Re: + memblock-add-assertion-for-zero-allocation-alignment.patch added to -mm tree

Re: + memblock-add-assertion-for-zero-allocation-alignment.patch added to -mm tree

[Yinghai Lu]

On Wed, Mar 6, 2013 at 2:44 PM,  <akpm@linux-foundation.org> wrote:
> ------------------------------------------------------
> From: Vineet Gupta <Vineet.Gupta1@synopsys.com>
> Subject: memblock: add assertion for zero allocation alignment
>
> This came to light when calling memblock allocator from arc port (for
> copying flattended DT).  If a "0" alignment is passed, the allocator
> round_up() call incorrectly rounds up the size to 0.
>
> round_up(num, alignto) => ((num - 1) | (alignto -1)) + 1
>
> While the obvious allocation failure causes kernel to panic, it is better
> to warn the caller to fix the code.
>
> Tejun suggested that instead of BUG_ON(!align) - which might be
> ineffective due to pending console init and such, it is better to WARN_ON,
> and continue the boot with a reasonable default align.
>
> Caller passing @size need not be handled similarly as the subsequent
> panic will indicate that anyhow.
>
> Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
> Cc: Yinghai Lu <yinghai@kernel.org>
> Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
> Cc: Ingo Molnar <mingo@kernel.org>
> Acked-by: Tejun Heo <tj@kernel.org>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
>
>  mm/memblock.c |    3 +++
>  1 file changed, 3 insertions(+)
>
> diff -puN mm/memblock.c~memblock-add-assertion-for-zero-allocation-alignment mm/memblock.c
> --- a/mm/memblock.c~memblock-add-assertion-for-zero-allocation-alignment
> +++ a/mm/memblock.c
> @@ -771,6 +771,9 @@ static phys_addr_t __init memblock_alloc
>  {
>         phys_addr_t found;
>
> +       if (WARN_ON(!align))
> +               align = __alignof__(long long);
> +
>         /* align @size to avoid excessive fragmentation on reserved array */
>         size = round_up(size, align);

Hi, Peter,

Do you agree that we should check align in round_up()?

Thanks

Yinghai

Re: + memblock-add-assertion-for-zero-allocation-alignment.patch added to -mm tree

[Andrew Morton]

On Wed, 6 Mar 2013 16:07:20 -0800 Yinghai Lu <yinghai@kernel.org> wrote:

> > --- a/mm/memblock.c~memblock-add-assertion-for-zero-allocation-alignment
> > +++ a/mm/memblock.c
> > @@ -771,6 +771,9 @@ static phys_addr_t __init memblock_alloc
> >  {
> >         phys_addr_t found;
> >
> > +       if (WARN_ON(!align))
> > +               align = __alignof__(long long);
> > +
> >         /* align @size to avoid excessive fragmentation on reserved array */
> >         size = round_up(size, align);
> 
> Hi, Peter,
> 
> Do you agree that we should check align in round_up()?

As you don't describe your reasoning it is hard to say.

But no, I don't think so.  Checking for zero would add a pile of
basically useless code to the 100+ round_up() callsites, and
round_up(x, 0) is kinda meaningful, in a strange way.

Re: + memblock-add-assertion-for-zero-allocation-alignment.patch added to -mm tree

[H. Peter Anvin]

On 03/06/2013 04:07 PM, Yinghai Lu wrote:
>>
>>  mm/memblock.c |    3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff -puN mm/memblock.c~memblock-add-assertion-for-zero-allocation-alignment mm/memblock.c
>> --- a/mm/memblock.c~memblock-add-assertion-for-zero-allocation-alignment
>> +++ a/mm/memblock.c
>> @@ -771,6 +771,9 @@ static phys_addr_t __init memblock_alloc
>>  {
>>         phys_addr_t found;
>>
>> +       if (WARN_ON(!align))
>> +               align = __alignof__(long long);
>> +
>>         /* align @size to avoid excessive fragmentation on reserved array */
>>         size = round_up(size, align);
> 
> Hi, Peter,
> 
> Do you agree that we should check align in round_up()?
> 

Not in round_up(), that is used in way too many places.  Doing it in
memblock_alloc() might make sense.

	-hpa