public inbox for linux-omap@vger.kernel.org
 help / color / mirror / Atom feed
From: Nishanth Menon <nm@ti.com>
To: "Kanigeri, Hari" <h-kanigeri2@ti.com>
Cc: "Ramirez Luna, Omar" <omar.ramirez@ti.com>,
	linux-omap <linux-omap@vger.kernel.org>,
	Hiroshi Doyu <Hiroshi.DOYU@nokia.com>,
	Ameya Palande <ameya.palande@nokia.com>,
	Felipe Contreras <felipe.contreras@nokia.com>,
	"Guzman Lugo, Fernando" <x0095840@ti.com>,
	"Ramos Falcon, Ernesto" <ernesto@ti.com>,
	"Aguilar Pena, Leed" <leed.aguilar@ti.com>
Subject: Re: [PATCH 3/8] DSPBRIDGE: Increased DMM size to 256MB
Date: Fri, 8 Jan 2010 09:33:26 -0600	[thread overview]
Message-ID: <4B475046.8020500@ti.com> (raw)
In-Reply-To: <8F7AF80515AF0D4D93307E594F3CB40E43C0459F@dlee03.ent.ti.com>

Kanigeri, Hari had written, on 01/08/2010 09:29 AM, the following:
> Nishant,
> 
>> Would it be better that we make this as a board specific memory
>> requirement? not all boards will have the same needs right?
> 
> For every 1MB of DSP Virtual address 1Kbytes of physical memory is required for house keeping. So, for 256MB we would be taking 256K as opposed to 64K when using 64MB DSP virtual memory.
> 
> The requirement comes more from Multimedia requirements, and with all the new phones having the requirement of running multiple multimedia applications in parallel 256K would be safe.

the point is you are moving to 256MB today (Still not reaching 100% of 
virtual memory available if I am not mistaken). if you make this a 
configurable parameter (board based/kconfig based),:
a) old platforms(using old mm components) dont need to use a larger 
virtual memory to maintain "old compatibility"
b) newer platform (with newer multimedia) can use what ever it chooses to be
c) future platforms (with even newer and more bloated mm components) can 
increase it without having to change code.

in short - make it flexible please.
> 
> Thank you, 
> Best regards,
> Hari
> 
>> -----Original Message-----
>> From: Menon, Nishanth
>> Sent: Friday, January 08, 2010 8:28 AM
>> To: Kanigeri, Hari
>> Cc: Ramirez Luna, Omar; linux-omap; Hiroshi Doyu; Ameya Palande; Felipe
>> Contreras; Guzman Lugo, Fernando; Ramos Falcon, Ernesto; Aguilar Pena,
>> Leed
>> Subject: Re: [PATCH 3/8] DSPBRIDGE: Increased DMM size to 256MB
>>
>> Kanigeri, Hari had written, on 01/07/2010 11:16 PM, the following:
>>> Nishant,
>>>
>>> With 64MB we were seeing cases of running out of DSP virtual memory when
>> running multiple Multimedia use cases in parallel at a time.
>> Would it be better that we make this as a board specific memory
>> requirement? not all boards will have the same needs right?
>>
>>> Thank you,
>>> Best regards,
>>> Hari
>>>
>>> -----Original Message-----
>>> From: Menon, Nishanth
>>> Sent: Friday, January 08, 2010 7:48 AM
>>> To: Ramirez Luna, Omar
>>> Cc: linux-omap; Hiroshi Doyu; Ameya Palande; Felipe Contreras; Guzman
>> Lugo, Fernando; Ramos Falcon, Ernesto; Kanigeri, Hari; Aguilar Pena, Leed
>>> Subject: Re: [PATCH 3/8] DSPBRIDGE: Increased DMM size to 256MB
>>>
>>> Ramirez Luna, Omar had written, on 01/07/2010 07:00 PM, the following:
>>>> From: Hari Kanigeri <h-kanigeri2@ti.com>
>>>>
>>>> This patch increases the DMM from 64MB to 256MB.
>>> begs the question: Why?
>>>
>>>> Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
>>>> Signed-off-by: Omar Ramirez Luna <omar.ramirez@ti.com>
>>>> Signed-off-by: Leed Aguilar <leed.aguilar@ti.com>
>>>> ---
>>>>  arch/arm/plat-omap/include/dspbridge/dmm.h |    2 +-
>>>>  drivers/dsp/bridge/pmgr/dmm.c              |    8 ++++----
>>>>  2 files changed, 5 insertions(+), 5 deletions(-)
>>>>
>>>> diff --git a/arch/arm/plat-omap/include/dspbridge/dmm.h
>> b/arch/arm/plat-omap/include/dspbridge/dmm.h
>>>> index 335edf8..af0c35a 100644
>>>> --- a/arch/arm/plat-omap/include/dspbridge/dmm.h
>>>> +++ b/arch/arm/plat-omap/include/dspbridge/dmm.h
>>>> @@ -41,7 +41,7 @@
>>>>  		u32 reserved;
>>>>  	} ;
>>>>
>>>> -#define DMMPOOLSIZE      0x4000000
>>>> +#define DMMPOOLSIZE      0x10000000
>>>>
>>>>  /*
>>>>   *  ======== DMM_GetHandle ========
>>>> diff --git a/drivers/dsp/bridge/pmgr/dmm.c
>> b/drivers/dsp/bridge/pmgr/dmm.c
>>>> index 46c05c6..f878855 100644
>>>> --- a/drivers/dsp/bridge/pmgr/dmm.c
>>>> +++ b/drivers/dsp/bridge/pmgr/dmm.c
>>>> @@ -103,10 +103,10 @@ static struct GT_Mask DMM_debugMask = { NULL,
>> NULL };	/* GT trace variable */
>>>>  static u32 cRefs;		/* module reference count */
>>>>  struct MapPage {
>>>> -	u32   RegionSize:15;
>>>> -	u32   MappedSize:15;
>>>> -	u32   bReserved:1;
>>>> -	u32   bMapped:1;
>>>> +	u64   RegionSize:31;
>>>> +	u64   MappedSize:31;
>>>> +	u64   bReserved:1;
>>>> +	u64   bMapped:1;
>> this does not make much sense meanwhile.. what does this have to do with
>> the dmmpool size increase and what are these unused fields being used for?
>>
>>>>  };
>>>>
>>>>  /*  Create the free list */
>>>
>>
>> --
>> Regards,
>> Nishanth Menon


-- 
Regards,
Nishanth Menon

  reply	other threads:[~2010-01-08 15:33 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-08  1:00 [PATCH 3/8] DSPBRIDGE: Increased DMM size to 256MB Omar Ramirez Luna
2010-01-08  2:17 ` Nishanth Menon
2010-01-08  5:16   ` Kanigeri, Hari
2010-01-08 14:28     ` Nishanth Menon
2010-01-08 15:29       ` Kanigeri, Hari
2010-01-08 15:33         ` Nishanth Menon [this message]
2010-01-08 14:08 ` Felipe Contreras
2010-01-08 14:14   ` Kanigeri, Hari

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4B475046.8020500@ti.com \
    --to=nm@ti.com \
    --cc=Hiroshi.DOYU@nokia.com \
    --cc=ameya.palande@nokia.com \
    --cc=ernesto@ti.com \
    --cc=felipe.contreras@nokia.com \
    --cc=h-kanigeri2@ti.com \
    --cc=leed.aguilar@ti.com \
    --cc=linux-omap@vger.kernel.org \
    --cc=omar.ramirez@ti.com \
    --cc=x0095840@ti.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox