summaryrefslogtreecommitdiffstats
path: root/drivers/nvdimm
diff options
context:
space:
mode:
authorChen Huang <chenhuang5@huawei.com>2021-04-29 07:03:48 +0000
committerPalmer Dabbelt <palmerdabbelt@google.com>2021-05-22 18:59:20 -0700
commiteac2f3059e02382d91f8c887462083841d6ea2a3 (patch)
treeb2a5e92f2ee0853bd578038c348efaea5a825cb6 /drivers/nvdimm
parent6efb943b8616ec53a5e444193dccf1af9ad627b5 (diff)
downloadlinux-stable-eac2f3059e02382d91f8c887462083841d6ea2a3.tar.gz
linux-stable-eac2f3059e02382d91f8c887462083841d6ea2a3.tar.bz2
linux-stable-eac2f3059e02382d91f8c887462083841d6ea2a3.zip
riscv: stacktrace: fix the riscv stacktrace when CONFIG_FRAME_POINTER enabled
As [1] and [2] said, the arch_stack_walk should not to trace itself, or it will leave the trace unexpectedly when called. The example is when we do "cat /sys/kernel/debug/page_owner", all pages' stack is the same. arch_stack_walk+0x18/0x20 stack_trace_save+0x40/0x60 register_dummy_stack+0x24/0x5e init_page_owner+0x2e So we use __builtin_frame_address(1) as the first frame to be walked. And mark the arch_stack_walk() noinline. We found that pr_cont will affact pages' stack whose task state is RUNNING when testing "echo t > /proc/sysrq-trigger". So move the place of pr_cont and mark the function dump_backtrace() noinline. Also we move the case when task == NULL into else branch, and test for it in "echo c > /proc/sysrq-trigger". [1] https://lore.kernel.org/lkml/20210319184106.5688-1-mark.rutland@arm.com/ [2] https://lore.kernel.org/lkml/20210317142050.57712-1-chenjun102@huawei.com/ Signed-off-by: Chen Huang <chenhuang5@huawei.com> Fixes: 5d8544e2d007 ("RISC-V: Generic library routines and assembly") Cc: stable@vger.kernel.org Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
Diffstat (limited to 'drivers/nvdimm')
0 files changed, 0 insertions, 0 deletions