Paths

Table of Contentst

hwpstate_intel(4): use CPU_FOREACH instead of an IPI
AcceptedPublic
Actions

Authored by rlibby on Sun, Jul 19, 3:44 AM.

Details

Reviewers

kib
olce
aokblast

Summary

hwpstate_intel(4): use CPU_FOREACH instead of an IPI

Test Plan

make buildkernel for amd64 and i386.

I do not have an appropriate CPU to test the actual functionality.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Skipped

Unit

Tests Skipped

Build Status

Buildable 74940
Build 71823: arc lint + arc unit

Event Timeline

rlibby created this revision.Sun, Jul 19, 3:44 AM

Herald added a subscriber: imp. · View Herald TranscriptSun, Jul 19, 3:44 AM

rlibby requested review of this revision.Sun, Jul 19, 3:44 AM

Harbormaster completed remote builds in B74924: Diff 182199.Sun, Jul 19, 3:44 AM

rlibby added a parent revision: D58335: i386: provide PCPU pc_small_core for amd64 compat.Sun, Jul 19, 3:45 AM

rlibby edited the test plan for this revision. (Show Details)Sun, Jul 19, 3:51 AM

rlibby added reviewers: kib, olce, aokblast.

kib added inline comments.Sun, Jul 19, 2:47 PM

sys/x86/cpufreq/hwpstate_intel.c
379–380	You could just do CPU_FOREACH() as well. I think it would be nicer without introducing a global useful only for the specific driver. What is the problem? Missing PCPU small_cores on i386? Why not add the member there, formally, without initializing?

rlibby added inline comments.Sun, Jul 19, 3:09 PM

sys/x86/cpufreq/hwpstate_intel.c
379–380	You could just do CPU_FOREACH() as well. I think it would be nicer without introducing a global useful only for the specific driver. What is the problem? Missing PCPU small_cores on i386? Why not add the member there, formally, without initializing? Yeah, no pc_small_core on i386. I'm fine with that as a fix. I didn't do that initially because I thought pc_small_cores would be nonsense on i386 and memorizing the presence of small cores seemed to address two problems, but I do like that that approach will only affect i386 sources. I'll update the diffs. And yes, ack on the CPU_FOREACH. @olce made a similar comment on D57377.

kib feedback and earlier olce comment: CPU_FOREACH

Harbormaster completed remote builds in B74931: Diff 182224.Sun, Jul 19, 3:37 PM

rlibby retitled this revision from hwpstate_intel(4): use memorized hybrid_small_cores instead of an IPI to hwpstate_intel(4): use CPU_FOREACH instead of an IPI.Sun, Jul 19, 3:40 PM

rlibby edited the summary of this revision. (Show Details)

kib accepted this revision.Sun, Jul 19, 3:40 PM

This revision is now accepted and ready to land.Sun, Jul 19, 3:40 PM

aokblast accepted this revision.Sun, Jul 19, 3:50 PM

Refresh after conflict resolution for 02c440e20404 ("hwpstate_intel: Minimize ifdef for i386 build")

This revision now requires review to proceed.Sun, Jul 19, 8:10 PM

Harbormaster completed remote builds in B74940: Diff 182257.Sun, Jul 19, 8:10 PM

kib accepted this revision.Sun, Jul 19, 8:35 PM

This revision is now accepted and ready to land.Sun, Jul 19, 8:35 PM

Fine, though please note I have a lot of rework of hwpstate_intel(4) incoming, in particular dropping this hybrid core detection code, so if you plan more/significant work in this area please liaise with me.

In D58336#1338226, @olce wrote:

Fine, though please note I have a lot of rework of hwpstate_intel(4) incoming, in particular dropping this hybrid core detection code, so if you plan more/significant work in this area please liaise with me.

Nothing planned. I was only trying to clean this up related to the build break, which is now solved. If you'll be cleaning this up soon anyway, I'll be happy to either push this now as-is or drop it entirely and leave it to you. LMK, or, if you don't have a preference, I'll commit tomorrow to clear it off my patch stack.

Revision Contents
Changeset List

Path

Size

sys/

x86/

cpufreq/

hwpstate_intel.c

24 lines

Diff 182257

View Options

sys/x86/cpufreq/hwpstate_intel.c

hwpstate_intel(4): use CPU_FOREACH instead of an IPIAcceptedPublicActions