| Message ID | 20220908130936.674-1-thunder.leizhen@huawei.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | kallsyms: Optimizes the performance of lookup symbols | expand |
On Thu, Sep 08, 2022 at 09:09:29PM +0800, Zhen Lei wrote: > Currently, to search for a symbol, we need to expand the symbols in > 'kallsyms_names' one by one, and then use the expanded string for > comparison. This is very slow. > > In fact, we can first compress the name being looked up and then use > it for comparison when traversing 'kallsyms_names'. > > This patch series optimizes the performance of function kallsyms_lookup_name(), > and function klp_find_object_symbol() in the livepatch module. Based on the > test results, the performance overhead is reduced to 5%. That is, the > performance of these functions is improved by 20 times. > > To avoid increasing the kernel size in non-debug mode, the optimization is only > for the case CONFIG_KALLSYMS_ALL=y. WIthout having time yet to reveiw the implementation details, it would seem this is an area we may want to test for future improvements easily, so a selftest better yet a kunit test may be nice for this. Can you write one so we can easily gather a simple metric for "how long does this take"? Luis
On 2022/9/9 8:07, Luis Chamberlain wrote: > On Thu, Sep 08, 2022 at 09:09:29PM +0800, Zhen Lei wrote: >> Currently, to search for a symbol, we need to expand the symbols in >> 'kallsyms_names' one by one, and then use the expanded string for >> comparison. This is very slow. >> >> In fact, we can first compress the name being looked up and then use >> it for comparison when traversing 'kallsyms_names'. >> >> This patch series optimizes the performance of function kallsyms_lookup_name(), >> and function klp_find_object_symbol() in the livepatch module. Based on the >> test results, the performance overhead is reduced to 5%. That is, the >> performance of these functions is improved by 20 times. >> >> To avoid increasing the kernel size in non-debug mode, the optimization is only >> for the case CONFIG_KALLSYMS_ALL=y. > > WIthout having time yet to reveiw the implementation details, it would > seem this is an area we may want to test for future improvements easily, > so a selftest better yet a kunit test may be nice for this. Can you > write one so we can easily gather a simple metric for "how long does > this take"? Good advice. I'll write it today. > > Luis > . >
Currently, to search for a symbol, we need to expand the symbols in 'kallsyms_names' one by one, and then use the expanded string for comparison. This is very slow. In fact, we can first compress the name being looked up and then use it for comparison when traversing 'kallsyms_names'. This patch series optimizes the performance of function kallsyms_lookup_name(), and function klp_find_object_symbol() in the livepatch module. Based on the test results, the performance overhead is reduced to 5%. That is, the performance of these functions is improved by 20 times. To avoid increasing the kernel size in non-debug mode, the optimization is only for the case CONFIG_KALLSYMS_ALL=y. Zhen Lei (7): scripts/kallsyms: don't compress symbol type when CONFIG_KALLSYMS_ALL=y scripts/kallsyms: rename build_initial_tok_table() kallsyms: Adjust the types of some local variables kallsyms: Improve the performance of kallsyms_lookup_name() kallsyms: Add helper kallsyms_on_each_match_symbol() livepatch: Use kallsyms_on_each_match_symbol() to improve performance livepatch: Improve the search performance of module_kallsyms_on_each_symbol() include/linux/kallsyms.h | 8 +++ kernel/kallsyms.c | 135 +++++++++++++++++++++++++++++++++++++-- kernel/livepatch/core.c | 25 ++++++-- kernel/module/kallsyms.c | 13 +++- scripts/kallsyms.c | 19 ++++-- 5 files changed, 184 insertions(+), 16 deletions(-)