On 11/30/23 10:55, Lukas Bulwahn wrote:
> Commit 1db9d06aaa55 ("mm/slab: remove CONFIG_SLAB from all Kconfig and
Note that's a -next (from slab/for-next) commit which might still change at
this point.
> Makefile") removes the config SLAB and makes the SLUB allocator the only
> default allocator in the kernel. Hence, the advice on reducing OS jitter
> due to kworker kernel threads to build with CONFIG_SLUB instead of
> CONFIG_SLAB is obsolete.
>
> Remove the obsolete advice to build with SLUB instead of SLAB.
>
> Signed-off-by: Lukas Bulwahn <lukas.bulwahn@xxxxxxxxx>
Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
But due to above, maybe it's best to wait after the removal goes through.
> ---
> .../admin-guide/kernel-per-CPU-kthreads.rst | 16 ++++++----------
> 1 file changed, 6 insertions(+), 10 deletions(-)
>
> diff --git a/Documentation/admin-guide/kernel-per-CPU-kthreads.rst b/Documentation/admin-guide/kernel-per-CPU-kthreads.rst
> index 993c2a05f5ee..b6aeae3327ce 100644
> --- a/Documentation/admin-guide/kernel-per-CPU-kthreads.rst
> +++ b/Documentation/admin-guide/kernel-per-CPU-kthreads.rst
> @@ -243,13 +243,9 @@ To reduce its OS jitter, do any of the following:
> 3. Do any of the following needed to avoid jitter that your
> application cannot tolerate:
>
> - a. Build your kernel with CONFIG_SLUB=y rather than
> - CONFIG_SLAB=y, thus avoiding the slab allocator's periodic
> - use of each CPU's workqueues to run its cache_reap()
> - function.
> - b. Avoid using oprofile, thus avoiding OS jitter from
> + a. Avoid using oprofile, thus avoiding OS jitter from
> wq_sync_buffer().
> - c. Limit your CPU frequency so that a CPU-frequency
> + b. Limit your CPU frequency so that a CPU-frequency
> governor is not required, possibly enlisting the aid of
> special heatsinks or other cooling technologies. If done
> correctly, and if you CPU architecture permits, you should
> @@ -259,7 +255,7 @@ To reduce its OS jitter, do any of the following:
>
> WARNING: Please check your CPU specifications to
> make sure that this is safe on your particular system.
> - d. As of v3.18, Christoph Lameter's on-demand vmstat workers
> + c. As of v3.18, Christoph Lameter's on-demand vmstat workers
> commit prevents OS jitter due to vmstat_update() on
> CONFIG_SMP=y systems. Before v3.18, is not possible
> to entirely get rid of the OS jitter, but you can
> @@ -274,7 +270,7 @@ To reduce its OS jitter, do any of the following:
> (based on an earlier one from Gilad Ben-Yossef) that
> reduces or even eliminates vmstat overhead for some
> workloads at https://lore.kernel.org/r/00000140e9dfd6bd-40db3d4f-c1be-434f-8132-7820f81bb586-000000@xxxxxxxxxxxxxxxxxxx.
> - e. If running on high-end powerpc servers, build with
> + d. If running on high-end powerpc servers, build with
> CONFIG_PPC_RTAS_DAEMON=n. This prevents the RTAS
> daemon from running on each CPU every second or so.
> (This will require editing Kconfig files and will defeat
> @@ -282,12 +278,12 @@ To reduce its OS jitter, do any of the following:
> due to the rtas_event_scan() function.
> WARNING: Please check your CPU specifications to
> make sure that this is safe on your particular system.
> - f. If running on Cell Processor, build your kernel with
> + e. If running on Cell Processor, build your kernel with
> CBE_CPUFREQ_SPU_GOVERNOR=n to avoid OS jitter from
> spu_gov_work().
> WARNING: Please check your CPU specifications to
> make sure that this is safe on your particular system.
> - g. If running on PowerMAC, build your kernel with
> + f. If running on PowerMAC, build your kernel with
> CONFIG_PMAC_RACKMETER=n to disable the CPU-meter,
> avoiding OS jitter from rackmeter_do_timer().
>
