[PATCH v2] ARM: dts: rockchip: temporarily remove emmc hs200 speed from rk3288-veyron-speedy.

[Vagrant Cascadian]

On 2016-10-18, Paul Kocialkowski wrote:
> Le mardi 18 octobre 2016 à 11:21 +0200, Heiko Stübner a écrit :
>> Am Sonntag, 16. Oktober 2016, 21:49:43 schrieb Paul Kocialkowski:
>> > Le mardi 27 septembre 2016 à 13:53 -0700, Vagrant Cascadian a écrit :
>> > > 
>> > > This essentially mimics what was done with rk3288-veyron-minnie in
>> > > commit 984926781122f034d5bc9962815d135b6c4a8e1d.
>> > > 
>> > > The eMMC of the speedy Chromebook also appears to need the same tuning
>> > > workaround, as it frequently fails to recognize the eMMC without it.
>> > 
>> > I have a device where (without this patch) eMMC sometimes fails, with:
>> > [    3.561010] dwmmc_rockchip ff0f0000.dwmmc: Successfully tuned phase to
>> > 175 [    3.571742] mmc2: new HS200 MMC card at address 0001
>> > [    3.571943] mmcblk2: mmc2:0001 HAG2e 14.7 GiB 
>> > [    3.572026] mmcblk2boot0: mmc2:0001 HAG2e partition 1 4.00 MiB
>> > [    3.572107] mmcblk2boot1: mmc2:0001 HAG2e partition 2 4.00 MiB
>> > [    3.572181] mmcblk2rpmb: mmc2:0001 HAG2e partition 3 4.00 MiB
>> > [    3.685647] mmcblk2: error -110 transferring data, sector 0, nr 8, cmd
>> > response 0x900, card status 0x0

>> > And sometimes works, with:
>> > [    3.451058] dwmmc_rockchip ff0f0000.dwmmc: Successfully tuned phase to
>> > 176 [    3.491093] mmc2: new HS200 MMC card at address 0001
>> > [    3.491277] mmcblk2: mmc2:0001 HAG2e 14.7 GiB 
>> > [    3.491345] mmcblk2boot0: mmc2:0001 HAG2e partition 1 4.00 MiB
>> > [    3.491409] mmcblk2boot1: mmc2:0001 HAG2e partition 2 4.00 MiB
>> > [    3.491474] mmcblk2rpmb: mmc2:0001 HAG2e partition 3 4.00 MiB
>> > [    3.493548]  mmcblk2: p1 p2

This is similar behavior to my veyron-speedy. Here are several boot logs
From running 4.8.0-rc7, with some failures and some successes:

  https://cascadia.aikidev.net/~vagrant/veyron-speedy/


>> > However, with this change, it always fails, with:
>> > [    3.322129] mmc_host mmc2: Bus speed (slot 0) = 50000000Hz (slot req
>> > 52000000Hz, actual 50000000HZ div = 0) [    3.333174] mmc2: error -110
>> > whilst initialising MMC card
>> > 
>> > I don't have so much time to investigate this issue, but it's clear that
>> > this patch doesn't fix the issue (and actually worsens it) for my device.

Well, that makes things more complicated.


>> As discussed on IRC we now have varying reports of the emmc working or not 
>> working with and without that patch applied. So it's not really a bandaid fix 
>> and I've thus dropped this patch again.
>
> Thanks for dropping it! For the record, my eMMC shows up as:
> mmcblk2: mmc2:0001 HAG2e 14.7 GiB

Mine looks similar, although reporting as mmc0:

  [    3.166550] mmcblk0: mmc0:0001 HAG2e 14.7 GiB


> Maybe it could help to share what each tested device reports as eMMC model and
> associate that with the current behavior, in spite of getting a clearer idea of
> what issue affects what model.

FWIW, I'm using a veyron-speedy with 4GB of ram, not sure what other
information might be useful to distinguish between different models.

On the bottom of the case, it claims to be a C201P. Near the serial,
it's marked as C201PA-DS02-LG.


live well,
  vagrant
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 800 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20161018/643b245b/attachment.sig>