Bug 60710
| Summary: | Radeon RV530 I2C failure with DVI monitor. | ||
|---|---|---|---|
| Product: | Drivers | Reporter: | Chris Rankin (rankincj) |
| Component: | Video(DRI - non Intel) | Assignee: | drivers_video-dri |
| Status: | NEW --- | ||
| Severity: | normal | CC: | alexdeucher, szg00000 |
| Priority: | P1 | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Kernel Version: | 3.10.5 | Subsystem: | |
| Regression: | No | Bisected commit-id: | |
| Attachments: |
dmesg log for 3.10.5 (WORKING)
Relevant slice of /var/log/messages |
||
|
Description
Chris Rankin
2013-08-06 22:10:05 UTC
Created attachment 107126 [details]
dmesg log for 3.10.5 (WORKING)
This doesn't demonstrate the problem. It's merely to describe the T60p hardware better.
Created attachment 107128 [details]
Relevant slice of /var/log/messages
I've trawled through my remaining logs, and after the initial error they really don't get any better than:
Aug 6 15:46:43 lantern kernel: [drm:r500_hw_i2c_xfer] *ERROR* failed to get i2c bus
I was originally trying to get my HD4650 card to drive my HDTV, while leaving my 1920x1200 monitor and LVDS panel driven by the internal M66GL chip. However, I'm not sure that I succeeded. More likely, I accidentally pushed XRandR into a configuration that the HW I2C was unable to handle gracefully instead.
hw i2c is diabled by default for a reason. The error message means the driver was not able to aquire the hw lock for the i2c engine. As to why, I'm not sure. The hw i2c engine somehow got into a bad state. (In reply to Alex Deucher from comment #3) > The hw i2c engine somehow got into a bad state. Not only that, but GNOME3 put the HW I2C engine back into that same "bad state" when I rebooted and logged back in. (The DVI-0 monitor was switched to a non-native 1680x1050 mode during the whole reboot process, but was then shut down and disconnected when GNOME3 was unable to reconfigure it for its native 1920x1200 mode). > hw i2c is diabled by default for a reason. Yes, presumably because it has bugs like this in ;-). But surely that doesn't mean you're not interested in people reporting those bugs when they find them? (In reply to Chris Rankin from comment #4) > (In reply to Alex Deucher from comment #3) > > The hw i2c engine somehow got into a bad state. > > Not only that, but GNOME3 put the HW I2C engine back into that same "bad > state" when I rebooted and logged back in. (The DVI-0 monitor was switched > to a non-native 1680x1050 mode during the whole reboot process, but was then > shut down and disconnected when GNOME3 was unable to reconfigure it for its > native 1920x1200 mode). Well a warm reboot may not have fuly reset the hardware; the bad state may have persisted. > > > hw i2c is diabled by default for a reason. > > Yes, presumably because it has bugs like this in ;-). But surely that > doesn't mean you're not interested in people reporting those bugs when they > find them? Sure, I just wouldn't hold out much hope for a fix :) r5xx hardware is pretty old. (In reply to Alex Deucher from comment #5) > Well a warm reboot may not have fuly reset the hardware; the bad state may > have persisted. What "warm reboot"? IIRC, I powered the laptop down and came back to it ~6 hours later (although admittedly, I didn't remote the battery). The DVI-0 monitor came back fine afterwards, until I actually logged in and (presumably) GNOME3 tried to apply my account's XRandR settings. |