On Tue, Jun 16, 2020 at 05:36:47PM +0200, Freek de Kruijf wrote:
> Op dinsdag 16 juni 2020 14:12:12 CEST schreef Michal Suchánek:
> > On Tue, Jun 16, 2020 at 02:02:44PM +0200, Matthias Brugger wrote:
> > > +nicolas +michal
> > > 
> > > On 16/06/2020 14:00, Freek de Kruijf wrote:
> > > > I am doing some testing with the newest Tumbleweed images for the
> > > > Raspberry Pi 4B. The micro-SD card I was using for the KDE version
> > > > showed a rather slow system. Long waiting between commands.
> > > > 
> > > > After trying to change the default language with "yast language", which
> > > > pulls in quite a lot of new packages, the system was very slow. Finally
> > > > dmesg showed error messages complaining:
> > > > 
> > > > mmc1 timeout waiting for hardware interrupt
> > > > 
> > > > Searching for a solution I found the following patch:
> > > > 
> > > > https://patchwork.kernel.org/patch/10219153/
> > > > 
> > > > The main cause seems to be the behaviour of certain SD cards of which I
> > > > apparently have one. I do have a similar card, same brand, same kind and
> > > > same size, which does not show this behaviour.
> > > > 
> > > > So my question is: Has this patch been implemented in the newest version
> > > > of
> > > > Tumbleweed for the RPi4?
> > 
> > Most likely.
> > 
> > > > If not, would it solve this problem?
> > 
> > No.
> > 
> > In the case that matters it seems most of the system is locked up. I
> > suspect something essential like the DMA controller locks up rather than
> > the MMC controller itself. I can see now why Chinese chips employ
> > multiple DMA cotrollers with key components like storage having a
> > private DMA each.
> > 
> > Given the low general interest and low occurence of these broken cards I
> > gave up on debugging this.
> > 
> > I may be able to find the one card that exhibits the problem after
> > covid-19 situation stabilizes some more but I did not seriously try to
> > debug a low-level problem like this yet.
> > 
> > Thanks
> > 
> > Michal
> 
> I can send you my SD card if you wish. It behaves normally in my desktop 
> computer. Send me your address by private email. It is a SanDisk Ultra 16 GB.

I have my broken Samsung EVO (orange) 16GB, thank you very much.

It also works fine in my desktop and other boards and fails in multiple
rPi 3 boards. Did not to try with Pi 4 yet.

I tried to get a few more EVO cards but none of them exhibits the
problem. This is rather rare issue.

Thanks

Michal
-- 
To unsubscribe, e-mail: [email protected]
To contact the owner, e-mail: [email protected]

Reply via email to