JakeHillion/scx-upstream
1
0
Fork 0
mirror of https://github.com/sched-ext/scx.git synced 2024-11-24 20:00:22 +00:00
Code Issues Packages Projects Releases Wiki Activity
1,755 Commits 30 Branches 18 Tags 20 MiB
cd7846f4d2
Commit Graph

206 Commits

Author SHA1 Message Date
likewhatevs
99d1179866
enable ide's etc. to work on bpf.c files (#668) 
* enable ide's etc. to work on the bpf.c files
this makes it so that clangd and ide tools which use clangd
can work on the bpf.c code.

nothing should actually be changed outside of that ide/editor
environment, all the changes are ifdef'ed on LSP which is set
in the added .clangd file.

* move intf include out of both sides of ifdef toggle
 2024-09-24 16:55:02 -04:00
Daniel Hodges
2805bb77a5
Merge pull request #683 from hodgesds/layered-idle-topo 
scx_layered: Make layered idle CPU selection topology aware
 2024-09-24 16:37:23 -04:00
Daniel Hodges
679dd5920c scx_layered: Fix comment 
Fix comment to be more accurate.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-24 12:31:03 -07:00
Daniel Hodges
87c6e276d9 scx_layered: Restrict idle selection to layer cpus 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-24 09:40:38 -07:00
Daniel Hodges
e68fccd26c scx_layered: Update comments on layer preemption 
Update comments on layer preemption to be more descriptive.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-24 08:24:04 -07:00
Daniel Hodges
6b966cda0c scx_layered: Restrict preemption to layer cpumask 
When preempting restrict preemption to the current layer cpumask. This
may reduce the amount of preemption, but cause better cache locality
of preempted tasks.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-24 06:23:01 -07:00
I Hsin Cheng
61cb3f7fc5 scx_common_bpf: Append cast_mask() 
Remove cast_mask() function distributed throughout different schedulers
and add it in common.bpf.h so every scheduler can reference it once they
need to.

Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
 2024-09-24 16:01:19 +08:00
Daniel Hodges
29fb647c93 scx_layered: Refactor idle core selection 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-23 12:01:42 -07:00
Daniel Hodges
380fd1f3b3 scx_layered: Make idle select topology aware 
Make idle CPU selection topology aware.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-23 10:10:43 -07:00
Daniel Hodges
35477970bd scx_layered: Cleanup dump format 
Cleanup the dump format for topology aware dumps in scx_layered.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-23 10:02:49 -07:00
Daniel Hodges
8b14e48994
Merge pull request #671 from hodgesds/layered-last-waker 
scx_layered: Add waker stats per layer
 2024-09-23 10:58:54 -04:00
Daniel Hodges
91d32663bd
scx_layered: Refactor waker tracking to only use last waker 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 18:05:54 -04:00
Daniel Hodges
1a2f82b91c
Merge pull request #666 from hodgesds/layered-local-llc 
scx_layered: Add topology aware preemption
 2024-09-22 17:36:32 -04:00
Daniel Hodges
326f3b7988
Merge pull request #667 from hodgesds/layered-pcore-grow 
scx_layered: Add Big/Little core growth algos
 2024-09-22 16:59:42 -04:00
Daniel Hodges
1ac9712d2e
scx_layered: Refactor preemption into a separate function 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 16:54:11 -04:00
Daniel Hodges
bc34bd867b
scx_layered: Add option to enable XNUMA preemption 
Disable XNUMA preemption by default and add an option to enable it.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 16:52:57 -04:00
Daniel Hodges
e105d9f8b1
scx_layered: Use cast_mask helper 
Use the cast_mask helper to clean up some of the bpf cpumask conversion
code for preemption.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 16:52:57 -04:00
Daniel Hodges
5d9d32b65c
scx_layered: Add stats for XLLC/XNUMA preemptions 
Add stats for XLLC/XNUMA preemptions.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 16:52:57 -04:00
Daniel Hodges
c15ecbb3a4
scx_layered: Add topology aware preemption 
Add topology aware preemption that begins in the local LLC and attempts
to preempt from cpus nearest in the topology.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 16:52:56 -04:00
Daniel Hodges
6fb2f0b2b4
scx_layered: Clean up waker code 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 06:43:10 -04:00
Daniel Hodges
c55b2c6e69
scx_layered: Add waker stats per layered 
Update the task context to keep a mask of wakers and add stats for wakes
across layers.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-22 06:43:03 -04:00
Daniel Hodges
140a101874
Merge pull request #449 from hodgesds/layered-dsq-fixes 
scx_layered: Add a hi fallback dsq per llc
 2024-09-22 06:39:46 -04:00
Daniel Hodges
4aa841de0a
scx_layered: Rename HI_FALLBACK_DSQ to HI_FALLBACK_DSQ_BASE 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-20 17:28:38 -04:00
Daniel Hodges
a3d1344293
scx_layered: Add core growth algo for core type 
Add core growth algos for Big/Little core support. The algos allow
layers to grow layers by preferring either big or little cores first.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-20 11:50:15 -04:00
I Hsin Cheng
7799b94f07 scx_layered: Add helper function to access cpumask within bpf_cpumask 
Before passing "nodec->cpumas" and "cachec->cpumask" into
"bpf_cpumask_test_cpu()", type conversion should be done first.
Implement "cast_mask()" to convert "struct bpf_cpumask *" into "const
struct cpumask *".

Reference from
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/testing/selftests/bpf/progs/cpumask_common.h#n63

Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
 2024-09-20 20:52:03 +08:00
Daniel Hodges
8532ba3f1e
scx_layered: Fix hi fallback dsq consumption 
Fix hi fallback dsq consumption to only consume from the cache local hi
fallback dsq.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-20 04:18:05 -04:00
I Hsin Cheng
e4bb99efc5 scx_layered: Refactor match_layer() 
Refactor match_layer() to prevent the compiling error caused by
uninitialization of the variable "nr_match_ors" before usage.

Move the checking of "nr_match_ors" after it access the value within
"layer->nr_match_ors" to make sure it's initiailized successfully.

Signed-off-by: I Hsin Cheng <richard120310@gmail.com>
 2024-09-19 22:20:03 +08:00
Jake Hillion
3848d87895 scx_layered: add round robin growth strategy 2024-09-12 23:27:21 +01:00
Daniel Hodges
632fcfe4ae
Merge pull request #648 from hodgesds/layered-llc-stats 
scx_layered: Add stats for XNUMA/XLLC migrations
 2024-09-12 13:23:23 -04:00
Daniel Hodges
dde6e0c7f9 scx_utils: Add node/llc id to core topology 
Add ids for node/llc in the Core topology struct.
 2024-09-12 10:05:02 -07:00
Daniel Hodges
aee19dd9a1 scx_layered: Add topology aware core growth selection 
Add topology aware core growth selection.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-12 06:48:51 -07:00
Daniel Hodges
14a19dc3ca scx_layered: Add random layer growth algo 
Add a random layer growth algo.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-12 05:35:54 -07:00
Daniel Hodges
ae57f8d1f9 scx_rusty: Initialize node cpumask 
Initialize the node cpumask, which was previously uninitialized causing
metric calculations to be wrong when attempting to lookup CPUs in the
node cpumask.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-11 13:14:44 -07:00
Jake Hillion
8ca45cfa37
lint: enable cargo fmt (#643) 
Use `cargo fmt` with a specific nightly branch in the CI to enforce formatting. Globally format these files while the diff is still small so we can stay on top of it.

Test plan:
- CI lint check passes.
 2024-09-11 10:03:20 +01:00
Daniel Hodges
43ec8bfe82 scx_layered: Add stats for XNUMA/XLLC migrations 
Add stats for XNUMA/XLLC migrations. An example of the output is shown:
```
  hodgesd  : util/frac=    5.4/  0.1 load/frac=    301.0/  0.3 tasks=   476
             tot=   3168 local=97.82 wake/exp/reenq= 2.18/ 0.00/ 0.00
             keep/max/busy= 0.03/ 0.00/ 0.03 kick= 0.00 yield/ign= 0.09/    0
             open_idle= 0.00 mig= 6.82 xnuma_mig= 6.82 xllc_mig= 4.86 affn_viol= 0.00
             preempt/first/idle/fail= 0.00/ 0.00/ 0.00/ 0.00 min_exec= 0.00/   0.00ms
             cpus=  2 [  2,  4] 00000000 00000010 00001000
  normal   : util/frac=   28.7/  0.7 load/frac= 101704.7/ 95.8 tasks=  2450
             tot=   4660 local=99.06 wake/exp/reenq= 0.88/ 0.06/ 0.00
             keep/max/busy= 1.03/ 0.00/ 0.00 kick= 0.06 yield/ign= 0.04/  400
             open_idle=15.73 mig=23.45 xnuma_mig=23.45 xllc_mig= 3.07 affn_viol= 0.00
             preempt/first/idle/fail= 0.00/ 0.00/ 0.00/ 0.88 min_exec= 0.00/   0.00ms
             cpus=  2 [  2,  2] 00000001 00000100 00000000
             excl_coll=12.55 excl_preempt= 0.00
  random   : util/frac=    0.0/  0.0 load/frac=      0.0/  0.0 tasks=     0
             tot=      0 local= 0.00 wake/exp/reenq= 0.00/ 0.00/ 0.00
             keep/max/busy= 0.00/ 0.00/ 0.00 kick= 0.00 yield/ign= 0.00/    0
             open_idle= 0.00 mig= 0.00 xnuma_mig= 0.00 xllc_mig= 0.00 affn_viol= 0.00
             preempt/first/idle/fail= 0.00/ 0.00/ 0.00/ 0.00 min_exec= 0.00/   0.00ms
             cpus=  0 [  0,  0] 00000000 00000000 00000000
             excl_coll= 0.00 excl_preempt= 0.00
  stress-ng: util/frac= 4189.1/ 99.2 load/frac=   4200.0/  4.0 tasks=    43
             tot=     62 local= 0.00 wake/exp/reenq= 0.00/100.0/ 0.00
             keep/max/busy=2433.9/177.4/ 0.00 kick=100.0 yield/ign= 3.23/    0
             open_idle= 0.00 mig=54.84 xnuma_mig=54.84 xllc_mig=35.48 affn_viol= 0.00
             preempt/first/idle/fail= 0.00/ 0.00/ 0.00/ 0.00 min_exec= 0.00/   0.00ms
             cpus=  4 [  4,  4] 00000300 00030000 00000000
             excl_coll= 0.00 excl_preempt= 0.00
```

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-10 19:53:28 -07:00
likewhatevs
85863d0e1c
Merge pull request #644 from hodgesds/layered-topo-order 
scx_layered: Pass layer spec for core growth algo
 2024-09-10 14:49:37 -04:00
Daniel Hodges
5fdd257862 scx_layered: Pass layer spec for core growth algo 
Pass in the layer spec when determining the layer core growth algo. This
should make it easier to implement layer growth algos that are spec
specific.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-10 10:27:08 -07:00
Samuel Nair
c6af1aa1c8 scx_layered: Fix typo in stats 2024-09-10 08:44:57 -07:00
Jake Hillion
2c008b2afa scx_layered: clean up Layer::new layer_growth_algo 2024-09-06 18:25:50 +01:00
Tejun Heo
46fc2e1a49 version: v1.0.4 2024-09-05 18:12:45 -10:00
Daniel Hodges
0fa369b914
Merge pull request #619 from hodgesds/stats-fixes 
scx_layered: Fix stats typo
 2024-09-05 15:44:15 -04:00
Daniel Hodges
25e1642bbc
scx_layered: Fix stats typo 
Small typo fix

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-05 14:12:28 -04:00
Daniel Hodges
76ad880475
scx_layered: Fix stats formatting 
Fix formatting precision of stats to have lower precision for
readability. The existing formatting is hard to read:

tot=   1538 local=31.27 open_idle= 2.73 affn_viol=23.80 proc=4ms
busy=  1.1 util=   16.6 load=     32.7 fallback_cpu=  6
excl_coll=0.06501950585175553 excl_preempt=0.26007802340702213 excl_idle=0.16384915474642392 excl_wakeup=0.25097529258777634

With this fix stats are far more readable formatting:

tot=    441 local=33.56 open_idle= 0.00 affn_viol=20.63 proc=3ms
busy=  0.4 util=    6.3 load=     33.6 fallback_cpu=  6
excl_coll=0.454 excl_preempt=0.000 excl_idle=0.132 excl_wakeup=0.200

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-09-05 06:44:54 -04:00
Tejun Heo
f010eda5c0 meson: Remove scheds/rust/*/meson.build 
These aren't used since 43950c65 ("build: Use workspace to group rust
sub-projects"). Drop them.
 2024-09-04 06:40:17 -10:00
Daniel Hodges
e04b612688 scx_layered: Fix layer timeslice not being applied 
Fix a small bug where the layer timeslice is not applied.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-08-30 11:53:42 -07:00
Daniel Hodges
47184e9d19
Merge pull request #582 from hodgesds/layered-growth-interface 
scx_layered: Add layer growth config
 2024-08-29 18:49:59 -04:00
Daniel Hodges
7e0329e45c scx_layered: Add layer growth config 
Add a per layer config for different implementations of layer growth
algorithms. Convert the existing default logic into a default layer
growth algorithm and add a linear implementation.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-08-28 19:17:24 -07:00
Daniel Hodges
cf765562c7
scx_layered: Update docs for layer slice setting 
Add docs for layer slice setting.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-08-28 22:12:07 -04:00
Daniel Hodges
a23308e7b0 scx_layered: Add more docs on tuning 
Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-08-28 12:38:05 -07:00
Daniel Hodges
96326b1ef3 scx_layered: Add additional docs 
Add some additional docs on tuning layered.

Signed-off-by: Daniel Hodges <hodges.daniel.scott@gmail.com>
 2024-08-28 12:27:26 -07:00