Re: [flashrom] [PATCH] Add programmer for the MSTAR I2C ISP protocol

[Stefan Tauner]

On Thu, 8 May 2014 00:07:29 +0200
Idwer Vollering  wrote:
> Adding the lead developers to CC.

And the mailing list... this should be public IMHO.
> 2014-05-07 23:59 GMT+02:00 Alexandre Boeglin :
> > Le mardi 06 mai 2014 à 16:09, Alexandre Boeglin a écrit:
> > > Le mardi 06 mai 2014 à 15:11, Idwer Vollering a écrit:
> > > > However, using my Dell screen, every file written/read with -p
> > > > mstarddc_spi will have the first byte rotated, so verify will always
> > > > fail:
> > >
> > > I have other screens available, I'll check if they use MSTAR SoCs…
> >
> > Well, I tried with a BenQ FP93GX I had, which also seem to have the same
> > ISP port.
> >
> > I tried to probe it, but with no luck.
> >
> > After closer inspection, it seems that I have the same rotation issue,
> > even in the RES command, which makes it impossible to identify the flash
> > chip:
> >
> > Probing for PMC Pm25LV010, 128 kB: programmer_map_flash_region: mapping
> > flash chip from 0x00000000fffe0000 to 0x0000000000000000
> > RES returned 0x9d 0x7c 0x7f. probe_spi_res3: id1 0x9d7c, id2 0x7f
> >
> > It returns "0x9d 0x7c 0x7f", whereas I think it should return "0x7f 0x9d
> > 0x7c"…
> >
> > But then, without documentation from MSTAR, I don't really know how to
> > differentiate between screens that have the bug, and those that don't.

We could add a programmer parameter and let the user decide/override
until we understand this better. Does the DDC tunnel still work with a
broken firmware or is an external programmer needed if the first try to
modify the firmware fails and the SoC inside the device is reset?

[PATCH] Add programmer for the MSTAR I2C ISP protocol

[Alexandre Boeglin]

Le jeudi 08 mai 2014 à 00:23, Stefan Tauner a écrit:
> We could add a programmer parameter and let the user decide/override
> until we understand this better. Does the DDC tunnel still work with a
> broken firmware or is an external programmer needed if the first try to
> modify the firmware fails and the SoC inside the device is reset?

Well, it seems it depends. One user of this mstarddc programmer had an
error during the write operation due to (it seems) an I2C driver working
at less than half of the normal 100kb/s I2Cspeed, but even after
power-cycling the screen, he managed to successfully reflash it using
another laptop.

But in the mean time, the screen was not working properly, and I'm
guessing it was maybe in some kind of emergency mode, maybe an
in-silicon bootloader that allows access to the ISP port.

OTOH, another person, using the "official" MSTAR tool was unable to
reflash his screen after the initial failure. But I didn't get him to
try the programmer I wrote, so I don't know for sure what his problem
was (maybe even just a bug in the official tool).

Anyway, I slightly modified the programmer in the patch revision I sent
this morning. It does not send the reset command (which causes the
screen to exit ISP mode and re-read the flash firmware) if an error
occured. So, as long as the screen remains powered, another attempt can
be made.

But then, with the limited amount of info I had, most of it is just
guesses.

Best regards,
Alex