media: s5p-mfc: Fix memdev DMA configuration
Commit Message
Having of_reserved_mem_device_init() forcibly reconfigure DMA for all
callers, potentially overriding the work done by a bus-specific
.dma_configure method earlier, is at best a bad idea and at worst
actively harmful. If drivers really need virtual devices to own
dma-coherent memory, they should explicitly configure those devices
based on the appropriate firmware node as they create them.
It looks like the only driver not passing in a proper OF platform device
is s5p-mfc, so move the rogue of_dma_configure() call into that driver
where it logically belongs.
CC: Smitha T Murthy <smitha.t@samsung.com>
CC: Marek Szyprowski <m.szyprowski@samsung.com>
CC: Rob Herring <robh@kernel.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
drivers/media/platform/s5p-mfc/s5p_mfc.c | 7 +++++++
drivers/of/of_reserved_mem.c | 4 ----
2 files changed, 7 insertions(+), 4 deletions(-)
Comments
Hi Robin,
On 2018-09-12 18:45, Robin Murphy wrote:
> Having of_reserved_mem_device_init() forcibly reconfigure DMA for all
> callers, potentially overriding the work done by a bus-specific
> .dma_configure method earlier, is at best a bad idea and at worst
> actively harmful. If drivers really need virtual devices to own
> dma-coherent memory, they should explicitly configure those devices
> based on the appropriate firmware node as they create them.
>
> It looks like the only driver not passing in a proper OF platform device
> is s5p-mfc, so move the rogue of_dma_configure() call into that driver
> where it logically belongs.
>
> CC: Smitha T Murthy <smitha.t@samsung.com>
> CC: Marek Szyprowski <m.szyprowski@samsung.com>
> CC: Rob Herring <robh@kernel.org>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Right, after recent cleanup dma ops initialization, MFC driver is
a better place for calling of_dma_configure() on virtual devices.
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
> ---
> drivers/media/platform/s5p-mfc/s5p_mfc.c | 7 +++++++
> drivers/of/of_reserved_mem.c | 4 ----
> 2 files changed, 7 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/media/platform/s5p-mfc/s5p_mfc.c b/drivers/media/platform/s5p-mfc/s5p_mfc.c
> index 927a1235408d..77eb4a4511c1 100644
> --- a/drivers/media/platform/s5p-mfc/s5p_mfc.c
> +++ b/drivers/media/platform/s5p-mfc/s5p_mfc.c
> @@ -1094,6 +1094,13 @@ static struct device *s5p_mfc_alloc_memdev(struct device *dev,
> child->dma_mask = dev->dma_mask;
> child->release = s5p_mfc_memdev_release;
>
> + /*
> + * The memdevs are not proper OF platform devices, so in order for them
> + * to be treated as valid DMA masters we need a bit of a hack to force
> + * them to inherit the MFC node's DMA configuration.
> + */
> + of_dma_configure(child, dev->of_node, true);
> +
> if (device_add(child) == 0) {
> ret = of_reserved_mem_device_init_by_idx(child, dev->of_node,
> idx);
> diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> index 895c83e0c7b6..4ef6f4485335 100644
> --- a/drivers/of/of_reserved_mem.c
> +++ b/drivers/of/of_reserved_mem.c
> @@ -350,10 +350,6 @@ int of_reserved_mem_device_init_by_idx(struct device *dev,
> mutex_lock(&of_rmem_assigned_device_mutex);
> list_add(&rd->list, &of_rmem_assigned_device_list);
> mutex_unlock(&of_rmem_assigned_device_mutex);
> - /* ensure that dma_ops is set for virtual devices
> - * using reserved memory
> - */
> - of_dma_configure(dev, np, true);
>
> dev_info(dev, "assigned reserved memory node %s\n", rmem->name);
> } else {
Best regards
Em Fri, 14 Sep 2018 14:19:29 +0200
Marek Szyprowski <m.szyprowski@samsung.com> escreveu:
> Hi Robin,
>
> On 2018-09-12 18:45, Robin Murphy wrote:
> > Having of_reserved_mem_device_init() forcibly reconfigure DMA for all
> > callers, potentially overriding the work done by a bus-specific
> > .dma_configure method earlier, is at best a bad idea and at worst
> > actively harmful. If drivers really need virtual devices to own
> > dma-coherent memory, they should explicitly configure those devices
> > based on the appropriate firmware node as they create them.
> >
> > It looks like the only driver not passing in a proper OF platform device
> > is s5p-mfc, so move the rogue of_dma_configure() call into that driver
> > where it logically belongs.
> >
> > CC: Smitha T Murthy <smitha.t@samsung.com>
> > CC: Marek Szyprowski <m.szyprowski@samsung.com>
> > CC: Rob Herring <robh@kernel.org>
> > Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>
> Right, after recent cleanup dma ops initialization, MFC driver is
> a better place for calling of_dma_configure() on virtual devices.
>
> Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
This patch seems to fit better via OF/DT tree. Not sure if it was
merged there yet. In any case:
Acked-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
>
> > ---
> > drivers/media/platform/s5p-mfc/s5p_mfc.c | 7 +++++++
> > drivers/of/of_reserved_mem.c | 4 ----
> > 2 files changed, 7 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/media/platform/s5p-mfc/s5p_mfc.c b/drivers/media/platform/s5p-mfc/s5p_mfc.c
> > index 927a1235408d..77eb4a4511c1 100644
> > --- a/drivers/media/platform/s5p-mfc/s5p_mfc.c
> > +++ b/drivers/media/platform/s5p-mfc/s5p_mfc.c
> > @@ -1094,6 +1094,13 @@ static struct device *s5p_mfc_alloc_memdev(struct device *dev,
> > child->dma_mask = dev->dma_mask;
> > child->release = s5p_mfc_memdev_release;
> >
> > + /*
> > + * The memdevs are not proper OF platform devices, so in order for them
> > + * to be treated as valid DMA masters we need a bit of a hack to force
> > + * them to inherit the MFC node's DMA configuration.
> > + */
> > + of_dma_configure(child, dev->of_node, true);
> > +
> > if (device_add(child) == 0) {
> > ret = of_reserved_mem_device_init_by_idx(child, dev->of_node,
> > idx);
> > diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> > index 895c83e0c7b6..4ef6f4485335 100644
> > --- a/drivers/of/of_reserved_mem.c
> > +++ b/drivers/of/of_reserved_mem.c
> > @@ -350,10 +350,6 @@ int of_reserved_mem_device_init_by_idx(struct device *dev,
> > mutex_lock(&of_rmem_assigned_device_mutex);
> > list_add(&rd->list, &of_rmem_assigned_device_list);
> > mutex_unlock(&of_rmem_assigned_device_mutex);
> > - /* ensure that dma_ops is set for virtual devices
> > - * using reserved memory
> > - */
> > - of_dma_configure(dev, np, true);
> >
> > dev_info(dev, "assigned reserved memory node %s\n", rmem->name);
> > } else {
>
> Best regards
Thanks,
Mauro
@@ -1094,6 +1094,13 @@ static struct device *s5p_mfc_alloc_memdev(struct device *dev,
child->dma_mask = dev->dma_mask;
child->release = s5p_mfc_memdev_release;
+ /*
+ * The memdevs are not proper OF platform devices, so in order for them
+ * to be treated as valid DMA masters we need a bit of a hack to force
+ * them to inherit the MFC node's DMA configuration.
+ */
+ of_dma_configure(child, dev->of_node, true);
+
if (device_add(child) == 0) {
ret = of_reserved_mem_device_init_by_idx(child, dev->of_node,
idx);
@@ -350,10 +350,6 @@ int of_reserved_mem_device_init_by_idx(struct device *dev,
mutex_lock(&of_rmem_assigned_device_mutex);
list_add(&rd->list, &of_rmem_assigned_device_list);
mutex_unlock(&of_rmem_assigned_device_mutex);
- /* ensure that dma_ops is set for virtual devices
- * using reserved memory
- */
- of_dma_configure(dev, np, true);
dev_info(dev, "assigned reserved memory node %s\n", rmem->name);
} else {