Re: [Qemu-devel] barebox on Canon A1100
On Fri, Jan 04, 2019 at 01:56:41AM +0300, Antony Pavlov wrote:
> Hi Sascha!
>
> I have some good news and some bad news.
>
> First the good news.
>
> Barebox was used for Day 18 of QEMU 2018 advent
> (https://www.qemu-advent-calendar.org/2018/).
>
> Here is a quote:
>
> > This is an image for QEMU's emulated Canon A1100 Arm machine:
> > Play 2048 in the Barebox firmware environment.
Hey nice :)
>
> The bad news.
>
> Current master branch barebox for Canon A1100 does not work under qemu,
> barebox v2017.02.0 was used for qemu advent.
>
> Thomas Huth reported the problem in the qemu-devel mailing
> list two months ago. Please see
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg571887.html
> for details.
>
> The commit a43e2bbc468a71 breaks barebox for Canon A1100:
>
> > commit a43e2bbc468a71fb8825e8acafb7fac7ad7c6dad
> > Author: Sascha Hauer
> > Date: Wed Mar 7 10:04:07 2018 +0100
> >
> >ARM: return positive offset in get_runtime_offset()
>
> Any suggestion?
Yes, try again with a recent barebox. There was a bug in this commit that
should be fixed with the following:
| commit 0230365df6601cd561c9e3e368c96222076ec9f5
| Author: Sascha Hauer
| Date: Wed Dec 12 08:55:31 2018 +0100
|
| ARM: fix setup_c when runtime offset is != 0
|
| The runtime offset has to be added to the memcpy source address and
| substracted from the return address. This should have been changed in
| a43e2bbc46 which changed from returning the negative runtime offset
| into changing the positive runtime offset. Instead a43e2bbc46 only
| changed a zero substraction ("subs r4, r0, #0") into a zero addition
| ("adds r4, r0, #0") which was used as a equal to zero test and changed
| nothing. This part is reverted here.
|
| Fixes wrong copy / return locations when setup_c is called with
| different runtime and link addresses.
|
| fixes: a43e2bbc46 ("ARM: return positive offset in get_runtime_offset()")
|
| Signed-off-by: Sascha Hauer
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0|
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917- |
Re: [Qemu-devel] barebox on Canon A1100
On Fri, 4 Jan 2019 09:17:41 +0100
Sascha Hauer wrote:
> On Fri, Jan 04, 2019 at 01:56:41AM +0300, Antony Pavlov wrote:
> > Hi Sascha!
> >
> > I have some good news and some bad news.
> >
> > First the good news.
> >
> > Barebox was used for Day 18 of QEMU 2018 advent
> > (https://www.qemu-advent-calendar.org/2018/).
> >
> > Here is a quote:
> >
> > > This is an image for QEMU's emulated Canon A1100 Arm machine:
> > > Play 2048 in the Barebox firmware environment.
>
> Hey nice :)
>
> >
> > The bad news.
> >
> > Current master branch barebox for Canon A1100 does not work under qemu,
> > barebox v2017.02.0 was used for qemu advent.
> >
> > Thomas Huth reported the problem in the qemu-devel mailing
> > list two months ago. Please see
> > https://www.mail-archive.com/qemu-devel@nongnu.org/msg571887.html
> > for details.
> >
> > The commit a43e2bbc468a71 breaks barebox for Canon A1100:
> >
> > > commit a43e2bbc468a71fb8825e8acafb7fac7ad7c6dad
> > > Author: Sascha Hauer
> > > Date: Wed Mar 7 10:04:07 2018 +0100
> > >
> > >ARM: return positive offset in get_runtime_offset()
> >
> > Any suggestion?
>
> Yes, try again with a recent barebox. There was a bug in this commit that
> should be fixed with the following:
>
> | commit 0230365df6601cd561c9e3e368c96222076ec9f5
> | Author: Sascha Hauer
> | Date: Wed Dec 12 08:55:31 2018 +0100
> |
> | ARM: fix setup_c when runtime offset is != 0
> |
> | The runtime offset has to be added to the memcpy source address and
> | substracted from the return address. This should have been changed in
> | a43e2bbc46 which changed from returning the negative runtime offset
> | into changing the positive runtime offset. Instead a43e2bbc46 only
> | changed a zero substraction ("subs r4, r0, #0") into a zero addition
> | ("adds r4, r0, #0") which was used as a equal to zero test and changed
> | nothing. This part is reverted here.
> |
> | Fixes wrong copy / return locations when setup_c is called with
> | different runtime and link addresses.
> |
> | fixes: a43e2bbc46 ("ARM: return positive offset in
> get_runtime_offset()")
> |
> | Signed-off-by: Sascha Hauer
Thanks for noting!
I have just tested recent master branch barebox with qemu-3.1.0 from Debian
unstable.
It works!
Here is the log:
$ qemu-system-arm --version
QEMU emulator version 3.1.0 (Debian 1:3.1+dfsg-2)
Copyright (c) 2003-2018 Fabrice Bellard and the QEMU Project developers
$ qemu-system-arm -M canon-a1100 -display none -serial stdio -bios
barebox.canon-a1100.bin
Switch to console [cs0]
barebox 2018.12.0-00140-g64fc4ac1b5 #1 Fri Jan 4 13:36:42 MSK 2019
Board: Canon PowerShot A1100 IS
digic-gpio c022.gpio.of: probed gpiochip-1 with base 0
cfi_flash f800.fl...@f800.of: found cfi flash at 0xf800, size 4 MiB
malloc space: 0x0010 -> 0x002f (size 2 MiB)
running /env/bin/init...
/env/bin/init not found
canon-a1100 > /
--
Best regards,
Antony Pavlov
[Qemu-devel] barebox on Canon A1100
Hi Sascha! I have some good news and some bad news. First the good news. Barebox was used for Day 18 of QEMU 2018 advent (https://www.qemu-advent-calendar.org/2018/). Here is a quote: > This is an image for QEMU's emulated Canon A1100 Arm machine: > Play 2048 in the Barebox firmware environment. The bad news. Current master branch barebox for Canon A1100 does not work under qemu, barebox v2017.02.0 was used for qemu advent. Thomas Huth reported the problem in the qemu-devel mailing list two months ago. Please see https://www.mail-archive.com/qemu-devel@nongnu.org/msg571887.html for details. The commit a43e2bbc468a71 breaks barebox for Canon A1100: > commit a43e2bbc468a71fb8825e8acafb7fac7ad7c6dad > Author: Sascha Hauer > Date: Wed Mar 7 10:04:07 2018 +0100 > >ARM: return positive offset in get_runtime_offset() Any suggestion? -- Best regards, Antony Pavlov
