Re: [Qemu-ppc] [PATCH v5 1/4] ppc4xx_i2c: Rewrite to model hardware more closely

[BALATON Zoltan]

On Thu, 28 Jun 2018, Cédric Le Goater wrote:
> I have tried an ISO from :
>
>         http://aros.sourceforge.net/nightly1.php
>
> which boots and writes :
>
>  U-Boot 2010.06.05 (Oct 22 2017 - 17:40:02)
>
>  CPU:   AMCC PowerPC 460EX Rev. B at 1150 MHz (PLB=230 OPB=115 EBC=115)
>         No Security/Kasumi support
>         Bootstrap Option A - Boot ROM Location EBC (8 bits)
>         Internal PCI arbiter disabled
>         32 kB I-Cache 32 kB D-Cache
>  Board: Sam460ex, PCIe 4x + SATA-2
>  I2C:   ready
>  DRAM:  512 MiB (ECC not enabled, 460 MHz, CL0)
>  *** Warning - bad CRC, using default environment
>
>  PCI:   Bus Dev VenId DevId Class Int
>          00  06  126f  0501  0380  00
>  PCIE1: successfully set as root-complex
>  Net:   ppc_4xx_eth0
>  FPGA:  Revision 00 (20 0-00-00)
>  SM502: found
>  VGA:   NO CARDS

This is the U-Boot firmware that prints this.

> and on the graphic window :
>
>  FLB: no SLB found in any of the designated boot sources, returning to u-boot.

And also this is from U-Boot saying it could not find bootable image so it 
did not boot. (One could test i2c from u_boot with date and i2c commands 
as well.)

> the graphical bool loader menu works but I couldn't exercise much more the
> system.

Have you by chance used the -cdrom shortcut? That tries to connect CDROM 
to secondary master but the sii3112 this board uses only has 2 SATA ports 
so this won't work. You need to use the long options like

-drive if=none,id=cd,file=aros-sam440-ppc.iso,format=raw \
-device ide-cd,drive=cd,bus=ide.0

I'm not sure how to fix -cdrom or have it print a meaningful warning or 
error.

> I then used the 4.5 kernel from :
>
>  http://www.supertuxkart-amiga.de/amiga/sam.html#downloads
>
> with this command line :
>
>   qemu-system-ppc -M sam460ex  -serial mon:stdio -nodefaults -kernel 
> ./Sam460ex-4.5.0/boot/uImage-sam460ex-4.5.0 -dtb Sam460ex-4.5.0/boot/canyonlands.dtb 
> -append "console=ttyS0,119200"
>
> output was :
>
>
>  [    0.000000] Using Canyonlands machine description
>  [    0.000000] Linux version 4.5.0-sam460ex (address@hidden) (gcc version 
> 5.3.1 20160205 (Ubuntu 5.3.1-8ubuntu2) ) #1 PREEMPT Mon Mar 14 06:27:13 AST 2016
>  ...
>  [    2.194282] i2c /dev entries driver
>  [    2.201326] rtc-m41t80 0-0068: rtc core: registered m41t80 as rtc0
>  [    2.202357] ibm-iic 4ef600700.i2c: using standard (100 kHz) mode
>  [    2.203796] ibm-iic 4ef600800.i2c: using standard (100 kHz) mode
>  ...
>  [    2.269934] rtc-m41t80 0-0068: setting system clock to 2018-06-28 09:49:28 
> UTC (1530179368)
>
>
> So I would say from the review of the code and the tests that we are fine.
>
> Could please document a little more the register and bit definitions and
> remove the raw hex values when possible ?

I'll check what can be dont but I think by now only the extsts state are 
not #define-d which are also not really implemented and they are not bits 
but 3 bit values so not sure what's the best way to define them. Other hex 
values are some reserved bit masks I think that probably should not need 
to be #defined.

> With these addressed :
>
> Reviewed-by: Cédric Le Goater <address@hidden>
>
> Thanks,
>
> C.
Thanks for testing,
BALATON Zoltan