Skip to contentSkip to conversation

sandbox: './u-boot -l ' fails

BUG Heinrich Schuchardt 3 participants 14 messages Sep 19, 2020 – Sep 29, 2020

lore
Share & tools
lore
#0 lore raw

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

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.

Best regards

Heinrich

#1 lore raw

Hi Heinrich,

Quoted 7 lines · Heinrich Schuchardt

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.

Quoted 34 lines

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

Quoted 5 lines

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.

Regards,
Simon

#2 lore raw

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 57 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 10 lines

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.

Quoted 8 lines
Quoted 1 line

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?

Quoted 3 lines

Regards,
Simon

#3 lore raw
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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.

Best regards

Heinrich

#4 lore raw

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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

#5 lore raw
Quoted 129 lines · Simon Glass

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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)

Quoted 9 lines

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.

Best regards

Heinrich

#6 lore raw

Hi Heinrich,

Quoted 141 lines · Heinrich Schuchardt
Quoted 129 lines · Simon Glass

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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)

Quoted 9 lines

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

Quoted 8 lines

#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

#7 lore raw
Quoted 157 lines · Simon Glass

Hi Heinrich,

Quoted 141 lines · Heinrich Schuchardt
Quoted 129 lines · Simon Glass

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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)

Quoted 9 lines

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

Quoted 8 lines

#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.

Quoted 9 lines

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

#8 lore raw

Hi Heinrich,

Quoted 168 lines · Heinrich Schuchardt
Quoted 157 lines · Simon Glass

Hi Heinrich,

Quoted 141 lines · Heinrich Schuchardt
Quoted 129 lines · Simon Glass

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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)

Quoted 9 lines

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

Quoted 8 lines

#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.

Quoted 9 lines

Try

#define getc(x) _sandbox_get(x)

which might get you closer.

Regards,
Simon

Quoted 7 lines

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.

What is shift supposed to do? Do you mean you cannot get capital letters or symbols?

Quoted 1 line

* The output is black and white only.

That's odd. I haven't tried it recently but earlier in the year I was able to write a colour BMP.

Quoted 3 lines

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

Well that should work. I adds Nuklear support a while back and everything seemed good. Perhaps there has been a regression? But there is a test for colour bitmaps. Perhaps it is just the SDL display side which is not working?

I just tried this and got a colour image:

=> host load hostfs - 0 tools/logos/denx.bmp
=> bmp display 0

Regards,
Simon

#9 lore raw
Quoted 196 lines · Simon Glass

Hi Heinrich,

Quoted 168 lines · Heinrich Schuchardt
Quoted 157 lines · Simon Glass

Hi Heinrich,

Quoted 141 lines · Heinrich Schuchardt
Quoted 129 lines · Simon Glass

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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)

Quoted 9 lines

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

Quoted 8 lines

#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.

Quoted 9 lines

Try

#define getc(x) _sandbox_get(x)

which might get you closer.

Regards,
Simon

Quoted 7 lines

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.

What is shift supposed to do? Do you mean you cannot get capital letters or symbols?

Quoted 1 line

* The output is black and white only.

That's odd. I haven't tried it recently but earlier in the year I was able to write a colour BMP.

Quoted 3 lines

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

Well that should work. I adds Nuklear support a while back and everything seemed good. Perhaps there has been a regression? But there is a test for colour bitmaps. Perhaps it is just the SDL display side which is not working?

Quoted 8 lines

I just tried this and got a colour image:

=> host load hostfs - 0 tools/logos/denx.bmp
=> bmp display 0

Regards,
Simon

Hello Simon,

the background of characters is not set correctly. I thought it was black and white.

To see what I mean:

CONFIG_EFI_SELFTEST=y

./u-boot -D -l

=> setenv efi_selftest block image transfer
=> bootefi selftest

You will need my two patches.

Things that are still wrong:

* We start with black on white instead white on black.
* Background for characters.
* The font in not mono-spaced.
* When typing a backslash you get always two of them.

Best regards

Heinrich

#10 lore raw
Quoted 220 lines · Heinrich Schuchardt
Quoted 196 lines · Simon Glass

Hi Heinrich,

Quoted 168 lines · Heinrich Schuchardt
Quoted 157 lines · Simon Glass

Hi Heinrich,

Quoted 141 lines · Heinrich Schuchardt
Quoted 129 lines · Simon Glass

Hi Heinrich,

Quoted 72 lines · Heinrich Schuchardt
Quoted 60 lines · Heinrich Schuchardt

Am 28. September 2020 06:24:38 MESZ schrieb Simon Glass <[email protected]>:

Quoted 59 lines

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

Hello Simon,

when I try to run ./u-boot -l the sandbox stalls. Shouldn't it run

out

Quoted 3 lines

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.

Quoted 2 lines

U-Boot 2020.10-rc4-00018-g21a10244f9-dirty (Sep 19 2020 - 19:55:39

+0200)

Quoted 18 lines
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:

Quoted 7 lines

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

Quoted 1 line

__serial_getc(). I wonder how it gets there without returning from

the

Quoted 11 lines

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.

Quoted 9 lines
Quoted 6 lines

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?

Quoted 7 lines

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)

Quoted 9 lines

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

Quoted 8 lines

#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.

Quoted 9 lines

Try

#define getc(x) _sandbox_get(x)

which might get you closer.

Regards,
Simon

Quoted 7 lines

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.

What is shift supposed to do? Do you mean you cannot get capital letters or symbols?

Quoted 1 line

* The output is black and white only.

That's odd. I haven't tried it recently but earlier in the year I was able to write a colour BMP.

Quoted 3 lines

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

Well that should work. I adds Nuklear support a while back and everything seemed good. Perhaps there has been a regression? But there is a test for colour bitmaps. Perhaps it is just the SDL display side which is not working?

Quoted 8 lines

I just tried this and got a colour image:

=> host load hostfs - 0 tools/logos/denx.bmp
=> bmp display 0

Regards,
Simon

Hello Simon,

the background of characters is not set correctly. I thought it was black and white.

To see what I mean:

CONFIG_EFI_SELFTEST=y

./u-boot -D -l

=> setenv efi_selftest block image transfer
=> bootefi selftest

Quoted 8 lines

You will need my two patches.

Things that are still wrong:

* We start with black on white instead white on black.
* Background for characters.
* The font in not mono-spaced.
* When typing a backslash you get always two of them.

* White on black is customizing.
Can we change that for the sandbox_defconfig to fit to Linux standards?

* The truetype console does not yet understand the escape sequences for colors. The normal console does.

* The normal console has a mono-space font. We also have alternative fonts for the truetype console.

* Both consoles fail on Unicode characters. But we probably do not want larger font files.

So the only problem I do not understood yet is the duplicate backspace.

Best regards

Heinrich

#11 lore raw
In reply to Heinrich Schuchardt

Dear Heinrich,

Quoted 4 lines · In message <[email protected]> you wrote:

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.

We should not invent any new names when there are existing good alterntives.

Why not use getchar()?

Quote fgetc(3) man page:

...

       int getchar(void);

...
getchar() is equivalent to getc(stdin).
...

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
I saw Susie sitting in a shoe shine shop. Where she sits she shines,
and where she shines she sits.
#12 lore raw

Am 29. September 2020 13:12:25 MESZ schrieb Wolfgang Denk <[email protected]>:

Quoted 17 lines

Dear Heinrich,

Quoted 3 lines · In message <[email protected]> you wrote:

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

Quoted 1 line

avoid confusion.

We should not invent any new names when there are existing good alterntives.

Why not use getchar()?

Quote fgetc(3) man page:

...

Quoted 8 lines
      int getchar(void);

...
getchar() is equivalent to getc(stdin).
...

Best regards,

Wolfgang Denk

Hello Wolfgang,

please, have a look at http://patchwork.ozlabs.org/project/uboot/patch/20200928234514.124945-1-xypron.glpk at gmx.de/

Is such a #define the way to go or should we globally rename getc?

Best regards

Heinrich

#13 lore raw

Dear Heinrich,

Quoted 5 lines · In message <[email protected]> you wrote:
Quoted 5 lines

We should not invent any new names when there are existing good alterntives.

Why not use getchar()?

...

Quoted 3 lines

please, have a look at http://patchwork.ozlabs.org/project/uboot/patch> /20200928234514.124945-1-xypron.glpk at gmx.de/

Is such a #define the way to go or should we globally rename getc?

I don't think this is a good idea; I would recommend to keep the getc() as used so far in normal U-Boot, and only "translate" it into getchar() for the sandbox environment.

Especially messing with include/_exports.h is critical, as this is used for standalone applications, which are the only allowed exception from GPL code, so we have to assume there is some amount of binary-only code in the field, and we should be careful not to break this without need. Changes here may break compatibility.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"Well I don't see why I have to make one man  miserable  when  I  can
make so many men happy."              - Ellyn Mustard, about marriage