,
airlied@linux.ie, daniel@ffwll.ch, sean@poorly.run,
kraxel@redhat.com, emil.l.velikov@gmail.com, sam@ravnborg.org
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [PATCH 1/2] drm/shmem: Use cached mappings by default
Date: Wed, 13 May 2020 17:06:08 +0200 [thread overview]
Message-ID: <f2bccb30-366a-9822-d6ed-345a96847735@suse.de> (raw)
In-Reply-To: <20200513150312.21421-2-tzimmermann@suse.de>
[-- Attachment #1.1.1: Type: text/plain, Size: 3451 bytes --]
Am 13.05.20 um 17:03 schrieb Thomas Zimmermann:
> SHMEM-buffer backing storage is allocated from system memory; which is
> typically cachable. Currently, only virtio uses cachable mappings; udl
> uses its own vmap/mmap implementation for cachable mappings. Other
> drivers default to writecombine mappings.
>
> Use cached mappings by default. The exception is pages imported via
> dma-buf. DMA memory is usually not cached.
>
> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
> ---
> drivers/gpu/drm/drm_gem_shmem_helper.c | 6 ++++--
> drivers/gpu/drm/virtio/virtgpu_object.c | 1 -
> include/drm/drm_gem_shmem_helper.h | 4 ++--
> 3 files changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_gem_shmem_helper.c
> index df31e5782eed1..1ce90325dfa31 100644
> --- a/drivers/gpu/drm/drm_gem_shmem_helper.c
> +++ b/drivers/gpu/drm/drm_gem_shmem_helper.c
> @@ -259,7 +259,7 @@ static void *drm_gem_shmem_vmap_locked(struct drm_gem_shmem_object *shmem)
> } else {
> pgprot_t prot = PAGE_KERNEL;
>
> - if (!shmem->map_cached)
> + if (shmem->map_wc)
> prot = pgprot_writecombine(prot);
> shmem->vaddr = vmap(shmem->pages, obj->size >> PAGE_SHIFT,
> VM_MAP, prot);
> @@ -546,7 +546,7 @@ int drm_gem_shmem_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
>
> vma->vm_flags |= VM_MIXEDMAP | VM_DONTEXPAND;
> vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
> - if (!shmem->map_cached)
> + if (shmem->map_wc)
> vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
> vma->vm_ops = &drm_gem_shmem_vm_ops;
>
> @@ -664,6 +664,8 @@ drm_gem_shmem_prime_import_sg_table(struct drm_device *dev,
> if (IS_ERR(shmem))
> return ERR_CAST(shmem);
>
> + shmem->map_wc = false; /* dma-buf mappings use writecombine */
Just when I posted the patch, I saw this bug. map_wc should be set to
true to enable writecombine mappings for dma-buf pages. Will be fixed in
the next iteration.
> +
> shmem->pages = kvmalloc_array(npages, sizeof(struct page *), GFP_KERNEL);
> if (!shmem->pages) {
> ret = -ENOMEM;
> diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
> index 6ccbd01cd888c..80ba6b2b61668 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_object.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_object.c
> @@ -132,7 +132,6 @@ struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev,
>
> dshmem = &shmem->base.base;
> dshmem->base.funcs = &virtio_gpu_shmem_funcs;
> - dshmem->map_cached = true;
> return &dshmem->base;
> }
>
> diff --git a/include/drm/drm_gem_shmem_helper.h b/include/drm/drm_gem_shmem_helper.h
> index 294b2931c4cc0..a5bc082a77c48 100644
> --- a/include/drm/drm_gem_shmem_helper.h
> +++ b/include/drm/drm_gem_shmem_helper.h
> @@ -98,9 +98,9 @@ struct drm_gem_shmem_object {
> unsigned int vmap_use_count;
>
> /**
> - * @map_cached: map object cached (instead of using writecombine).
> + * @map_wc: map object using writecombine (instead of cached).
> */
> - bool map_cached;
> + bool map_wc;
> };
>
> #define to_drm_gem_shmem_obj(obj) \
>
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Felix Imendörffer
[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]
[-- Attachment #2: Type: text/plain, Size: 160 bytes --]
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2020-05-13 15:07 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-13 15:03 [PATCH 0/2] Default to cachable mappings for GEM SHMEM Thomas Zimmermann
2020-05-13 15:03 ` [PATCH 1/2] drm/shmem: Use cached mappings by default Thomas Zimmermann
2020-05-13 15:06 ` Thomas Zimmermann [this message]
2020-05-14 12:40 ` Daniel Vetter
2020-05-14 15:27 ` Thomas Zimmermann
2020-05-14 20:36 ` Rob Herring
2020-05-15 6:58 ` Thomas Zimmermann
2020-05-15 14:10 ` Daniel Vetter
2020-05-18 8:13 ` Thomas Zimmermann
2020-05-18 8:23 ` Gerd Hoffmann
2020-05-18 8:50 ` Thomas Zimmermann
2020-05-18 10:11 ` Gerd Hoffmann
2020-05-18 14:40 ` Daniel Vetter
2020-05-19 6:31 ` Thomas Zimmermann
2020-05-13 15:03 ` [PATCH 2/2] drm/udl: Use GEM vmap/mmap function from SHMEM helpers Thomas Zimmermann
2020-05-13 15:49 ` Daniel Vetter
2020-05-13 17:19 ` Thomas Zimmermann
2020-05-14 12:44 ` Daniel Vetter
2020-05-19 6:36 ` Thomas Zimmermann
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=f2bccb30-366a-9822-d6ed-345a96847735@suse.de \
--to=tzimmermann@suse.de \
--cc=airlied@linux.ie \
--cc=daniel@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=emil.l.velikov@gmail.com \
--cc=kraxel@redhat.com \
--cc=maarten.lankhorst@linux.intel.com \
--cc=mripard@kernel.org \
--cc=sam@ravnborg.org \
--cc=sean@poorly.run \
/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