Re: mmc: sdhci-of-esdhc: fix P2020 errata handling

Re: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Daniel Walker (danielwa)]

Hi,

We got your patch from stable. On p2020 we had mmc issues, and bisected the
issue to your patch.

Both prior patches seems to modify quirks2. 

    Fixes: 05cb6b2a66fa ("mmc: sdhci-of-esdhc: add erratum eSDHC-A001 and A-008358 support")
    Fixes: a46e42712596 ("mmc: sdhci-of-esdhc: add erratum eSDHC5 support")

So I'm not sure if your changes were done correctly.


Daniel

RE: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Y.b. Lu]

Hi Daniel,

Sorry for the trouble.
I think you were saying below patch introduced issue for you.
fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling

Actually it looked fine from code. Probably it resolved its own issue but caused another.
I don't know which kernel you were using. Could you have a check whether my other fix-up patches which may be related to P2020 in your kernel after that one?
I remembered P2020 should work.

80c7482 mmc: sdhci-of-esdhc: fix serious issue clock is always disabled
429d939 mmc: sdhci-of-esdhc: fix transfer mode register reading
1b21a70 mmc: sdhci-of-esdhc: fix clock setting for different controller versions
2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller versions
f667216 mmc: sdhci-of-esdhc: re-implement erratum A-009204 workaround

Thanks a lot.

Best regards,
Yangbo Lu


> -----Original Message-----
> From: Daniel Walker (danielwa) <danielwa@cisco.com>
> Sent: Wednesday, April 1, 2020 4:50 AM
> To: Y.b. Lu <yangbo.lu@nxp.com>
> Cc: stable@vger.kernel.org; Shyam More (shymore) <shymore@cisco.com>;
> xe-linux-external(mailer list) <xe-linux-external@cisco.com>; Ulf Hansson
> <ulf.hansson@linaro.org>
> Subject: Re: mmc: sdhci-of-esdhc: fix P2020 errata handling
> 
> Hi,
> 
> We got your patch from stable. On p2020 we had mmc issues, and bisected
> the
> issue to your patch.
> 
> Both prior patches seems to modify quirks2.
> 
>     Fixes: 05cb6b2a66fa ("mmc: sdhci-of-esdhc: add erratum eSDHC-A001
> and A-008358 support")
>     Fixes: a46e42712596 ("mmc: sdhci-of-esdhc: add erratum eSDHC5
> support")
> 
> So I'm not sure if your changes were done correctly.
> 
> 
> Daniel

Re: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Daniel Walker (danielwa)]

On Wed, Apr 01, 2020 at 05:12:44AM +0000, Y.b. Lu wrote:
> Hi Daniel,
> 
> Sorry for the trouble.
> I think you were saying below patch introduced issue for you.
> fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling

Yes, this patch cased mmc to stop functioning on p2020.

> Actually it looked fine from code. Probably it resolved its own issue but caused another.
> I don't know which kernel you were using. Could you have a check whether my other fix-up patches which may be related to P2020 in your kernel after that one?
> I remembered P2020 should work.

We are using 4.9 stable. Your patch changes from quirks2 to quirks, but the
original patches all use quirks2.

> 80c7482 mmc: sdhci-of-esdhc: fix serious issue clock is always disabled
> 429d939 mmc: sdhci-of-esdhc: fix transfer mode register reading
> 1b21a70 mmc: sdhci-of-esdhc: fix clock setting for different controller versions
> 2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller versions
> f667216 mmc: sdhci-of-esdhc: re-implement erratum A-009204 workaround
 
These are all 5.5 time frame patches, unless it's in stable we don't have them.

Daniel

Re: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Daniel Walker (danielwa)]

On Wed, Apr 01, 2020 at 03:26:35PM +0000, Daniel Walker (danielwa) wrote:
> On Wed, Apr 01, 2020 at 05:12:44AM +0000, Y.b. Lu wrote:
> > Hi Daniel,
> > 
> > Sorry for the trouble.
> > I think you were saying below patch introduced issue for you.
> > fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling
> 
> Yes, this patch cased mmc to stop functioning on p2020.
> 

Have you investigated this ?

Daniel

RE: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Y.b. Lu]

Hi Daniel,

> -----Original Message-----
> From: Daniel Walker (danielwa) <danielwa@cisco.com>
> Sent: Wednesday, April 8, 2020 12:35 AM
> To: Y.b. Lu <yangbo.lu@nxp.com>
> Cc: stable@vger.kernel.org; Shyam More (shymore) <shymore@cisco.com>;
> xe-linux-external(mailer list) <xe-linux-external@cisco.com>; Ulf Hansson
> <ulf.hansson@linaro.org>
> Subject: Re: mmc: sdhci-of-esdhc: fix P2020 errata handling
> 
> On Wed, Apr 01, 2020 at 03:26:35PM +0000, Daniel Walker (danielwa) wrote:
> > On Wed, Apr 01, 2020 at 05:12:44AM +0000, Y.b. Lu wrote:
> > > Hi Daniel,
> > >
> > > Sorry for the trouble.
> > > I think you were saying below patch introduced issue for you.
> > > fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling
> >
> > Yes, this patch cased mmc to stop functioning on p2020.
> >
> 
> Have you investigated this ?

As I said, this patch was proper fix-up, fixing mistake which used host->quirks2. It was host->quirks that should be used.
But after this patch, the other potential issues for P2020 appeared.

I had shared you all other related fix-ups for P2020. If your linux doesn't not have, you need to cherry-pick by yourself.

80c7482 mmc: sdhci-of-esdhc: fix serious issue clock is always disabled
429d939 mmc: sdhci-of-esdhc: fix transfer mode register reading
1b21a70 mmc: sdhci-of-esdhc: fix clock setting for different controller versions
2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller versions
f667216 mmc: sdhci-of-esdhc: re-implement erratum A-009204 workaround

At least, if you try to cherry-pick the two patches, P2020 eSDHC would work again.

2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller versions
f667216 mmc: sdhci-of-esdhc: re-implement erratum A-009204 workaround

Thanks.

> 
> Daniel

Re: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Daniel Walker (danielwa)]

On Wed, Apr 08, 2020 at 05:53:52AM +0000, Y.b. Lu wrote:
> Hi Daniel,
> 
> > -----Original Message-----
> > From: Daniel Walker (danielwa) <danielwa@cisco.com>
> > Sent: Wednesday, April 8, 2020 12:35 AM
> > To: Y.b. Lu <yangbo.lu@nxp.com>
> > Cc: stable@vger.kernel.org; Shyam More (shymore) <shymore@cisco.com>;
> > xe-linux-external(mailer list) <xe-linux-external@cisco.com>; Ulf Hansson
> > <ulf.hansson@linaro.org>
> > Subject: Re: mmc: sdhci-of-esdhc: fix P2020 errata handling
> > 
> > On Wed, Apr 01, 2020 at 03:26:35PM +0000, Daniel Walker (danielwa) wrote:
> > > On Wed, Apr 01, 2020 at 05:12:44AM +0000, Y.b. Lu wrote:
> > > > Hi Daniel,
> > > >
> > > > Sorry for the trouble.
> > > > I think you were saying below patch introduced issue for you.
> > > > fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling
> > >
> > > Yes, this patch cased mmc to stop functioning on p2020.
> > >
> > 
> > Have you investigated this ?
> 
> As I said, this patch was proper fix-up, fixing mistake which used host->quirks2. It was host->quirks that should be used.
> But after this patch, the other potential issues for P2020 appeared.
 
Your including a change into stable which breaks p2020, that's not acceptable. If
more changes are needed to make p2020 stable can you send those additional
patches to stable also ?

Otherwise we need to revert your current patch.

Daniel

RE: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Y.b. Lu]

Thanks Daniel for pointing out it.

Hi Uffe,

As Daniel reported, below commit introduced issue on P2020 platform.
fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling

And the fix-up for the issue is,
2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller versions

It seemed fe0acab was applied to linux-stable for 5.5, 5.4, 4.19, 4.14, 4.9, and 4.4, without the fix-up 2aa3d82.
I tried to cherry-pick the fix-up to all these stable branches, but got many conflicts except 5.5 and 5.4.

May I have your suggestion from safe perspective should I rework the fix-up for these branches, or request to just revert fe0acab.
The patch fe0acab is just for errata handling while the errata are hard to trigger. It is not strongly required.

Thanks a lot.

Best regards,
Yangbo Lu

> -----Original Message-----
> From: Daniel Walker (danielwa) <danielwa@cisco.com>
> Sent: Wednesday, April 8, 2020 11:31 PM
> To: Y.b. Lu <yangbo.lu@nxp.com>
> Cc: stable@vger.kernel.org; Shyam More (shymore) <shymore@cisco.com>;
> xe-linux-external(mailer list) <xe-linux-external@cisco.com>; Ulf Hansson
> <ulf.hansson@linaro.org>
> Subject: Re: mmc: sdhci-of-esdhc: fix P2020 errata handling
> 
> On Wed, Apr 08, 2020 at 05:53:52AM +0000, Y.b. Lu wrote:
> > Hi Daniel,
> >
> > > -----Original Message-----
> > > From: Daniel Walker (danielwa) <danielwa@cisco.com>
> > > Sent: Wednesday, April 8, 2020 12:35 AM
> > > To: Y.b. Lu <yangbo.lu@nxp.com>
> > > Cc: stable@vger.kernel.org; Shyam More (shymore)
> <shymore@cisco.com>;
> > > xe-linux-external(mailer list) <xe-linux-external@cisco.com>; Ulf Hansson
> > > <ulf.hansson@linaro.org>
> > > Subject: Re: mmc: sdhci-of-esdhc: fix P2020 errata handling
> > >
> > > On Wed, Apr 01, 2020 at 03:26:35PM +0000, Daniel Walker (danielwa)
> wrote:
> > > > On Wed, Apr 01, 2020 at 05:12:44AM +0000, Y.b. Lu wrote:
> > > > > Hi Daniel,
> > > > >
> > > > > Sorry for the trouble.
> > > > > I think you were saying below patch introduced issue for you.
> > > > > fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling
> > > >
> > > > Yes, this patch cased mmc to stop functioning on p2020.
> > > >
> > >
> > > Have you investigated this ?
> >
> > As I said, this patch was proper fix-up, fixing mistake which used
> host->quirks2. It was host->quirks that should be used.
> > But after this patch, the other potential issues for P2020 appeared.
> 
> Your including a change into stable which breaks p2020, that's not acceptable.
> If
> more changes are needed to make p2020 stable can you send those additional
> patches to stable also ?
> 
> Otherwise we need to revert your current patch.
> 
> Daniel

Re: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Ulf Hansson]

On Thu, 9 Apr 2020 at 05:42, Y.b. Lu <yangbo.lu@nxp.com> wrote:
>
> Thanks Daniel for pointing out it.
>
> Hi Uffe,
>
> As Daniel reported, below commit introduced issue on P2020 platform.
> fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling
>
> And the fix-up for the issue is,
> 2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller versions
>
> It seemed fe0acab was applied to linux-stable for 5.5, 5.4, 4.19, 4.14, 4.9, and 4.4, without the fix-up 2aa3d82.
> I tried to cherry-pick the fix-up to all these stable branches, but got many conflicts except 5.5 and 5.4.
>
> May I have your suggestion from safe perspective should I rework the fix-up for these branches, or request to just revert fe0acab.
> The patch fe0acab is just for errata handling while the errata are hard to trigger. It is not strongly required.

It sounds to me, like a revert may be the best option, at least for
4.19 and backwards.

Perhaps from 5.4 and onwards, you can send the needed additional
patches as plain backports to stable, as they should be easy to apply,
right!?

[...]

Kind regards
Uffe

RE: mmc: sdhci-of-esdhc: fix P2020 errata handling

[Y.b. Lu]

> -----Original Message-----
> From: Ulf Hansson <ulf.hansson@linaro.org>
> Sent: Thursday, April 9, 2020 6:42 PM
> To: Y.b. Lu <yangbo.lu@nxp.com>
> Cc: Daniel Walker (danielwa) <danielwa@cisco.com>; stable@vger.kernel.org;
> Shyam More (shymore) <shymore@cisco.com>; xe-linux-external(mailer list)
> <xe-linux-external@cisco.com>
> Subject: Re: mmc: sdhci-of-esdhc: fix P2020 errata handling
> 
> On Thu, 9 Apr 2020 at 05:42, Y.b. Lu <yangbo.lu@nxp.com> wrote:
> >
> > Thanks Daniel for pointing out it.
> >
> > Hi Uffe,
> >
> > As Daniel reported, below commit introduced issue on P2020 platform.
> > fe0acab mmc: sdhci-of-esdhc: fix P2020 errata handling
> >
> > And the fix-up for the issue is,
> > 2aa3d82 mmc: sdhci-of-esdhc: fix esdhc_reset() for different controller
> versions
> >
> > It seemed fe0acab was applied to linux-stable for 5.5, 5.4, 4.19, 4.14, 4.9,
> and 4.4, without the fix-up 2aa3d82.
> > I tried to cherry-pick the fix-up to all these stable branches, but got many
> conflicts except 5.5 and 5.4.
> >
> > May I have your suggestion from safe perspective should I rework the fix-up
> for these branches, or request to just revert fe0acab.
> > The patch fe0acab is just for errata handling while the errata are hard to
> trigger. It is not strongly required.
> 
> It sounds to me, like a revert may be the best option, at least for
> 4.19 and backwards.
> 
> Perhaps from 5.4 and onwards, you can send the needed additional
> patches as plain backports to stable, as they should be easy to apply,
> right!?

Thanks a lot for your suggestion, Uffe.
Sent out the fix-up to kernel 5.5. and 5.4. And requested help to revert the issue patch on other kernel versions.

> 
> [...]
> 
> Kind regards
> Uffe