I've spent the last week or so getting ftrace et al. working
again on sparc64. The NMI scheme I use there was causing all
sorts of issues, but luckily the final fix was very simple.
While I was here debugging this stuff I added function graph
tracer support, as well as support for kmemleak. Kmemleak
even found a bug :-)
All combinations of features seems to work (perf top with tracers,
lockdep, and kmemleak all enabled for example) and wow is this stuff
I merged in your tree to resolve a conflict in lib/Kconfig.debug
since Microblaze added kmemleak support recently too.
Please pull, thanks a lot!
The following changes since commit 0d0fb0f9c5fddef4a10242fe3337f00f528a3099:
Linus Torvalds (1):
are available in the git repository at:
David S. Miller (14):
sparc64: Use kstack_valid() in die_if_kernel().
sparc64: Adjust __raw_local_irq_save() to cooperate in NMIs.
sparc64: Add HAVE_FUNCTION_TRACE_MCOUNT_TEST and tidy up.
sparc64: Kill CONFIG_STACK_DEBUG code.
sparc64: Kill unnecessary static on local var in ftrace_call_replace().
sparc64: Remove profiling from some low-level bits.
sparc64: Use a seperate counter for timer interrupts and NMI checks, like x86.
sparc64: Give a stack frame to the ftrace call sites.
sparc64: Add function graph tracer support.
sparc64: Support kmemleak.
sparc64: Add kmemleak annotation to sun4v_build_virq()
sparc64: Fix memory leak in pci_register_iommu_region().
Merge branch 'master' of /home/davem/src/GIT/linux-2.6/
sparc64: Add some more commentary to __raw_local_irq_save()
arch/sparc/Kconfig | 3 +
arch/sparc/Kconfig.debug | 5 +-
arch/sparc/include/asm/cpudata_64.h | 2 +-
arch/sparc/include/asm/irqflags_64.h | 23 ++++-
arch/sparc/kernel/Makefile | 10 ++-