[PATCH] mtd: rawnand: gpmi: Set WAIT_FOR_READY timeout based on program/erase times

[Greg Kroah-Hartman]

On Fri, Jul 15, 2022 at 07:27:02AM +0200, Tomasz Moń wrote:
> On Mon, 2022-07-11 at 11:12 +0200, Tomasz Moń wrote:
> > On Fri, 2022-07-01 at 13:03 +0200, Sascha Hauer wrote:
> > > 06781a5026350 Fixes the calculation of the DEVICE_BUSY_TIMEOUT register
> > > value from busy_timeout_cycles. busy_timeout_cycles is calculated wrong
> > > though: It is calculated based on the maximum page read time, but the
> > > timeout is also used for page write and block erase operations which
> > > require orders of magnitude bigger timeouts.
> > > 
> > > Fix this by calculating busy_timeout_cycles from the maximum of
> > > tBERS_max and tPROG_max.
> > 
> > 06781a5026350 was merged in v5.19-rc4 and then was picked up by several
> > stable kernels, including v5.15.51. After we have upgraded to v5.15.51
> > we have observed the issue that Sascha mentioned in his email [1].
> > 
> > As the v5.19-rc6 was released yesterday and this fix is still not
> > applied, the v5.19-rc6 (and all stable kernels that picked up the
> > backport) causes NAND flash data loss on imx targets.
> > 
> > I have backported this patch to our internal v5.15.51 based kernel on
> > 4th July 2022 and I can confirm that it does indeed solve the NAND data
> > loss on imx targets.
> > 
> > Is it possible for this patch to make it to the v5.19-rc7?
> 
> No response, so sending the email to more people so the voice is heard.
> Sorry if this is not the proper way, but I think the issue is serious.
> 
> Current prepatch kernels starting with v5.19-rc4 and stable kernels
> starting with v5.4.202. v5.10.127, v5.15.51, v5.18.8 contain a
>   "[PATCH] [REALLY REALLY BROKEN] mtd: rawnand: gpmi: Fix setting busy timeout setting"
> that is wreaking havoc to i.MX[678] or i.MX28 devices with NAND
>   "** THIS PATCH WILL CAUSE DATA LOSS ON YOUR NAND!! **" [1]
> 
> The solution is to either:
>   * Revert 06781a5026350 ("mtd: rawnand: gpmi: Fix setting busy timeout
> setting") and all its cherry-picks to stable branches, *OR*
>   * Apply the fix ("mtd: rawnand: gpmi: Set WAIT_FOR_READY timeout
> based on program/erase times") [2]
> 
> Please do whatever you see fit.

I can do do a stable release with this reverted, but I really expected
to see the fix in linux-next by now at the very least.  Does this driver
not have an active maintainer and subsystem maintainer for some reason?

thanks,

greg k-h