* [PATCH] of: alloc anywhere from memblock if range not specified
@ 2016-02-22 13:45 Vinayak Menon
2016-03-03 22:52 ` Rob Herring
2017-01-10 16:16 ` Leonard Crestez
0 siblings, 2 replies; 4+ messages in thread
From: Vinayak Menon @ 2016-02-22 13:45 UTC (permalink / raw)
To: devicetree, linux-kernel, linux-arm-kernel
Cc: robh+dt, frowand.list, grant.likely, m.szyprowski, vinayakm.list,
Vinayak Menon
early_init_dt_alloc_reserved_memory_arch passes end as 0 to
__memblock_alloc_base, when limits are not specified. But
__memblock_alloc_base takes end value of 0 as MEMBLOCK_ALLOC_ACCESSIBLE
and limits the end to memblock.current_limit. This results in regions
never being placed in HIGHMEM area, for e.g. CMA.
Let __memblock_alloc_base allocate from anywhere in memory if limits are
not specified.
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
---
drivers/of/of_reserved_mem.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
index 1a3556a..ed01c01 100644
--- a/drivers/of/of_reserved_mem.c
+++ b/drivers/of/of_reserved_mem.c
@@ -32,11 +32,13 @@ int __init __weak early_init_dt_alloc_reserved_memory_arch(phys_addr_t size,
phys_addr_t align, phys_addr_t start, phys_addr_t end, bool nomap,
phys_addr_t *res_base)
{
+ phys_addr_t base;
/*
* We use __memblock_alloc_base() because memblock_alloc_base()
* panic()s on allocation failure.
*/
- phys_addr_t base = __memblock_alloc_base(size, align, end);
+ end = !end ? MEMBLOCK_ALLOC_ANYWHERE : end;
+ base = __memblock_alloc_base(size, align, end);
if (!base)
return -ENOMEM;
--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
member of the Code Aurora Forum, hosted by The Linux Foundation
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] of: alloc anywhere from memblock if range not specified
2016-02-22 13:45 [PATCH] of: alloc anywhere from memblock if range not specified Vinayak Menon
@ 2016-03-03 22:52 ` Rob Herring
2017-01-10 16:16 ` Leonard Crestez
1 sibling, 0 replies; 4+ messages in thread
From: Rob Herring @ 2016-03-03 22:52 UTC (permalink / raw)
To: Vinayak Menon
Cc: devicetree, linux-kernel, linux-arm-kernel, frowand.list,
grant.likely, m.szyprowski, vinayakm.list
On Mon, Feb 22, 2016 at 07:15:44PM +0530, Vinayak Menon wrote:
> early_init_dt_alloc_reserved_memory_arch passes end as 0 to
> __memblock_alloc_base, when limits are not specified. But
> __memblock_alloc_base takes end value of 0 as MEMBLOCK_ALLOC_ACCESSIBLE
> and limits the end to memblock.current_limit. This results in regions
> never being placed in HIGHMEM area, for e.g. CMA.
> Let __memblock_alloc_base allocate from anywhere in memory if limits are
> not specified.
>
> Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
> Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
Applied and marked for stable.
Rob
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] of: alloc anywhere from memblock if range not specified
2016-02-22 13:45 [PATCH] of: alloc anywhere from memblock if range not specified Vinayak Menon
2016-03-03 22:52 ` Rob Herring
@ 2017-01-10 16:16 ` Leonard Crestez
2017-01-10 19:49 ` Laura Abbott
1 sibling, 1 reply; 4+ messages in thread
From: Leonard Crestez @ 2017-01-10 16:16 UTC (permalink / raw)
To: Vinayak Menon
Cc: devicetree, linux-kernel, linux-arm-kernel, robh+dt,
frowand.list, grant.likely, m.szyprowski, vinayakm.list,
Octavian Purdila
Hello,
I have some trouble with this patch.
It seems the intention is to allow CMA to be placed in highmem. If the
CMA area is larger than highmem and no alloc-ranges is specified (just a
size) it is possible to end up allocating a area that spans from
multiple zones. This later breaks checks in cma_activate_area and makes
most dma allocations fail.
Am I missing something or this a bug?
On Mon, Feb 22, 2016 at 3:45 PM, Vinayak Menon <vinmenon@codeaurora.org>
wrote:
>
> early_init_dt_alloc_reserved_memory_arch passes end as 0 to
> __memblock_alloc_base, when limits are not specified. But
> __memblock_alloc_base takes end value of 0 as MEMBLOCK_ALLOC_ACCESSIBLE
> and limits the end to memblock.current_limit. This results in regions
> never being placed in HIGHMEM area, for e.g. CMA.
> Let __memblock_alloc_base allocate from anywhere in memory if limits are
> not specified.
>
> Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
> Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
> ---
> drivers/of/of_reserved_mem.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> index 1a3556a..ed01c01 100644
> --- a/drivers/of/of_reserved_mem.c
> +++ b/drivers/of/of_reserved_mem.c
> @@ -32,11 +32,13 @@ int __init __weak early_init_dt_alloc_reserved_memory_arch(phys_addr_t size,
> phys_addr_t align, phys_addr_t start, phys_addr_t end, bool nomap,
> phys_addr_t *res_base)
> {
> + phys_addr_t base;
> /*
> * We use __memblock_alloc_base() because memblock_alloc_base()
> * panic()s on allocation failure.
> */
> - phys_addr_t base = __memblock_alloc_base(size, align, end);
> + end = !end ? MEMBLOCK_ALLOC_ANYWHERE : end;
> + base = __memblock_alloc_base(size, align, end);
> if (!base)
> return -ENOMEM;
>
> --
> QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
> member of the Code Aurora Forum, hosted by The Linux Foundation
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] of: alloc anywhere from memblock if range not specified
2017-01-10 16:16 ` Leonard Crestez
@ 2017-01-10 19:49 ` Laura Abbott
0 siblings, 0 replies; 4+ messages in thread
From: Laura Abbott @ 2017-01-10 19:49 UTC (permalink / raw)
To: Leonard Crestez, Vinayak Menon
Cc: devicetree, linux-kernel, linux-arm-kernel, robh+dt,
frowand.list, m.szyprowski, vinayakm.list, Octavian Purdila
On 01/10/2017 08:16 AM, Leonard Crestez wrote:
> Hello,
>
> I have some trouble with this patch.
>
> It seems the intention is to allow CMA to be placed in highmem. If the CMA area is
> larger than highmem and no alloc-ranges is specified (just a size) it is possible
> to end up allocating a area that spans from multiple zones. This later breaks
> checks in cma_activate_area and makes most dma allocations fail.
>
> Am I missing something or this a bug?
>
This has been discussed in previous threads
https://marc.info/?l=linux-kernel&m=147990760506179&w=2
https://marc.info/?l=linux-kernel&m=147928325113103&w=2
I haven't seen any follow up since then though.
> On Mon, Feb 22, 2016 at 3:45 PM, Vinayak Menon <vinmenon@codeaurora.org> wrote:
>>
>> early_init_dt_alloc_reserved_memory_arch passes end as 0 to
>> __memblock_alloc_base, when limits are not specified. But
>> __memblock_alloc_base takes end value of 0 as MEMBLOCK_ALLOC_ACCESSIBLE
>> and limits the end to memblock.current_limit. This results in regions
>> never being placed in HIGHMEM area, for e.g. CMA.
>> Let __memblock_alloc_base allocate from anywhere in memory if limits are
>> not specified.
>>
>> Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
>> Signed-off-by: Vinayak Menon <vinmenon@codeaurora.org>
>> ---
>> drivers/of/of_reserved_mem.c | 4 +++-
>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
>> index 1a3556a..ed01c01 100644
>> --- a/drivers/of/of_reserved_mem.c
>> +++ b/drivers/of/of_reserved_mem.c
>> @@ -32,11 +32,13 @@ int __init __weak early_init_dt_alloc_reserved_memory_arch(phys_addr_t size,
>> phys_addr_t align, phys_addr_t start, phys_addr_t end, bool nomap,
>> phys_addr_t *res_base)
>> {
>> + phys_addr_t base;
>> /*
>> * We use __memblock_alloc_base() because memblock_alloc_base()
>> * panic()s on allocation failure.
>> */
>> - phys_addr_t base = __memblock_alloc_base(size, align, end);
>> + end = !end ? MEMBLOCK_ALLOC_ANYWHERE : end;
>> + base = __memblock_alloc_base(size, align, end);
>> if (!base)
>> return -ENOMEM;
>>
>> --
>> QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
>> member of the Code Aurora Forum, hosted by The Linux Foundation
>>
> --
> To unsubscribe from this list: send the line "unsubscribe devicetree" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2017-01-10 19:49 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-02-22 13:45 [PATCH] of: alloc anywhere from memblock if range not specified Vinayak Menon
2016-03-03 22:52 ` Rob Herring
2017-01-10 16:16 ` Leonard Crestez
2017-01-10 19:49 ` Laura Abbott
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).