Skip to content

Remove assert()

Tom Rini u-boot 1 version New

Get series with b4
b4 am 20171129023047.GL3587@bill-the-cat

Versions newest first

  1. v1 Latest0 of 5 patches 0 repliesNew

Patches in v1 No applies recorded yet

None of the 5 patches the cover letter announces are in this archive.

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [U-Boot] [PATCH 0/5] Remove assert()
Quoted 41 lines · Masahiro Yamada

2017-11-29 10:42 GMT+09:00 Tom Rini <[email protected]>:

Quoted 40 lines · Masahiro Yamada
Quoted 17 lines

2017-11-28 23:52 GMT+09:00 Tom Rini <[email protected]>:

Quoted 12 lines · Masahiro Yamada
Quoted 10 lines

Buildman test passed.

Masahiro Yamada (5):
Move CONFIG_PANIC_HANG to Kconfig
ARM: openrd: set CONFIG_LOGLEVEL to 2
Enable CONFIG_PANIC_HANG for boards without do_reset()
treewide: convert assert() to BUG_ON()
Remove assert()

Quoted 3 lines

Have you done a size check on this series? If not, I'll fire one up, thanks!

Quoted 6 lines

I think I did it correctly.

At least, I needed the following patch for openrd boards:
http://patchwork.ozlabs.org/patch/842106/

Looks like the NAND core pull-request needs it, too.

So, in my testing (which is gcc-6 for everything thanks to Debian/9):
06: Remove assert()
arm: + clearfog peach-pi snow smdk5250 smdk5420 turris_omnia spring omap3_evm peach-pit

(quoted diff) +5 -1
-(am335x_hs_evm) arm-linux-gnueabihf-ld.bfd: region `.sram' overflowed by 760 bytes
+(clearfog,turris_omnia) arm-linux-gnueabihf-ld.bfd: SPL image too big
+(am335x_hs_evm) arm-linux-gnueabihf-ld.bfd: region `.sram' overflowed by 1080 bytes
+(snow,smdk5250,peach-pi,smdk5420,spring,peach-pit) arch/arm/mach-exynos/built-in.o: In function `clock_calc_best_scalar':
+(snow,smdk5250,peach-pi,smdk5420,spring,peach-pit) build/../arch/arm/mach-exynos/clock.c:1408: undefined reference to `panic'
+(omap3_evm) arm-linux-gnueabihf-ld.bfd: region `.sram' overflowed by 28 bytes

And then a lot of platforms grow in size. Taking a harder peek, BUG_ON is always set. However, assert() is only set when DEBUG is set. Now, I can certainly see an argument along the lines of that being fairly unexpected, and I don't know that I would disagree.

Quoted 7 lines
Quoted 5 lines

After a very quick look over what git grep -l says, perhaps we want to introduce a new (default off) option to preserve the current behavior of assert() in BUG_ON, as it looks like almost only kernel imported code is using BUG_ON.

Do you have an idea for the option name?

ENABLE_BUG_ON_CHECKS ?

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20171128/fc4e4803/attachment.sig>

Read the cover letter in its thread