* Re: [PATCH] ACPI: scan: Fix DMA range assignment
2022-10-18 13:14 [PATCH] ACPI: scan: Fix DMA range assignment Robin Murphy
@ 2022-10-18 13:31 ` Jianmin Lv
2022-10-19 0:41 ` Jeremy Linton
` (2 subsequent siblings)
3 siblings, 0 replies; 6+ messages in thread
From: Jianmin Lv @ 2022-10-18 13:31 UTC (permalink / raw)
To: Robin Murphy, rafael
Cc: yangyicong, lpieralisi, chenhuacai, linux-acpi, linux-kernel,
lenb, jeremy.linton
Seems good. Thanks very much.
Reviewed-by: Jianmin Lv <lvjianmin@loongson.cn>
On 2022/10/18 下午9:14, Robin Murphy wrote:
> Assigning the device's dma_range_map from the iterator variable after
> the loop means it always points to the empty terminator at the end of
> the map, which is not what we want. Similarly, freeing the iterator on
> error when it points to somwhere in the middle of the allocated array
> won't work either. Fix this.
>
> Fixes: bf2ee8d0c385 ("ACPI: scan: Support multiple DMA windows with different offsets")
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/acpi/scan.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> index 558664d169fc..024cc373a197 100644
> --- a/drivers/acpi/scan.c
> +++ b/drivers/acpi/scan.c
> @@ -1509,9 +1509,12 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> goto out;
> }
>
> + *map = r;
> +
> list_for_each_entry(rentry, &list, node) {
> if (rentry->res->start >= rentry->res->end) {
> - kfree(r);
> + kfree(*map);
> + *map = NULL;
> ret = -EINVAL;
> dev_dbg(dma_dev, "Invalid DMA regions configuration\n");
> goto out;
> @@ -1523,8 +1526,6 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> r->offset = rentry->offset;
> r++;
> }
> -
> - *map = r;
> }
> out:
> acpi_dev_free_resource_list(&list);
>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH] ACPI: scan: Fix DMA range assignment
2022-10-18 13:14 [PATCH] ACPI: scan: Fix DMA range assignment Robin Murphy
2022-10-18 13:31 ` Jianmin Lv
@ 2022-10-19 0:41 ` Jeremy Linton
2022-10-19 2:07 ` Yicong Yang
2022-10-19 8:13 ` Lorenzo Pieralisi
3 siblings, 0 replies; 6+ messages in thread
From: Jeremy Linton @ 2022-10-19 0:41 UTC (permalink / raw)
To: Robin Murphy, rafael
Cc: lvjianmin, yangyicong, lpieralisi, chenhuacai, linux-acpi,
linux-kernel, lenb
Hi,
On 10/18/22 08:14, Robin Murphy wrote:
> Assigning the device's dma_range_map from the iterator variable after
> the loop means it always points to the empty terminator at the end of
> the map, which is not what we want. Similarly, freeing the iterator on
> error when it points to somwhere in the middle of the allocated array
> won't work either. Fix this.
This fixes the boot problem on both SoC generations of the rpi4+ACPI,
Thanks,
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
>
> Fixes: bf2ee8d0c385 ("ACPI: scan: Support multiple DMA windows with different offsets")
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/acpi/scan.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> index 558664d169fc..024cc373a197 100644
> --- a/drivers/acpi/scan.c
> +++ b/drivers/acpi/scan.c
> @@ -1509,9 +1509,12 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> goto out;
> }
>
> + *map = r;
> +
> list_for_each_entry(rentry, &list, node) {
> if (rentry->res->start >= rentry->res->end) {
> - kfree(r);
> + kfree(*map);
> + *map = NULL;
> ret = -EINVAL;
> dev_dbg(dma_dev, "Invalid DMA regions configuration\n");
> goto out;
> @@ -1523,8 +1526,6 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> r->offset = rentry->offset;
> r++;
> }
> -
> - *map = r;
> }
> out:
> acpi_dev_free_resource_list(&list);
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH] ACPI: scan: Fix DMA range assignment
2022-10-18 13:14 [PATCH] ACPI: scan: Fix DMA range assignment Robin Murphy
2022-10-18 13:31 ` Jianmin Lv
2022-10-19 0:41 ` Jeremy Linton
@ 2022-10-19 2:07 ` Yicong Yang
2022-10-19 8:13 ` Lorenzo Pieralisi
3 siblings, 0 replies; 6+ messages in thread
From: Yicong Yang @ 2022-10-19 2:07 UTC (permalink / raw)
To: Robin Murphy, rafael
Cc: yangyicong, lvjianmin, lpieralisi, chenhuacai, linux-acpi,
linux-kernel, lenb, jeremy.linton
On 2022/10/18 21:14, Robin Murphy wrote:
> Assigning the device's dma_range_map from the iterator variable after
> the loop means it always points to the empty terminator at the end of
> the map, which is not what we want. Similarly, freeing the iterator on
> error when it points to somwhere in the middle of the allocated array
> won't work either. Fix this.
>
> Fixes: bf2ee8d0c385 ("ACPI: scan: Support multiple DMA windows with different offsets")
Thanks for fixing this. Works on my platform.
Tested-by: Yicong Yang <yangyicong@hisilicon.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/acpi/scan.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> index 558664d169fc..024cc373a197 100644
> --- a/drivers/acpi/scan.c
> +++ b/drivers/acpi/scan.c
> @@ -1509,9 +1509,12 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> goto out;
> }
>
> + *map = r;
> +
> list_for_each_entry(rentry, &list, node) {
> if (rentry->res->start >= rentry->res->end) {
> - kfree(r);
> + kfree(*map);
> + *map = NULL;
> ret = -EINVAL;
> dev_dbg(dma_dev, "Invalid DMA regions configuration\n");
> goto out;
> @@ -1523,8 +1526,6 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> r->offset = rentry->offset;
> r++;
> }
> -
> - *map = r;
> }
> out:
> acpi_dev_free_resource_list(&list);
>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH] ACPI: scan: Fix DMA range assignment
2022-10-18 13:14 [PATCH] ACPI: scan: Fix DMA range assignment Robin Murphy
` (2 preceding siblings ...)
2022-10-19 2:07 ` Yicong Yang
@ 2022-10-19 8:13 ` Lorenzo Pieralisi
2022-10-19 18:30 ` Rafael J. Wysocki
3 siblings, 1 reply; 6+ messages in thread
From: Lorenzo Pieralisi @ 2022-10-19 8:13 UTC (permalink / raw)
To: Robin Murphy
Cc: rafael, lvjianmin, yangyicong, chenhuacai, linux-acpi,
linux-kernel, lenb, jeremy.linton
On Tue, Oct 18, 2022 at 02:14:04PM +0100, Robin Murphy wrote:
> Assigning the device's dma_range_map from the iterator variable after
> the loop means it always points to the empty terminator at the end of
> the map, which is not what we want. Similarly, freeing the iterator on
> error when it points to somwhere in the middle of the allocated array
> won't work either. Fix this.
>
> Fixes: bf2ee8d0c385 ("ACPI: scan: Support multiple DMA windows with different offsets")
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/acpi/scan.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
A quick comment below, otherwise:
Reviewed-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> index 558664d169fc..024cc373a197 100644
> --- a/drivers/acpi/scan.c
> +++ b/drivers/acpi/scan.c
> @@ -1509,9 +1509,12 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> goto out;
> }
>
> + *map = r;
I wonder whether having a local variable to stash the base pointer
would make code easier to read (so that we avoid using *map for that
purpose and also to return the array to the caller).
Thanks for fixing it so promptly.
Lorenzo
> +
> list_for_each_entry(rentry, &list, node) {
> if (rentry->res->start >= rentry->res->end) {
> - kfree(r);
> + kfree(*map);
> + *map = NULL;
> ret = -EINVAL;
> dev_dbg(dma_dev, "Invalid DMA regions configuration\n");
> goto out;
> @@ -1523,8 +1526,6 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> r->offset = rentry->offset;
> r++;
> }
> -
> - *map = r;
> }
> out:
> acpi_dev_free_resource_list(&list);
> --
> 2.36.1.dirty
>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH] ACPI: scan: Fix DMA range assignment
2022-10-19 8:13 ` Lorenzo Pieralisi
@ 2022-10-19 18:30 ` Rafael J. Wysocki
0 siblings, 0 replies; 6+ messages in thread
From: Rafael J. Wysocki @ 2022-10-19 18:30 UTC (permalink / raw)
To: Lorenzo Pieralisi, Robin Murphy
Cc: rafael, lvjianmin, yangyicong, chenhuacai, linux-acpi,
linux-kernel, lenb, jeremy.linton
On Wed, Oct 19, 2022 at 10:13 AM Lorenzo Pieralisi
<lpieralisi@kernel.org> wrote:
>
> On Tue, Oct 18, 2022 at 02:14:04PM +0100, Robin Murphy wrote:
> > Assigning the device's dma_range_map from the iterator variable after
> > the loop means it always points to the empty terminator at the end of
> > the map, which is not what we want. Similarly, freeing the iterator on
> > error when it points to somwhere in the middle of the allocated array
> > won't work either. Fix this.
> >
> > Fixes: bf2ee8d0c385 ("ACPI: scan: Support multiple DMA windows with different offsets")
> > Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> > ---
> > drivers/acpi/scan.c | 7 ++++---
> > 1 file changed, 4 insertions(+), 3 deletions(-)
>
> A quick comment below, otherwise:
>
> Reviewed-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Applied as is and the code may be cleaned up later.
Thanks!
> > diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> > index 558664d169fc..024cc373a197 100644
> > --- a/drivers/acpi/scan.c
> > +++ b/drivers/acpi/scan.c
> > @@ -1509,9 +1509,12 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> > goto out;
> > }
> >
> > + *map = r;
>
> I wonder whether having a local variable to stash the base pointer
> would make code easier to read (so that we avoid using *map for that
> purpose and also to return the array to the caller).
>
> Thanks for fixing it so promptly.
>
> Lorenzo
>
> > +
> > list_for_each_entry(rentry, &list, node) {
> > if (rentry->res->start >= rentry->res->end) {
> > - kfree(r);
> > + kfree(*map);
> > + *map = NULL;
> > ret = -EINVAL;
> > dev_dbg(dma_dev, "Invalid DMA regions configuration\n");
> > goto out;
> > @@ -1523,8 +1526,6 @@ int acpi_dma_get_range(struct device *dev, const struct bus_dma_region **map)
> > r->offset = rentry->offset;
> > r++;
> > }
> > -
> > - *map = r;
> > }
> > out:
> > acpi_dev_free_resource_list(&list);
> > --
> > 2.36.1.dirty
> >
^ permalink raw reply [flat|nested] 6+ messages in thread