> CQHCI has its own DMA descriptors, so maybe a similar change is needed for CQHCI?
I didn't find this limitation in synopsys DW IP databook.
>> I found it on page 84
-----Original Message-----
From: Shawn Lin <shawn.lin@xxxxxxxxxxxxxx>
Sent: Tuesday, November 8, 2022 5:15 PM
To: Adrian Hunter <adrian.hunter@xxxxxxxxx>
Cc: shawn.lin@xxxxxxxxxxxxxx; linux-mmc <linux-mmc@xxxxxxxxxxxxxxx>; Jyan Chou [周芷安] <jyanchou@xxxxxxxxxxx>; Jisheng.Zhang@xxxxxxxxxxxxx
Subject: Re: Question about EMMC CMDQ HOST CONTROLLER INTERFACE (CQHCI) DRIVER
+ Jisheng Zhang
On 2022/11/8 16:34, Adrian Hunter wrote:
> On 8/11/22 10:28, Shawn Lin wrote:
>> On 2022/11/8 16:20, Adrian Hunter wrote:
>>> On 8/11/22 09:25, Jyan Chou [周芷安] wrote:
>>>> Hello Adrian Hunter,
>>>>
>>>> We are now using the upstream code of EMMC CMDQ HOST CONTROLLER INTERFACE (CQHCI) DRIVER, but we found that the existing driver cannot support the limitation of Synopsys IP.
>>>>
>>>> Synopsys IP has a description on their data book " While using DMA, the host memory data buffer size and start address must not exceed 128 MB".
>>
>> Synopsys-based SDHCI IP does have a limitation of 128MB boundary. But
>> it has already been solved by upstream driver.
>>
>> FYI:
>>
>> commit b85c997d2cfefe7d1f706b85ae46e35a50e3131c ("mmc:
>> sdhci-of-dwcmshc: solve 128MB DMA boundary limitation")
>
> CQHCI has its own DMA descriptors, so maybe a similar change is needed for CQHCI?
I didn't find this limitation in synopsys DW IP databook.
>
>>
>>>>
>>>> I am wondering whether there is a method or patch that can fix this boundary limitation.
>>>> Thanks.
>>>> Best Regards,
>>>> Jyan Chou
>>>
>>> Hello Jyan Chou
>>>
>>> I am not clear on what the exact limitation is. The driver never uses buffers as big as 128 MB.
>>> To restrict DMA to low memory addresses a DMA mask can be used.
>>>
>>> But perhaps you mean not to cross a 128 MB boundary?
>>>
>>> Please cc your questions to the linux kernel mmc mailing list:
>>> linux-mmc@xxxxxxxxxxxxxxx because others can answer too.
>>>
>>> Regards
>>> Adrian
>
------Please consider the environment before printing this e-mail.
