2018-04-11 03:39:48

by Laura Abbott

[permalink] [raw]
Subject: [PATCHv2] drm/amdkfd: Remove vla

There's an ongoing effort to remove VLAs[1] from the kernel to eventually
turn on -Wvla. Switch to a constant value that covers all hardware.

[1] https://lkml.org/lkml/2018/3/7/621

Signed-off-by: Laura Abbott <[email protected]>
---
v2: Switch to a larger size to account for other hardware
---
drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
index 035c351f47c5..c3a5a80e31ae 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
@@ -139,10 +139,12 @@ static void interrupt_wq(struct work_struct *work)
{
struct kfd_dev *dev = container_of(work, struct kfd_dev,
interrupt_work);
+ uint32_t ih_ring_entry[8];

- uint32_t ih_ring_entry[DIV_ROUND_UP(
- dev->device_info->ih_ring_entry_size,
- sizeof(uint32_t))];
+ if (dev->device_info->ih_ring_entry_size > (8 * sizeof(uint32_t))) {
+ dev_err(kfd_chardev(), "Ring entry too small\n");
+ return;
+ }

while (dequeue_ih_ring_entry(dev, ih_ring_entry))
dev->device_info->event_interrupt_class->interrupt_wq(dev,
--
2.14.3



2018-04-11 09:35:16

by Christian König

[permalink] [raw]
Subject: Re: [PATCHv2] drm/amdkfd: Remove vla

Am 11.04.2018 um 03:02 schrieb Laura Abbott:
> There's an ongoing effort to remove VLAs[1] from the kernel to eventually
> turn on -Wvla. Switch to a constant value that covers all hardware.
>
> [1] https://lkml.org/lkml/2018/3/7/621
>
> Signed-off-by: Laura Abbott <[email protected]>

It would be nicer to have a define for that somewhere.

Anyway the patch is Acked-by: Christian König <[email protected]>
for now.

Regards,
Christian.

> ---
> v2: Switch to a larger size to account for other hardware
> ---
> drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
> index 035c351f47c5..c3a5a80e31ae 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
> @@ -139,10 +139,12 @@ static void interrupt_wq(struct work_struct *work)
> {
> struct kfd_dev *dev = container_of(work, struct kfd_dev,
> interrupt_work);
> + uint32_t ih_ring_entry[8];
>
> - uint32_t ih_ring_entry[DIV_ROUND_UP(
> - dev->device_info->ih_ring_entry_size,
> - sizeof(uint32_t))];
> + if (dev->device_info->ih_ring_entry_size > (8 * sizeof(uint32_t))) {
> + dev_err(kfd_chardev(), "Ring entry too small\n");
> + return;
> + }
>
> while (dequeue_ih_ring_entry(dev, ih_ring_entry))
> dev->device_info->event_interrupt_class->interrupt_wq(dev,


2018-04-11 15:35:18

by Felix Kuehling

[permalink] [raw]
Subject: Re: [PATCHv2] drm/amdkfd: Remove vla

On 2018-04-10 09:02 PM, Laura Abbott wrote:
> There's an ongoing effort to remove VLAs[1] from the kernel to eventually
> turn on -Wvla. Switch to a constant value that covers all hardware.
>
> [1] https://lkml.org/lkml/2018/3/7/621
>
> Signed-off-by: Laura Abbott <[email protected]>
> ---
> v2: Switch to a larger size to account for other hardware
> ---
> drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
> index 035c351f47c5..c3a5a80e31ae 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
> @@ -139,10 +139,12 @@ static void interrupt_wq(struct work_struct *work)
> {
> struct kfd_dev *dev = container_of(work, struct kfd_dev,
> interrupt_work);
> + uint32_t ih_ring_entry[8];
>
> - uint32_t ih_ring_entry[DIV_ROUND_UP(
> - dev->device_info->ih_ring_entry_size,
> - sizeof(uint32_t))];
> + if (dev->device_info->ih_ring_entry_size > (8 * sizeof(uint32_t))) {
> + dev_err(kfd_chardev(), "Ring entry too small\n");

When the ring entry size really is too small, this will potentially
flood the kernel log. Maybe a WARN_ONCE would be more helpful. With that
fixed, this patch is Reviewed-by: Felix Kuehling <[email protected]>

Regards,
  Felix

> + return;
> + }
>
> while (dequeue_ih_ring_entry(dev, ih_ring_entry))
> dev->device_info->event_interrupt_class->interrupt_wq(dev,