Re: pre-release 2.6.11-0.1 debian/powerpc kernel.
Sven Luther wrote:
[...]
> And altough it is not yet fully ready to be uploaded (mostly because
> kernel-source-2.6.11 needs cleaning up a bit yet), i would encourage you to
> try it out and give me feedback on it.
>
> I have been running it since a couple of days, and didn't see any particular
> problems.
In short words: sleep support is broken.
I have an ibook2.2:
| debby3:/home/frank# cat /proc/cpuinfo
| cpu : 750FX
| temperature : 11-13 C (uncalibrated)
| clock : 600MHz
| revision : 1.18 (pvr 7000 0112)
| bogomips : 1189.47
| machine : PowerBook4,3
| motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh
| board revision : 00000000
| detected as : 257 (iBook 2 rev. 2)
| pmac flags : 0000000b
| L2 cache : 512K unified
| memory : 384MB
| pmac-generation : NewWorld
The ibook enters sleep via the 'snooze' command.
The sleep LED pulses normal.
After touching the keybord the LED stops to pulse.
Sometimes the backlight is powered on, sometimes not.
Then the wake process stops.
There is some debug output on the screen in rare cases.
For instance:
| cpufreq: resume failed to assert current freqency is what timing core
| thinks it is
| Oops: exception in kernel mod, sig4 [#1]
|
| NIP C003F298 LR C003F2D0 SP C1CE9E99
| REGS: C1CE9DE0 TRAP 0700 Not tainted
[...]
| # less /boot/System.map-2.6.11-powerpc
[...]
| c003f094 t __drain_pages
| c003f148 T mark_free_pages
| c003f258 T drain_local_pages
| c003f294 t zone_statistics
| c003f298 t free_hot_cold_page
| c003f3dc T free_hot_page
| c003f3e4 T free_cold_page
| c003f3ec t buffered_rmqueue
| c003f6a0 T zone_watermark_ok
[...]
I have had the same behavior if I use self-compiled vanilla-2.6.11
kernels. I tested to wake up the ibook with and without usb support,
with and without cpufreq support and with and without connected usb
devices.
The backlight powers on in most cases, sometimes I get some lines on
the screen, e.g.:
| suspending type 'cpu':
| cpu0
| HID1, before 60000044
| radeonfb: (0000:00:10.0): resuming from state:2 ...
| radeonfb: switching to D0 state ...
| HID1, after 60000044
| Resuming System Devices
| Resuming type 'cpu":
| cpu0
| cpufreq: resume failed to assert current frequncy what timing core
| thinks it is
| Resuming typ 'openpic':
| openpic0
| _
Since xmon was compiled in, I got in this case some more lines:
| vector 400 at pc = 20202020, lr = c003b390
| msr 40001032, sp = c4201d30 [c4201c80]
| current = cc2877b0, pid = 1950, comm = pmud
| mon>
(The keyboard was not usable.)
| # less /boot/System.map-2.6.11
[...]
| c003b230 T no_action
| c003b238 T handle_IRQ_event
| c003b2e0 T __do_IRQ
| c003b440 T disable_irq_nosync
| c003b440 t gcc2_compiled.
| c003b4b4 T disable_irq
[...]
And, of course I have had no problems using self-compiled
vanilla-2.6.10 kernels.
Frank
Reply to: