public inbox for u-boot@lists.denx.de
 help / color / mirror / Atom feed
From: Hong Xu <hong.xu@atmel.com>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH v2] ARM926ejs: Add routines to invalidate D-Cache
Date: Fri, 05 Aug 2011 14:17:18 +0800	[thread overview]
Message-ID: <4E3B8AEE.8050305@atmel.com> (raw)
In-Reply-To: <4E3B7B6C.5060507@emk-elektronik.de>

H Reinhard,

On 08/05/2011 01:11 PM, Reinhard Meyer wrote:
> Dear Hong Xu,
>> After DMA operation, we need to maintain D-Cache coherency.
>> We need to clean cache (write back the dirty lines) and then
>> make the cache invalidate as well(hence CPU will fetch data
>> written by DMA controller from RAM).
>>
>> Tested on AT91SAM9261EK with Peripheral DMA controller.
>>
>> Signed-off-by: Hong Xu<hong.xu@atmel.com>
>> Tested-by: Elen Song<elen.song@atmel.com>
>> CC: Albert Aribaud<albert.aribaud@free.fr>
>> CC: Heiko Schocher<hs@denx.de>
>> ---
>> Changes since v1
>> ~ Per Albert's suggestion, add invalidate_dcache_range originally defined
>> in include/common.h
>>
>> arch/arm/lib/cache.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++
>> 1 files changed, 46 insertions(+), 0 deletions(-)
>>
>> diff --git a/arch/arm/lib/cache.c b/arch/arm/lib/cache.c
>> index 92b61a2..0436443 100644
>> --- a/arch/arm/lib/cache.c
>> +++ b/arch/arm/lib/cache.c
>> @@ -53,3 +53,49 @@ void __flush_dcache_all(void)
>> }
>> void flush_dcache_all(void)
>> __attribute__((weak, alias("__flush_dcache_all")));
>> +
>> +void __invalidate_dcache_range(unsigned long start, unsigned long stop)
>> +{
>> + int cache_line_len;
>> + unsigned long mva;
>> +
>> +#ifdef CONFIG_ARM926EJS
>> +
>> +#ifdef CONFIG_SYS_CACHELINE_SIZE
>> + cache_line_len = CONFIG_SYS_CACHELINE_SIZE;
>> +#else
>> + cache_line_len = 32;
>> +#endif
>> + /*
>> + * If start and stop are not aligned to cache-line,
>> + * the adjacent lines will be cleaned
>> + */
>> + if ((start& (cache_line_len - 1)) != 0)
>> + asm("mcr p15, 0, %0, c7, c10, 1" : : "r" (start));
>> + if ((stop& (cache_line_len - 1)) != 0)
>> + asm("mcr p15, 0, %0, c7, c10, 1" : : "r" (stop));
>
> Why so complicated?

In case of mis-aligned (start, stop), this is to clean the cache line, 
aka. write back the potential dirty lines before successive invalidating

BR,
Eric

> How about:
> /* round down to the start of the cache line */
> start &= (cache_line_len - 1);
> /* round up to the end of the cache line */
> note: if, what I assume, the range to be invalidated is
> [start, stop) - that means stop is the first address not to be
> invalidated, the next statement is not necessary
> stop |= (cache_line_len - 1);
> while (start < stop) {
> asm("mcr p15, 0, %0, c7, c6, 1" : : "r"(start));
> start += cache_line_len;
> }
> /* Drain the WB */
> asm("mcr p15, 0, %0, c7, c10, 4" : : "r" (0));
>> +#endif
>
> Best Regards,
> Reinhard

  reply	other threads:[~2011-08-05  6:17 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-08-05  4:44 [U-Boot] [PATCH v2] ARM926ejs: Add routines to invalidate D-Cache Hong Xu
2011-08-05  5:11 ` Reinhard Meyer
2011-08-05  6:17   ` Hong Xu [this message]
2011-08-05  6:22     ` Albert ARIBAUD
2011-08-05  6:13 ` Albert ARIBAUD
2011-08-05  6:38   ` Hong Xu
2011-08-05  6:46     ` Albert ARIBAUD
2011-08-05  7:02       ` Hong Xu
2011-08-05  7:10       ` Aneesh V
2011-08-05  9:20         ` Albert ARIBAUD
2011-08-05  9:56           ` Aneesh V
2011-08-05 10:33         ` Hong Xu
2011-08-05 10:47           ` Aneesh V
2011-08-05 11:03             ` Albert ARIBAUD
2011-08-05 11:23               ` Reinhard Meyer
2011-08-05 11:26                 ` Albert ARIBAUD
2011-08-05 11:51               ` Aneesh V
2011-08-05 13:17                 ` Albert ARIBAUD
2011-08-05 14:59                   ` Aneesh V
2011-08-07  6:55                     ` Albert ARIBAUD
2011-08-08  8:24                       ` Aneesh V
2011-08-08  9:39                         ` Albert ARIBAUD
2011-08-08  9:51                           ` Aneesh V
2011-08-08  9:59                           ` Reinhard Meyer
2011-08-08 10:12                             ` Aneesh V
2011-08-08 10:25                               ` Reinhard Meyer
2011-08-08 10:27                                 ` Aneesh V
2011-08-08 11:05                                   ` Albert ARIBAUD
2011-08-05 23:04                 ` J. William Campbell
2011-08-07  7:07                   ` Albert ARIBAUD

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=4E3B8AEE.8050305@atmel.com \
    --to=hong.xu@atmel.com \
    --cc=u-boot@lists.denx.de \
    /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