config-linux: RFC 2119 wording for intelRdt #787
Conversation
| defined 'subset' of L3 cache which may be overlapping with other 'subsets'. | ||
| The different subsets are identified by class of service (CLOS) and each CLOS | ||
| has a capacity bitmask (CBM). | ||
| If `intelRdt` is not set, the runtime MUST NOT manipulate any `resctrl` psuedo-filesystems. |
There was a problem hiding this comment.
+ If `intelRdt` is not set, the runtime MUST NOT manipulate any `resctrl` psuedo-filesystems.
How about using these words? There should not be more than one resctrl filesystems.
+ If `intelRdt` is not set, the runtime MUST NOT manipulate the `resctrl` psuedo-filesystem.
There was a problem hiding this comment.
There should not be more than one resctrl filesystems.
But there might be more than one (just like proc). I expect you'd only actually have multiple resctrl filesystems mounted when you were setting up nested containers (e.g. docker-in-docker), but the “any” wording isn't much more complicated and avoids confusion in the multiple-mount corner case.
There was a problem hiding this comment.
@wking I got it. It makes sense in nested containers case.
config-linux.md
Outdated
|
|
||
| The following parameters can be specified for the container: | ||
| Consider a two-socket machine with two L3 caches where the default CBM is 0xfffff and the max CBM length is 20 bits. | ||
| This configuration assigns 4/5 of L3 cache id 0 and the whole L3 cache id 1 for the container: |
There was a problem hiding this comment.
Since the inline tutorial will be removed, I would like to improve the original wording to prevent confusion to new users of Intel RDT, and provide a more sophisticated example.
I will submit a new PR for it.
+Tasks inside the container only have access to the "upper" 80% of L3 cache id 0 and the "lower" 50% L3 cache id 1:
"linux": {
"intelRdt": {
"l3CacheSchema": "L3:0=ffff0;1=3ff"
}
}
There was a problem hiding this comment.
@wking Thanks. It looks good to me.
So we can compliance-test runtimes for these settings. Also remove the tutorial, since the kernel docs should provide sufficient documentation on that front. The kernel can be patched if they do not, and we do not include tutorials for other config-linux settings in this spec. The updated example was recommended by Xiaochen to compensate for the removed inline tutorial [1]. [1]: opencontainers#787 (comment) Signed-off-by: W. Trevor King <[email protected]>
|
LGTM |
So we can compliance-test runtimes for these settings. Also remove the tutorial, since the kernel docs should provide sufficient documentation on that front. The kernel can be patched if they do not, and we do not include tutorials for other config-linux settings in this spec. The updated example was recommended by Xiaochen to compensate for the removed inline tutorial [1]. [1]: opencontainers#787 (comment) Signed-off-by: W. Trevor King <[email protected]>
|
LGTM |
1 similar comment
|
LGTM |
So we can compliance-test runtimes for these settings.
Also remove the tutorial, since the kernel docs should provide sufficient documentation on that front. The kernel can be patched if they do not, and we do not include tutorials for other config-linux settings in this spec.
See also opencontainers/runc#433 and opencontainers/runc#1279.
CC @xiaochenshen.