2020-03-27 16:24:19

by Marek Szyprowski

[permalink] [raw]
Subject: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

Scatterlist elements contains both pages and DMA addresses, but one
should not assume 1:1 relation between them. The sg->length is the size
of the physical memory chunk described by the sg->page, while
sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
the sg_dma_address(sg).

The proper way of extracting both: pages and DMA addresses of the whole
buffer described by a scatterlist it to iterate independently over the
sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.

Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
Signed-off-by: Marek Szyprowski <[email protected]>
Reviewed-by: Alex Deucher <[email protected]>
---
drivers/gpu/drm/drm_prime.c | 37 +++++++++++++++++++++++++------------
1 file changed, 25 insertions(+), 12 deletions(-)

diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 1de2cde2277c..282774e469ac 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -962,27 +962,40 @@ int drm_prime_sg_to_page_addr_arrays(struct sg_table *sgt, struct page **pages,
unsigned count;
struct scatterlist *sg;
struct page *page;
- u32 len, index;
+ u32 page_len, page_index;
dma_addr_t addr;
+ u32 dma_len, dma_index;

- index = 0;
+ /*
+ * Scatterlist elements contains both pages and DMA addresses, but
+ * one shoud not assume 1:1 relation between them. The sg->length is
+ * the size of the physical memory chunk described by the sg->page,
+ * while sg_dma_len(sg) is the size of the DMA (IO virtual) chunk
+ * described by the sg_dma_address(sg).
+ */
+ page_index = 0;
+ dma_index = 0;
for_each_sg(sgt->sgl, sg, sgt->nents, count) {
- len = sg_dma_len(sg);
+ page_len = sg->length;
page = sg_page(sg);
+ dma_len = sg_dma_len(sg);
addr = sg_dma_address(sg);

- while (len > 0) {
- if (WARN_ON(index >= max_entries))
+ while (pages && page_len > 0) {
+ if (WARN_ON(page_index >= max_entries))
return -1;
- if (pages)
- pages[index] = page;
- if (addrs)
- addrs[index] = addr;
-
+ pages[page_index] = page;
page++;
+ page_len -= PAGE_SIZE;
+ page_index++;
+ }
+ while (addrs && dma_len > 0) {
+ if (WARN_ON(dma_index >= max_entries))
+ return -1;
+ addrs[dma_index] = addr;
addr += PAGE_SIZE;
- len -= PAGE_SIZE;
- index++;
+ dma_len -= PAGE_SIZE;
+ dma_index++;
}
}
return 0;
--
2.17.1


2020-03-27 18:32:24

by Michael J. Ruhl

[permalink] [raw]
Subject: RE: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

>-----Original Message-----
>From: Marek Szyprowski <[email protected]>
>Sent: Friday, March 27, 2020 12:21 PM
>To: [email protected]; [email protected];
>[email protected]
>Cc: Marek Szyprowski <[email protected]>;
>[email protected]; Bartlomiej Zolnierkiewicz
><[email protected]>; Maarten Lankhorst
><[email protected]>; Maxime Ripard
><[email protected]>; Thomas Zimmermann <[email protected]>;
>David Airlie <[email protected]>; Daniel Vetter <[email protected]>; Alex Deucher
><[email protected]>; Shane Francis <[email protected]>;
>Ruhl, Michael J <[email protected]>
>Subject: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a
>scatterlist
>
>Scatterlist elements contains both pages and DMA addresses, but one
>should not assume 1:1 relation between them. The sg->length is the size
>of the physical memory chunk described by the sg->page, while
>sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
>the sg_dma_address(sg).
>
>The proper way of extracting both: pages and DMA addresses of the whole
>buffer described by a scatterlist it to iterate independently over the
>sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.
>
>Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
>Signed-off-by: Marek Szyprowski <[email protected]>
>Reviewed-by: Alex Deucher <[email protected]>
>---
> drivers/gpu/drm/drm_prime.c | 37 +++++++++++++++++++++++++-----------
>-
> 1 file changed, 25 insertions(+), 12 deletions(-)
>
>diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
>index 1de2cde2277c..282774e469ac 100644
>--- a/drivers/gpu/drm/drm_prime.c
>+++ b/drivers/gpu/drm/drm_prime.c
>@@ -962,27 +962,40 @@ int drm_prime_sg_to_page_addr_arrays(struct
>sg_table *sgt, struct page **pages,
> unsigned count;
> struct scatterlist *sg;
> struct page *page;
>- u32 len, index;
>+ u32 page_len, page_index;
> dma_addr_t addr;
>+ u32 dma_len, dma_index;
>
>- index = 0;
>+ /*
>+ * Scatterlist elements contains both pages and DMA addresses, but
>+ * one shoud not assume 1:1 relation between them. The sg->length
>is
>+ * the size of the physical memory chunk described by the sg->page,
>+ * while sg_dma_len(sg) is the size of the DMA (IO virtual) chunk
>+ * described by the sg_dma_address(sg).
>+ */

Is there an example of what the scatterlist would look like in this case?

Does each SG entry always have the page and dma info? or could you have
entries that have page information only, and entries that have dma info only?

If the same entry has different size info (page_len = PAGE_SIZE,
dma_len = 4 * PAGE_SIZE?), are we guaranteed that the arrays (page and addrs) have
been sized correctly?

Just trying to get my head wrapped around this.

Thanks,

Mike

>+ page_index = 0;
>+ dma_index = 0;
> for_each_sg(sgt->sgl, sg, sgt->nents, count) {
>- len = sg_dma_len(sg);
>+ page_len = sg->length;
> page = sg_page(sg);
>+ dma_len = sg_dma_len(sg);
> addr = sg_dma_address(sg);
>
>- while (len > 0) {
>- if (WARN_ON(index >= max_entries))
>+ while (pages && page_len > 0) {
>+ if (WARN_ON(page_index >= max_entries))
> return -1;
>- if (pages)
>- pages[index] = page;
>- if (addrs)
>- addrs[index] = addr;
>-
>+ pages[page_index] = page;
> page++;
>+ page_len -= PAGE_SIZE;
>+ page_index++;
>+ }
>+ while (addrs && dma_len > 0) {
>+ if (WARN_ON(dma_index >= max_entries))
>+ return -1;
>+ addrs[dma_index] = addr;
> addr += PAGE_SIZE;
>- len -= PAGE_SIZE;
>- index++;
>+ dma_len -= PAGE_SIZE;
>+ dma_index++;
> }
> }
> return 0;
>--
>2.17.1

2020-03-28 07:18:59

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

On Fri, Mar 27, 2020 at 05:21:26PM +0100, Marek Szyprowski wrote:
> Scatterlist elements contains both pages and DMA addresses, but one
> should not assume 1:1 relation between them. The sg->length is the size
> of the physical memory chunk described by the sg->page, while
> sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
> the sg_dma_address(sg).
>
> The proper way of extracting both: pages and DMA addresses of the whole
> buffer described by a scatterlist it to iterate independently over the
> sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.
>
> Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
> Signed-off-by: Marek Szyprowski <[email protected]>
> Reviewed-by: Alex Deucher <[email protected]>
> ---
> drivers/gpu/drm/drm_prime.c | 37 +++++++++++++++++++++++++------------
> 1 file changed, 25 insertions(+), 12 deletions(-)

<formletter>

This is not the correct way to submit patches for inclusion in the
stable kernel tree. Please read:
https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
for how to do this properly.

</formletter>

2020-03-28 18:38:32

by Shane Francis

[permalink] [raw]
Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

On Fri, Mar 27, 2020 at 6:31 PM Ruhl, Michael J
<[email protected]> wrote:
> Is there an example of what the scatterlist would look like in this case?
>
> Does each SG entry always have the page and dma info? or could you have
> entries that have page information only, and entries that have dma info only?
>
> If the same entry has different size info (page_len = PAGE_SIZE,
> dma_len = 4 * PAGE_SIZE?), are we guaranteed that the arrays (page and addrs) have
> been sized correctly?
>
> Just trying to get my head wrapped around this.
>
> Thanks,
>
> Mike
>

My understanding is that page_len and dma_len in this case could have
different values (looking at iommu_dma_map_sg within dma-iommu.c),
this seems to add some padding calculated by using the device iova
domain to s_length but sg_dma_len is set to the original length

The scatterlists table can also get reduced down within
"__finalise_sg" possibly causing (if reduced) the dma_len of the last
table elements to be 0 (page_len would not be 0 in this case).

Documentation around looping & accessing scatterlists in DMA-API.txt
states that sg_dma_address() and sg_dma_len() should be used when
accessing addr and len rather than sg->address and sg->length.

Maybe it would be worth splitting this out into 2 functions to avoid
potential issues with the above use case ?

Regards,

Shane Francis

2020-03-29 10:06:45

by Marek Szyprowski

[permalink] [raw]
Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

Hi Michael,

On 2020-03-27 19:31, Ruhl, Michael J wrote:
>> -----Original Message-----
>> From: Marek Szyprowski <[email protected]>
>> Sent: Friday, March 27, 2020 12:21 PM
>> To: [email protected]; [email protected];
>> [email protected]
>> Cc: Marek Szyprowski <[email protected]>;
>> [email protected]; Bartlomiej Zolnierkiewicz
>> <[email protected]>; Maarten Lankhorst
>> <[email protected]>; Maxime Ripard
>> <[email protected]>; Thomas Zimmermann <[email protected]>;
>> David Airlie <[email protected]>; Daniel Vetter <[email protected]>; Alex Deucher
>> <[email protected]>; Shane Francis <[email protected]>;
>> Ruhl, Michael J <[email protected]>
>> Subject: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a
>> scatterlist
>>
>> Scatterlist elements contains both pages and DMA addresses, but one
>> should not assume 1:1 relation between them. The sg->length is the size
>> of the physical memory chunk described by the sg->page, while
>> sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
>> the sg_dma_address(sg).
>>
>> The proper way of extracting both: pages and DMA addresses of the whole
>> buffer described by a scatterlist it to iterate independently over the
>> sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.
>>
>> Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
>> Signed-off-by: Marek Szyprowski <[email protected]>
>> Reviewed-by: Alex Deucher <[email protected]>
>> ---
>> drivers/gpu/drm/drm_prime.c | 37 +++++++++++++++++++++++++-----------
>> -
>> 1 file changed, 25 insertions(+), 12 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
>> index 1de2cde2277c..282774e469ac 100644
>> --- a/drivers/gpu/drm/drm_prime.c
>> +++ b/drivers/gpu/drm/drm_prime.c
>> @@ -962,27 +962,40 @@ int drm_prime_sg_to_page_addr_arrays(struct
>> sg_table *sgt, struct page **pages,
>> unsigned count;
>> struct scatterlist *sg;
>> struct page *page;
>> - u32 len, index;
>> + u32 page_len, page_index;
>> dma_addr_t addr;
>> + u32 dma_len, dma_index;
>>
>> - index = 0;
>> + /*
>> + * Scatterlist elements contains both pages and DMA addresses, but
>> + * one shoud not assume 1:1 relation between them. The sg->length
>> is
>> + * the size of the physical memory chunk described by the sg->page,
>> + * while sg_dma_len(sg) is the size of the DMA (IO virtual) chunk
>> + * described by the sg_dma_address(sg).
>> + */
> Is there an example of what the scatterlist would look like in this case?

DMA framework or IOMMU is allowed to join consecutive chunks while
mapping if such operation is supported by the hw. Here is the example:

Lets assume that we have a scatterlist with 4 4KiB pages of the physical
addresses: 0x12000000, 0x13011000, 0x13012000, 0x11011000. The total
size of the buffer is 16KiB. After mapping this scatterlist to a device
behind an IOMMU it may end up as a contiguous buffer in the DMA (IOVA)
address space. at 0xf0010000. The scatterlist will look like this:

sg[0].page = 0x12000000
sg[0].len = 4096
sg[0].dma_addr = 0xf0010000
sg[0].dma_len = 16384
sg[1].page = 0x13011000
sg[1].len = 4096
sg[1].dma_addr = 0
sg[1].dma_len = 0
sg[2].page = 0x13012000
sg[2].len = 4096
sg[2].dma_addr = 0
sg[2].dma_len = 0
sg[3].page = 0x11011000
sg[3].len = 4096
sg[3].dma_addr = 0
sg[3].dma_len = 0

(I've intentionally wrote page as physical address to make it easier to
understand, in real SGs it is stored a struct page pointer).

> Does each SG entry always have the page and dma info? or could you have
> entries that have page information only, and entries that have dma info only?
When SG is not mapped yet it contains only the ->pages and ->len
entries. I'm not aware of the SGs with the DMA information only, but in
theory it might be possible to have such.
> If the same entry has different size info (page_len = PAGE_SIZE,
> dma_len = 4 * PAGE_SIZE?), are we guaranteed that the arrays (page and addrs) have
> been sized correctly?

There are always no more DMA related entries than the phys pages. If
there is 1:1 mapping between physical memory and DMA (IOVA) space, then
each SG entry will have len == dma_len, and dma_addr will be describing
the same as page entry. DMA mapping framework is allowed only to join
entries while mapping to DMA (IOVA).

> Just trying to get my head wrapped around this.

Sure, I hope my explanation helps a bit.

Best regards
--
Marek Szyprowski, PhD
Samsung R&D Institute Poland

2020-03-30 13:11:00

by Robin Murphy

[permalink] [raw]
Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

On 2020-03-29 10:55 am, Marek Szyprowski wrote:
> Hi Michael,
>
> On 2020-03-27 19:31, Ruhl, Michael J wrote:
>>> -----Original Message-----
>>> From: Marek Szyprowski <[email protected]>
>>> Sent: Friday, March 27, 2020 12:21 PM
>>> To: [email protected]; [email protected];
>>> [email protected]
>>> Cc: Marek Szyprowski <[email protected]>;
>>> [email protected]; Bartlomiej Zolnierkiewicz
>>> <[email protected]>; Maarten Lankhorst
>>> <[email protected]>; Maxime Ripard
>>> <[email protected]>; Thomas Zimmermann <[email protected]>;
>>> David Airlie <[email protected]>; Daniel Vetter <[email protected]>; Alex Deucher
>>> <[email protected]>; Shane Francis <[email protected]>;
>>> Ruhl, Michael J <[email protected]>
>>> Subject: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a
>>> scatterlist
>>>
>>> Scatterlist elements contains both pages and DMA addresses, but one
>>> should not assume 1:1 relation between them. The sg->length is the size
>>> of the physical memory chunk described by the sg->page, while
>>> sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
>>> the sg_dma_address(sg).
>>>
>>> The proper way of extracting both: pages and DMA addresses of the whole
>>> buffer described by a scatterlist it to iterate independently over the
>>> sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.
>>>
>>> Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
>>> Signed-off-by: Marek Szyprowski <[email protected]>
>>> Reviewed-by: Alex Deucher <[email protected]>
>>> ---
>>> drivers/gpu/drm/drm_prime.c | 37 +++++++++++++++++++++++++-----------
>>> -
>>> 1 file changed, 25 insertions(+), 12 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
>>> index 1de2cde2277c..282774e469ac 100644
>>> --- a/drivers/gpu/drm/drm_prime.c
>>> +++ b/drivers/gpu/drm/drm_prime.c
>>> @@ -962,27 +962,40 @@ int drm_prime_sg_to_page_addr_arrays(struct
>>> sg_table *sgt, struct page **pages,
>>> unsigned count;
>>> struct scatterlist *sg;
>>> struct page *page;
>>> - u32 len, index;
>>> + u32 page_len, page_index;
>>> dma_addr_t addr;
>>> + u32 dma_len, dma_index;
>>>
>>> - index = 0;
>>> + /*
>>> + * Scatterlist elements contains both pages and DMA addresses, but
>>> + * one shoud not assume 1:1 relation between them. The sg->length
>>> is
>>> + * the size of the physical memory chunk described by the sg->page,
>>> + * while sg_dma_len(sg) is the size of the DMA (IO virtual) chunk
>>> + * described by the sg_dma_address(sg).
>>> + */
>> Is there an example of what the scatterlist would look like in this case?
>
> DMA framework or IOMMU is allowed to join consecutive chunks while
> mapping if such operation is supported by the hw. Here is the example:
>
> Lets assume that we have a scatterlist with 4 4KiB pages of the physical
> addresses: 0x12000000, 0x13011000, 0x13012000, 0x11011000. The total
> size of the buffer is 16KiB. After mapping this scatterlist to a device
> behind an IOMMU it may end up as a contiguous buffer in the DMA (IOVA)
> address space. at 0xf0010000. The scatterlist will look like this:
>
> sg[0].page = 0x12000000
> sg[0].len = 4096
> sg[0].dma_addr = 0xf0010000
> sg[0].dma_len = 16384
> sg[1].page = 0x13011000
> sg[1].len = 4096
> sg[1].dma_addr = 0
> sg[1].dma_len = 0
> sg[2].page = 0x13012000
> sg[2].len = 4096
> sg[2].dma_addr = 0
> sg[2].dma_len = 0
> sg[3].page = 0x11011000
> sg[3].len = 4096
> sg[3].dma_addr = 0
> sg[3].dma_len = 0
>
> (I've intentionally wrote page as physical address to make it easier to
> understand, in real SGs it is stored a struct page pointer).
>
>> Does each SG entry always have the page and dma info? or could you have
>> entries that have page information only, and entries that have dma info only?
> When SG is not mapped yet it contains only the ->pages and ->len
> entries. I'm not aware of the SGs with the DMA information only, but in
> theory it might be possible to have such.
>> If the same entry has different size info (page_len = PAGE_SIZE,
>> dma_len = 4 * PAGE_SIZE?), are we guaranteed that the arrays (page and addrs) have
>> been sized correctly?
>
> There are always no more DMA related entries than the phys pages. If
> there is 1:1 mapping between physical memory and DMA (IOVA) space, then
> each SG entry will have len == dma_len, and dma_addr will be describing
> the same as page entry. DMA mapping framework is allowed only to join
> entries while mapping to DMA (IOVA).

Nit: even in a 1:1 mapping, merging would still be permitted (subject to
dma_parms constraints) during a bounce-buffer copy, or if the caller
simply generates a naive list like so:

sg[0].page = 0x12000000
sg[0].len = 4096
sg[1].page = 0x12001000
sg[1].len = 4096

dma_map_sg() =>

sg[0].dma_addr = 0x12000000
sg[0].dma_len = 8192
sg[1].dma_addr = 0
sg[1].dma_len = 0

I'm not sure that any non-IOMMU DMA API implementations actually take
advantage of this, but they are *allowed* to ;)

Robin.

2020-03-30 14:05:24

by Michael J. Ruhl

[permalink] [raw]
Subject: RE: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

>-----Original Message-----
>From: dri-devel <[email protected]> On Behalf Of
>Marek Szyprowski
>Sent: Sunday, March 29, 2020 5:56 AM
>To: Ruhl, Michael J <[email protected]>; dri-
>[email protected]; [email protected]; linux-
>[email protected]
>Cc: Bartlomiej Zolnierkiewicz <[email protected]>; David Airlie
><[email protected]>; Shane Francis <[email protected]>;
>[email protected]; Thomas Zimmermann <[email protected]>;
>Alex Deucher <[email protected]>
>Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from
>a scatterlist
>
>Hi Michael,
>

<snip>

>> Is there an example of what the scatterlist would look like in this case?
>
>DMA framework or IOMMU is allowed to join consecutive chunks while
>mapping if such operation is supported by the hw. Here is the example:
>
>Lets assume that we have a scatterlist with 4 4KiB pages of the physical
>addresses: 0x12000000, 0x13011000, 0x13012000, 0x11011000. The total
>size of the buffer is 16KiB. After mapping this scatterlist to a device
>behind an IOMMU it may end up as a contiguous buffer in the DMA (IOVA)
>address space. at 0xf0010000. The scatterlist will look like this:
>
>sg[0].page = 0x12000000
>sg[0].len = 4096
>sg[0].dma_addr = 0xf0010000
>sg[0].dma_len = 16384
>sg[1].page = 0x13011000
>sg[1].len = 4096
>sg[1].dma_addr = 0
>sg[1].dma_len = 0
>sg[2].page = 0x13012000
>sg[2].len = 4096
>sg[2].dma_addr = 0
>sg[2].dma_len = 0
>sg[3].page = 0x11011000
>sg[3].len = 4096
>sg[3].dma_addr = 0
>sg[3].dma_len = 0
>
>(I've intentionally wrote page as physical address to make it easier to
>understand, in real SGs it is stored a struct page pointer).
>
>> Does each SG entry always have the page and dma info? or could you have
>> entries that have page information only, and entries that have dma info
>only?
>When SG is not mapped yet it contains only the ->pages and ->len
>entries. I'm not aware of the SGs with the DMA information only, but in
>theory it might be possible to have such.
>> If the same entry has different size info (page_len = PAGE_SIZE,
>> dma_len = 4 * PAGE_SIZE?), are we guaranteed that the arrays (page and
>addrs) have
>> been sized correctly?
>
>There are always no more DMA related entries than the phys pages. If
>there is 1:1 mapping between physical memory and DMA (IOVA) space, then
>each SG entry will have len == dma_len, and dma_addr will be describing
>the same as page entry. DMA mapping framework is allowed only to join
>entries while mapping to DMA (IOVA).
>
>> Just trying to get my head wrapped around this.
>
>Sure, I hope my explanation helps a bit.

That is a great example! Thank you very much for the explanation.

I was somehow seeing it as the dma side getting split and extended (rather
than consolidated) into more possible entries. This clarifies the issue for me.

Thanks!

Mike

>Best regards
>--
>Marek Szyprowski, PhD
>Samsung R&D Institute Poland
>
>_______________________________________________
>dri-devel mailing list
>[email protected]
>https://lists.freedesktop.org/mailman/listinfo/dri-devel

2020-04-05 14:53:48

by Alex Deucher

[permalink] [raw]
Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

On Fri, Mar 27, 2020 at 12:23 PM Marek Szyprowski
<[email protected]> wrote:
>
> Scatterlist elements contains both pages and DMA addresses, but one
> should not assume 1:1 relation between them. The sg->length is the size
> of the physical memory chunk described by the sg->page, while
> sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
> the sg_dma_address(sg).
>
> The proper way of extracting both: pages and DMA addresses of the whole
> buffer described by a scatterlist it to iterate independently over the
> sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.
>
> Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
> Signed-off-by: Marek Szyprowski <[email protected]>
> Reviewed-by: Alex Deucher <[email protected]>

Applied. Thanks and sorry for the breakage.

Alex

> ---
> drivers/gpu/drm/drm_prime.c | 37 +++++++++++++++++++++++++------------
> 1 file changed, 25 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
> index 1de2cde2277c..282774e469ac 100644
> --- a/drivers/gpu/drm/drm_prime.c
> +++ b/drivers/gpu/drm/drm_prime.c
> @@ -962,27 +962,40 @@ int drm_prime_sg_to_page_addr_arrays(struct sg_table *sgt, struct page **pages,
> unsigned count;
> struct scatterlist *sg;
> struct page *page;
> - u32 len, index;
> + u32 page_len, page_index;
> dma_addr_t addr;
> + u32 dma_len, dma_index;
>
> - index = 0;
> + /*
> + * Scatterlist elements contains both pages and DMA addresses, but
> + * one shoud not assume 1:1 relation between them. The sg->length is
> + * the size of the physical memory chunk described by the sg->page,
> + * while sg_dma_len(sg) is the size of the DMA (IO virtual) chunk
> + * described by the sg_dma_address(sg).
> + */
> + page_index = 0;
> + dma_index = 0;
> for_each_sg(sgt->sgl, sg, sgt->nents, count) {
> - len = sg_dma_len(sg);
> + page_len = sg->length;
> page = sg_page(sg);
> + dma_len = sg_dma_len(sg);
> addr = sg_dma_address(sg);
>
> - while (len > 0) {
> - if (WARN_ON(index >= max_entries))
> + while (pages && page_len > 0) {
> + if (WARN_ON(page_index >= max_entries))
> return -1;
> - if (pages)
> - pages[index] = page;
> - if (addrs)
> - addrs[index] = addr;
> -
> + pages[page_index] = page;
> page++;
> + page_len -= PAGE_SIZE;
> + page_index++;
> + }
> + while (addrs && dma_len > 0) {
> + if (WARN_ON(dma_index >= max_entries))
> + return -1;
> + addrs[dma_index] = addr;
> addr += PAGE_SIZE;
> - len -= PAGE_SIZE;
> - index++;
> + dma_len -= PAGE_SIZE;
> + dma_index++;
> }
> }
> return 0;
> --
> 2.17.1
>
> _______________________________________________
> dri-devel mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

2020-04-05 16:42:08

by Greg KH

[permalink] [raw]
Subject: Re: [PATCH v2] drm/prime: fix extracting of the DMA addresses from a scatterlist

On Sun, Apr 05, 2020 at 10:47:49AM -0400, Alex Deucher wrote:
> On Fri, Mar 27, 2020 at 12:23 PM Marek Szyprowski
> <[email protected]> wrote:
> >
> > Scatterlist elements contains both pages and DMA addresses, but one
> > should not assume 1:1 relation between them. The sg->length is the size
> > of the physical memory chunk described by the sg->page, while
> > sg_dma_len(sg) is the size of the DMA (IO virtual) chunk described by
> > the sg_dma_address(sg).
> >
> > The proper way of extracting both: pages and DMA addresses of the whole
> > buffer described by a scatterlist it to iterate independently over the
> > sg->pages/sg->length and sg_dma_address(sg)/sg_dma_len(sg) entries.
> >
> > Fixes: 42e67b479eab ("drm/prime: use dma length macro when mapping sg")
> > Signed-off-by: Marek Szyprowski <[email protected]>
> > Reviewed-by: Alex Deucher <[email protected]>
>
> Applied. Thanks and sorry for the breakage.


<formletter>

This is not the correct way to submit patches for inclusion in the
stable kernel tree. Please read:
https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
for how to do this properly.

</formletter>