Bug 246341
| Summary: | newhidups wakes up 1000 times per second | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Fabrice Bellet <fabrice> |
| Component: | nut | Assignee: | Tomas Smetana <tsmetana> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | low | Docs Contact: | |
| Priority: | low | ||
| Version: | 7 | CC: | fabrice |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | 2.2.1-1 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2008-01-25 11:01:10 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 204948 | ||
I'm not sure that newhidups (or usbhid-ups in nut 2.2.0) could be more resource-friendly... They keep polling the ups for changes. The polling frequency is configurable (-i parameter) but setting it to high means that any power-outage or other problem might be discovered too late and nut may not initiate the shutdown in time. I'll try to tweak the default frequency. rawhide version 2.2.1 lowered the timeout value from 5000 to 250, so usbhid-ups now only generates ~100 wakeups per seconds, which is much better :) This bug can probably be closed now. Will version 2.2.1 be pushed for Fedora 8 eventually ? I'll take a look at the new version and if it doesn't break API (I think I saw something in the new release I didn't like) I will push it into F-8. Closing the bug as per comment #2. |
Under powertop, newhidups generates a large amount of wakeups/second: Wakeups-from-idle per second : 2381.6 no ACPI power usage estimate available Top causes for wakeups: 65.9% (936.8) newhidups : schedule_timeout (process_timeout) [...] Here is a strace sample: 20:12:13 select(87, NULL, [], NULL, {0, 1000}) = 0 (Timeout) 20:12:13 gettimeofday({1183227133, 69667}, NULL) = 0 20:12:13 ioctl(86, USBDEVFS_REAPURBNDELAY, 0xbfaed4f8) = -1 EAGAIN (Resource temporarily unavailable) 20:12:13 select(87, NULL, [], NULL, {0, 1000}) = 0 (Timeout) 20:12:13 gettimeofday({1183227133, 70666}, NULL) = 0 20:12:13 ioctl(86, USBDEVFS_REAPURBNDELAY, 0xbfaed4f8) = -1 EAGAIN (Resource temporarily unavailable) 20:12:13 select(87, NULL, [], NULL, {0, 1000}) = 0 (Timeout) 20:12:13 gettimeofday({1183227133, 71692}, NULL) = 0 20:12:13 ioctl(86, USBDEVFS_REAPURBNDELAY, 0xbfaed4f8) = -1 EAGAIN (Resource temporarily unavailable) 20:12:13 select(87, NULL, [], NULL, {0, 1000}) = 0 (Timeout) 20:12:13 gettimeofday({1183227133, 72667}, NULL) = 0 20:12:13 ioctl(86, USBDEVFS_REAPURBNDELAY, 0xbfaed4f8) = -1 EAGAIN (Resource temporarily unavailable) 20:12:13 select(87, NULL, [], NULL, {0, 1000}) = 0 (Timeout) 20:12:13 gettimeofday({1183227133, 73745}, NULL) = 0 [...] newhidups probably could be more resource-friendly.