Re: [PATCH] [RFC] Revert "mmc: core: Fixup support for writeback-cache for eMMC and SD"

[Marek Vasut <marex@denx.de>]

On 5/31/23 08:01, Avri Altman wrote:
>> This reverts commit 08ebf903af57cda6d773f3dd1671b64f73b432b8.
>>
>> On STM32MP153C DHCOR DRC Compact with microSD card Kingston Canvas
>> Go! Plus MicroSDXC I 64 GiB U3/V30/A2 SDCG3/64GB it is no longer possible
>> to mount ext4 filesystem.
>>
>> Git bisect points to this commit which is being reverted here. A bit of
>> investigation shows that for non-working microSD cards
>>    cache_enabled=TRUE fua_enabled=false
>> For working microSD card (Kingston Canvas React Plus, MicroSDXC II
>> 64 GiB U3/V90/A1 SDCR2/64GB)
>>    cache_enabled=FALSE fua_enabled=false
>>
>> It seems enabling the cache for SD cards causes the issue. Before the reverted
>> commit, the cache was not reported enabled for SD cards. I am not sure
>> whether such blunt revert is the correct solution however, any ideas ?
> The commit you are reverting has a couple of fixes tags - so reverting it will re-introduce those issues it aims to fix.
> If I recall correctly, I asked our sd system engineering to follow up on this,
> And they had no problem with the logic.
> 
> I think that we need more data to establish the root cause of this issue,
> Provided that it is not a specific-vendor issue.

I just sent Adrian the command log, but I don't see anything obvious in it.