Skip to content

Kernel memory leak detector 0.5

Catalin Marinas lkml 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest8 patches 7 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/8Base support for kmemleakinclude/linux/kernel.h include/linux/memleak.h init/main.c +3 more+1149 -30 replies
2/8Some documentation for kmemleakDocumentation/kmemleak.txt+92 -00 replies
3/8Add the memory allocation/freeing hooks for kmemleakdrivers/base/platform.c include/linux/slab.h mm/page_alloc.c +2 more+49 -40 replies
4/8Modules support for kmemleakkernel/module.c+32 -00 replies
5/8Add kmemleak support for i386arch/i386/kernel/vmlinux.lds.S include/asm-i386/processor.h+16 -00 replies
6/8Add kmemleak support for ARMarch/arm/kernel/vmlinux.lds.S include/asm-arm/processor.h+19 -00 replies
7/8Remove some of the kmemleak false positivesarch/i386/kernel/setup.c ipc/util.c kernel/params.c +1 more+9 -20 replies
8/8Simple testing for kmemleaklib/Kconfig.debug mm/Makefile mm/memleak-test.c+93 -00 replies

Reviews trailers per patch

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

Read the v1 thread

Cover letter [PATCH 2.6.17-rc5 0/8] Kernel memory leak detector 0.5

This is a new version (0.5) of the kernel memory leak detector. It is mainly a bug-fix release after testing it with a wider range of kernel modules. See the Documentation/kmemleak.txt file for a more detailed description. The patches are downloadable from (the bundled patch or the series):

http://homepage.ntlworld.com/cmarinas/kmemleak/patch-2.6.17-rc5-kmemleak-0.5.bz2
http://homepage.ntlworld.com/cmarinas/kmemleak/patches-kmemleak-0.5.tar.bz2

What's new in this version:

- fixed panic caused by some secondary offsets in modules
- removed the panic when not all aliases are found for freed pointers as the aliases list can be modified by modules insertion
- following Pavel's suggestion, shortened the function names by removing "_debug"

To do:

- more testing
- test Ingo's suggestion on task stacks scanning
- NUMA support
- (support for ioremap tracking)

Note that gcc4 cannot compile the kernels with this patch because of a bug with __builtin_constant_p (always returning true).

Any bug reports or suggestions are more than welcome as I can only test it with a limited number of kernel configurations.

Thanks.

Signature · 1 line
Catalin