Re: [PATCH v3 4/8] m25p80: Add the mx25l25635f SFPD table

[Cédric Le Goater]

On 10/10/22 11:58, Michael Walle wrote:
> Am 2022-10-10 08:23, schrieb Cédric Le Goater:
> > On 10/7/22 16:44, Francisco Iglesias wrote:
>
> > > > --- a/hw/block/m25p80.c
> > > > +++ b/hw/block/m25p80.c
> > > > @@ -234,6 +234,8 @@ static const FlashPartInfo known_devices[] = {
> > > >       { INFO("mx25l12855e", 0xc22618,      0,  64 << 10, 256, 0) },
> > > >       { INFO6("mx25l25635e", 0xc22019,     0xc22019,  64 << 10, 512, 0),
> > > >         .sfdp_read = m25p80_sfdp_mx25l25635e },
> > > > +    { INFO6("mx25l25635f", 0xc22019,     0xc22019,  64 << 10, 512, 0),
> > >
> > > I think I'm not seeing the extended id part in the datasheet I've found so
> > > might be that you can switch to just INFO and _ext_id 0 above
> >
> > This was added by commit 6bbe036f32dc ("m25p80: Return the JEDEC ID twice for
> > mx25l25635e") to fix a real breakage on HW.
>
>  From my experience, the ID has a particular length, at least three bytes
> and if you read past that length for some (all?) devices the id bytes just
> get repeated. I.e. the counter in the device will just wrap to offset 0
> again. 
At the time, I did some specific tests on these mx25l25635e chips and
this how they behaved. I can not check anymore as I don't have access
to these systems.

> If you want to emulate the hardware correctly, you would have to
> take that into consideration.

But this is not the case for all chips, most return 0 and don't
wrap around. I just did on w25q512jvq and w25q256 chips with a 6.0.

> But I don't think it's worth it, OTOH there seems to be some broken
> software which rely on that (undefined?) behavior.

I can't really tell :/ The driver (kernel based Linux 2.6.28) reads
4 bytes and expect the last to be the Manufacturer id: 0xC2. That's
valid for this chip.

Setting an extended ID in the flash info is an alternative :

  { INFO("mx25l25635e", 0xc22019,     0xc200,  ...

which does not seem to break other machines using mx25l25635e, with
recent kernels at least.

To reproduce, grab :

  http://smbserver.frankfurt.de.velia.net/ipmi/SMT_X11_158.bin

Run :

  qemu-system-arm -M supermicrox11-bmc -nic user -drive 
file=./SMT_X11_158.bin,format=raw,if=mtd -nographic -serial mon:stdio
  ...
  BMC flash ID:0xc21920c2
   platform_flash: MX25L25635E (32768 Kbytes)
  Creating 7 MTD partitions on "spi0.0":
  ...

C.