[PATCH v6 1/3] genalloc:support memory-allocation with bytes-alignment to genalloc

Zhao Qiang qiang.zhao at freescale.com
Tue Aug 25 18:09:56 AEST 2015


On 08/25/2015 12:01 PM, Laura Abbott wrote:

> -----Original Message-----
> From: Laura Abbott [mailto:labbott at redhat.com]
> Sent: Tuesday, August 25, 2015 12:01 PM
> To: Zhao Qiang-B45475; Wood Scott-B07421
> Cc: linux-kernel at vger.kernel.org; linuxppc-dev at lists.ozlabs.org;
> lauraa at codeaurora.org; Xie Xiaobo-R63061; benh at kernel.crashing.org; Li
> Yang-Leo-R58472; paulus at samba.org
> Subject: Re: [PATCH v6 1/3] genalloc:support memory-allocation with
> bytes-alignment to genalloc
> 
> On 08/24/2015 07:40 PM, Zhao Qiang wrote:
> > On 08/25/2015 07:11 AM, Laura Abbott wrote:
> >> -----Original Message-----
> >> From: Laura Abbott [mailto:labbott at redhat.com]
> >> Sent: Tuesday, August 25, 2015 7:11 AM
> >> To: Zhao Qiang-B45475; Wood Scott-B07421
> >> Cc: linux-kernel at vger.kernel.org; linuxppc-dev at lists.ozlabs.org;
> >> lauraa at codeaurora.org; Xie Xiaobo-R63061; benh at kernel.crashing.org;
> >> Li Yang-Leo-R58472; paulus at samba.org
> >> Subject: Re: [PATCH v6 1/3] genalloc:support memory-allocation with
> >> bytes-alignment to genalloc
> >>
> >> On 08/24/2015 02:31 AM, Zhao Qiang wrote:
> >>> Bytes alignment is required to manage some special RAM, so add
> >>> gen_pool_first_fit_align to genalloc, meanwhile add
> >>> gen_pool_alloc_data to pass data to gen_pool_first_fit_align(modify
> >>> gen_pool_alloc as a wrapper)
> >>>
> >>> Signed-off-by: Zhao Qiang <qiang.zhao at freescale.com>
> >>> ---
> >>> Changes for v6:
> >>> 	- patches set v6 include a new patch because of using
> >>> 	- genalloc to manage QE MURAM, patch 0001 is the new
> >>> 	- patch, adding bytes alignment for allocation for use.
> >>>
> >>>    include/linux/genalloc.h | 23 +++++++++++++++----
> >>>    lib/genalloc.c           | 58
> >> +++++++++++++++++++++++++++++++++++++++++++-----
> >>>    2 files changed, 72 insertions(+), 9 deletions(-)
> >>>
> >>> diff --git a/include/linux/genalloc.h b/include/linux/genalloc.h
> >>> index 1ccaab4..55da07e 100644
> >>> --- a/include/linux/genalloc.h
> >>> +++ b/include/linux/genalloc.h
> >>> @@ -34,6 +34,7 @@
> >>>
> >>>    struct device;
> >>>    struct device_node;
> >>> +struct gen_pool;
> >>>
> >>>    /**
> >>>     * Allocation callback function type definition @@ -47,7 +48,7 @@
> >>> typedef unsigned long (*genpool_algo_t)(unsigned long *map,
> >>>    			unsigned long size,
> >>>    			unsigned long start,
> >>>    			unsigned int nr,
> >>> -			void *data);
> >>> +			void *data, struct gen_pool *pool);
> >>>
> >>>    /*
> >>>     *  General purpose special memory pool descriptor.
> >>> @@ -73,6 +74,13 @@ struct gen_pool_chunk {
> >>>    	unsigned long bits[0];		/* bitmap for allocating memory
> chunk
> >> */
> >>>    };
> >>>
> >>> +/*
> >>> + *  gen_pool data descriptor for gen_pool_first_fit_align.
> >>> + */
> >>> +struct genpool_data_align {
> >>> +	int align;		/* alignment by bytes for starting address */
> >>> +};
> >>> +
> >>
> >> (sorry for chiming in late, I've been traveling)
> >>
> >> Is there an advantage here to wrapping this in a structure instead of
> >> just passing a pointer to an align integer?
> >
> >
> > Please look at the commit message for
> > ca279cf1065fb689abea1dc7d8c11787729bb185 which adds "data":
> >
> > "As I can't predict all the possible requirements/needs for all
> > allocation uses cases, I add a "free" field 'void *data' to pass any
> > needed information to the allocation function.  For example 'data'
> > could be used to handle a structure where you store the alignment, the
> > expected memory bank, the requester device, or any information that
> > could influence the allocation algorithm."
> >
> 
> Right, I understand what the purpose is but I'm not sure what you're
> getting from the structure vs passing a pointer, e.g.
> 
> int align;
> 
> align = 4;
> 
> gen_pool_alloc_data(&pool, size, &align);
> 
> it just seems to obfuscate what's going on by wrapping a single integer
> in a structure that's narrowly defined in a generic function right now. I
> guess it could change later which would necessitate having the structure
> but again it's so generic I'm not sure what else you would pass that
> would be applicable to all clients.

Scott and me have discussed about this issue in my RFC patch.
Please review: http://patchwork.ozlabs.org/patch/493297/

> 
> Thanks,
> Laura


More information about the Linuxppc-dev mailing list