[BUG] sandbox: './u-boot -l ' fails

[Heinrich Schuchardt]

On 9/28/20 3:42 PM, Simon Glass wrote:
> Hi Heinrich,
>
> On Mon, 28 Sep 2020 at 07:30, Heinrich Schuchardt <xypron.glpk at gmx.de> wrote:
>>
>> On 28.09.20 15:22, Simon Glass wrote:
>>> Hi Heinrich,
>>>
>>> On Mon, 28 Sep 2020 at 05:31, Heinrich Schuchardt <xypron.glpk at gmx.de> wrote:
>>>>
>>>> On 28.09.20 06:46, Heinrich Schuchardt wrote:
>>>>> Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <sjg at chromium.org>:
>>>>>> Hi Heinrich,
>>>>>>
>>>>>> On Sat, 19 Sep 2020 at 13:48, Heinrich Schuchardt <xypron.glpk at gmx.de>
>>>>>> wrote:
>>>>>>>
>>>>>>> Hello Simon,
>>>>>>>
>>>>>>> when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run
>>>>>> out
>>>>>>> of the box?
>>>>>>>
>>>>>>> $ ./u-boot -l -d arch/sandbox/dts/sandbox.dtb
>>>>>>
>>>>>> For the record you should be able to use -D to get the same effect as
>>>>>> your -d above.
>>>>>>
>>>>>>>
>>>>>>> U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39
>>>>>> +0200)
>>>>>>>
>>>>>>> Model: sandbox
>>>>>>> DRAM:  128 MiB
>>>>>>>
>>>>>>> Warning: host_lo MAC addresses don't match:
>>>>>>> Address in ROM is               26:4b:ca:6c:98:f4
>>>>>>> Address in environment is       00:00:11:22:33:44
>>>>>>>
>>>>>>> Warning: host_virbr0 MAC addresses don't match:
>>>>>>> Address in ROM is               ee:3e:c9:ce:1f:9c
>>>>>>> Address in environment is       00:00:11:22:33:45
>>>>>>>
>>>>>>> Warning: host_docker0 MAC addresses don't match:
>>>>>>> Address in ROM is               c2:85:07:7b:9a:18
>>>>>>> Address in environment is       00:00:11:22:33:46
>>>>>>> WDT:   Not found!
>>>>>>> MMC:
>>>>>>>
>>>>>>> No output after this point.
>>>>>>>
>>>>>>> The problem also exists with U-Boot v2020.07, v2019.10, v2018.11.
>>>>>>>
>>>>>>> CONFIG_SANDBOX_SDL=y
>>>>>>>
>>>>>>> SDL_InitSubSystem() never returns. It is looping somewhere in
>>>>>> U-Boot's
>>>>>>> __serial_getc(). I wonder how it gets there without returning from
>>>>>> the
>>>>>>> function.
>>>>>>>
>>>>>>> I compiled SDL2.cpp from
>>>>>>> https://gist.github.com/miguelmartin75/6946310#file-sdl2-cpp-L18
>>>>>>> with
>>>>>>>
>>>>>>> g++ SDL2.cpp -D_REENTRANT -I/usr/include/SDL2 -lSDL2 -lGL -o test
>>>>>>>
>>>>>>> and it runs fine showing an X11 windows with red background.
>>>>>>>
>>>>>>> So there seems to be no general problem with the SDL2 library.
>>>>>>
>>>>>> I hit this myself on another computer and it turned out to be that SDL
>>>>>> defined getc(), as does U-Boot, and things get confused. At least I
>>>>>> think it is getc.
>>>>>>
>>>>>> I hacked around with changing the name of getc (I think it was getc)
>>>>>> in U-Boot and the problem went away.
>>>>>
>>>>> Should we include a patched SDL2 with U-Boot for static linking?
>>>>>
>>>>
>>>> I cannot find a symbol getc() nor a reference to it in the SDL
>>>> repository. getc() is defined in stdio.h.
>>>>
>>>> Our getc() takes no argument, while the stdio one wants a FILE *.
>>>>
>>>> But changing the U-Boot definition to "int getc(void *)" does not solve
>>>> the issue.
>>>
>>> I just tried it on the machine where it doesn't work:
>>>
>>> $ gdb --args /tmp/b/sandbox/u-boot -l -D
>>> GNU gdb (Debian 9.2-1) 9.2
>>> Copyright (C) 2020 Free Software Foundation, Inc.
>>> License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
>>> This is free software: you are free to change and redistribute it.
>>> There is NO WARRANTY, to the extent permitted by law.
>>> Type "show copying" and "show warranty" for details.
>>> This GDB was configured as "x86_64-linux-gnu".
>>> Type "show configuration" for configuration details.
>>> For bug reporting instructions, please see:
>>> <http://www.gnu.org/software/gdb/bugs/>.
>>> Find the GDB manual and other documentation resources online at:
>>>     <http://www.gnu.org/software/gdb/documentation/>.
>>>
>>> For help, type "help".
>>> Type "apropos word" to search for commands related to "word"...
>>> Reading symbols from /tmp/b/sandbox/u-boot...
>>> (gdb) br getc
>>> Breakpoint 1 at 0x5b6d6: getc. (2 locations)
>>> (gdb) r
>>> Starting program: /tmp/b/sandbox/u-boot -l -D
>>> [Thread debugging using libthread_db enabled]
>>> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
>>>
>>>
>>> U-Boot 2020.10-rc5-00196-g0ac83d080a0 (Sep 28 2020 - 07:11:52 -0600)
>>>
>>> Model: sandbox
>>> DRAM:  128 MiB
>>>
>>> Warning: host_lo MAC addresses don't match:
>>> Address in ROM is 1a:34:9b:48:aa:53
>>> Address in environment is 00:00:11:22:33:44
>>> WDT:   Not found!
>>> MMC:
>>>
>>> Breakpoint 1, getc () at
>>> /home/sjg/cosarm/src/third_party/u-boot/files/common/console.c:414
>>> 414 if (!gd->have_console)
>>> (gdb) up
>>> #1  0x00007ffff7914d48 in ?? () from /lib/x86_64-linux-gnu/libX11.so.6
>>> (gdb)
>>>
>>>
>>> So it seems that it is libX11 causing the problem.
>>>
>>> I don't think the best fix is to remove getc() from that library,
>>> although I do wonder if it is a bug. Renaming the symbol in U-Boot
>>> with #define (only on sandbox) might be one option.
>>>
>>> Regards,
>>> Simon
>>>
>>
>> #define getc _uboot_getc
>>
>> is what I tried but it runs you into the trouble that many other symbols
>> contain the getc substring.
>>
>> As we always try to use the same definition in U-Boot as in glibc we
>> should really replace getc() by another symbol, e.g. chget() so that we
>> avoid confusion.
>
> Well the first question is whether this is a bug with x11.
>
> I'm not sure that renaming is a good idea. getc() is a pretty standard name.
>
> Try
>
> #define getc(x) _sandbox_get(x)
>
> which might get you closer.
>
> Regards,
> Simon
>
https://gitlab.denx.de/u-boot/custodians/u-boot-efi/-/commit/7b302bedb21f94b9eb619cb38ef985f7bbb9be1f

seems to fix the crash. I get an output window.

But there are still problems:

* The shift key does not work on the keyboard.
* The output is black and white only.

What I would like to have is a 32bit color framebuffer on which I can
draw with the EFI graphical output protocol.

Best regards

Heinrich