Print an error message to help debug when such an issue happen, since it's
not so obvious.
Signed-off-by: Sui Jingfeng <sui.jingfeng@linux.dev>
---
v0 -> v1: Use dev_err_ratelimited() to prevent spamming the logs
v0 is at https://lore.kernel.org/dri-devel/20240930221706.399139-1-sui.jingfeng@linux.dev/
---
drivers/gpu/drm/etnaviv/etnaviv_mmu.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
index a52ec5eb0e3d..37866ed05c13 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
@@ -300,8 +300,12 @@ int etnaviv_iommu_map_gem(struct etnaviv_iommu_context *context,
ret = etnaviv_iommu_insert_exact(context, node, user_size, va);
else
ret = etnaviv_iommu_find_iova(context, node, user_size);
- if (ret < 0)
+ if (ret < 0) {
+ dev_err_ratelimited(context->global->dev,
+ "Insert iova failed: 0x%llx(0x%x)\n",
+ va, user_size);
goto unlock;
+ }
mapping->iova = node->start;
ret = etnaviv_iommu_map(context, node->start, user_size, sgt,
--
2.34.1
Am Samstag, dem 05.10.2024 um 03:42 +0800 schrieb Sui Jingfeng: > Print an error message to help debug when such an issue happen, since it's > not so obvious. > > Signed-off-by: Sui Jingfeng <sui.jingfeng@linux.dev> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> > --- > v0 -> v1: Use dev_err_ratelimited() to prevent spamming the logs > > v0 is at https://lore.kernel.org/dri-devel/20240930221706.399139-1-sui.jingfeng@linux.dev/ > --- > drivers/gpu/drm/etnaviv/etnaviv_mmu.c | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c > index a52ec5eb0e3d..37866ed05c13 100644 > --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c > +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c > @@ -300,8 +300,12 @@ int etnaviv_iommu_map_gem(struct etnaviv_iommu_context *context, > ret = etnaviv_iommu_insert_exact(context, node, user_size, va); > else > ret = etnaviv_iommu_find_iova(context, node, user_size); > - if (ret < 0) > + if (ret < 0) { > + dev_err_ratelimited(context->global->dev, > + "Insert iova failed: 0x%llx(0x%x)\n", > + va, user_size); > goto unlock; > + } > > mapping->iova = node->start; > ret = etnaviv_iommu_map(context, node->start, user_size, sgt,
Hi Sui, Am Montag, dem 07.10.2024 um 12:20 +0200 schrieb Lucas Stach: > Am Samstag, dem 05.10.2024 um 03:42 +0800 schrieb Sui Jingfeng: > > Print an error message to help debug when such an issue happen, since it's > > not so obvious. > > > > Signed-off-by: Sui Jingfeng <sui.jingfeng@linux.dev> > > Reviewed-by: Lucas Stach <l.stach@pengutronix.de> > What happened to this patch? It's not part of the updated series anymore. Even though the problem at hand is solved right now, I still think this patch is useful to have. Regards, Lucas > > --- > > v0 -> v1: Use dev_err_ratelimited() to prevent spamming the logs > > > > v0 is at https://lore.kernel.org/dri-devel/20240930221706.399139-1-sui.jingfeng@linux.dev/ > > --- > > drivers/gpu/drm/etnaviv/etnaviv_mmu.c | 6 +++++- > > 1 file changed, 5 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c > > index a52ec5eb0e3d..37866ed05c13 100644 > > --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c > > +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c > > @@ -300,8 +300,12 @@ int etnaviv_iommu_map_gem(struct etnaviv_iommu_context *context, > > ret = etnaviv_iommu_insert_exact(context, node, user_size, va); > > else > > ret = etnaviv_iommu_find_iova(context, node, user_size); > > - if (ret < 0) > > + if (ret < 0) { > > + dev_err_ratelimited(context->global->dev, > > + "Insert iova failed: 0x%llx(0x%x)\n", > > + va, user_size); > > goto unlock; > > + } > > > > mapping->iova = node->start; > > ret = etnaviv_iommu_map(context, node->start, user_size, sgt, >
Hi, On 2024/10/29 00:01, Lucas Stach wrote: > Hi Sui, > > Am Montag, dem 07.10.2024 um 12:20 +0200 schrieb Lucas Stach: >> Am Samstag, dem 05.10.2024 um 03:42 +0800 schrieb Sui Jingfeng: >>> Print an error message to help debug when such an issue happen, since it's >>> not so obvious. >>> >>> Signed-off-by: Sui Jingfeng <sui.jingfeng@linux.dev> >> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> >> > What happened to this patch? It's not part of the updated series > anymore. Even though the problem at hand is solved right now, I still > think this patch is useful to have. I was thinking this patch should be a separate patch while making the v2. In the future, I guess there may have 64K CPU and/or GPU page size combinations from other users or contributors. I think I should resend this. But don't know take which Linux kernel as code base. etnaviv-next or etnaviv-fixes? If you have convenient code base at hand now, you could pick it up, I don't mind. > Regards, > Lucas > >>> --- >>> v0 -> v1: Use dev_err_ratelimited() to prevent spamming the logs >>> >>> v0 is at https://lore.kernel.org/dri-devel/20240930221706.399139-1-sui.jingfeng@linux.dev/ >>> --- >>> drivers/gpu/drm/etnaviv/etnaviv_mmu.c | 6 +++++- >>> 1 file changed, 5 insertions(+), 1 deletion(-) >>> >>> diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c >>> index a52ec5eb0e3d..37866ed05c13 100644 >>> --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c >>> +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c >>> @@ -300,8 +300,12 @@ int etnaviv_iommu_map_gem(struct etnaviv_iommu_context *context, >>> ret = etnaviv_iommu_insert_exact(context, node, user_size, va); >>> else >>> ret = etnaviv_iommu_find_iova(context, node, user_size); >>> - if (ret < 0) >>> + if (ret < 0) { >>> + dev_err_ratelimited(context->global->dev, >>> + "Insert iova failed: 0x%llx(0x%x)\n", >>> + va, user_size); >>> goto unlock; >>> + } >>> >>> mapping->iova = node->start; >>> ret = etnaviv_iommu_map(context, node->start, user_size, sgt, -- Best regards, Sui
© 2016 - 2024 Red Hat, Inc.