Re: 2.6.16-rc1-mm4

[Al Boldi <a1426z@gawab.com>]

Andrew Morton wrote:
> Al Boldi <a1426z@gawab.com> wrote:
> > Andrew Morton wrote:
> > > Al Boldi <a1426z@gawab.com> wrote:
> > > > Andrew Morton wrote:
> > > > > - Various other random bits and pieces.  Things have been pretty
> > > > > quiet lately - most activity seems to be concentrated about
> > > > > putting bugs into the various subsystem trees.
> > > >
> > > > Does it fix the DRM_i810 hang during a suspend-to-ram/disk cycle?
> > >
> > > I don't know - I don't watch every patch which goes into 49 different
> > > trees.  Did you try it?
> >
> > It still hangs w/ drm.
> >
> > w/o drm STD works like a charm.
>
> OK.
>
> > w/o drm STR works lest this t/o and a noisy display in X.
>
> What do "lest" and "t/o" mean?  Please use less abbreviations.

Sorry! lest=except ; t/o=timeout

>
> > Stopping tasks: ========================|
> > Suspending device 0.1
> > Suspending device 0.0
> > Suspending device ide0
> > Suspending device floppy.0
> > Suspending device serio1
> > Suspending device serio0
> > Suspending device i8042
> > Suspending device 0000:01:0a.0
> > Suspending device 0000:01:05.0
> > Suspending device 0000:00:1f.2
> > Suspending device 0000:00:1f.1
> > Suspending device 0000:00:1f.0
> > Suspending device 0000:00:1e.0
> > Suspending device 0000:00:01.0
> > Suspending device 0000:00:00.0
> > Suspending device pci0000:00
> > Suspending device platform
> > Intel machine check architecture supported.
> > Intel machine check reporting enabled on CPU#0.
> > Back to C!
> > ACPI: PCI Interrupt 0000:00:01.0[A] -> Link [LNKA] -> GSI 11 (level,
> > low) -> IRQ 11
> > PCI: Setting latency timer of device 0000:00:1e.0 to 64
> > eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
> > hda: drive not ready on wakeup
> > hda: status timeout: status=0xd0 { Busy }
> > ide: failed opcode was: unknown
> > hda: DMA disabled
> > hdb: DMA disabled
> > hda: drive not ready for command
> > ide0: reset: success
> > BUG: warning at drivers/ide/ide-iops.c:1235/ide_wait_not_busy()
> >  <c02612e2> ide_wait_not_busy+0xa2/0xb0   <c025ee87>
> > start_request+0x1a7/0x230
> >  <c025f17a> ide_do_request+0x23a/0x3c0   <c026440a>
> > set_multmode_intr+0x1a/0x70
> >  <c025f79b> ide_intr+0xeb/0x130   <c02643f0> set_multmode_intr+0x0/0x70
> >  <c0135610> handle_IRQ_event+0x30/0x70   <c01356a5> __do_IRQ+0x55/0xc0
> >  <c0105672> do_IRQ+0x42/0x70
> >  =======================
> >  <c0103976> common_interrupt+0x1a/0x20   <c021886d>
> > acpi_processor_idle+0x2b6/0x332
> >  <c0101030> default_idle+0x0/0x70   <c0101118> cpu_idle+0x58/0x70
> >  <c03c47bd> start_kernel+0x14d/0x170   <c03c4310>
> > unknown_bootoption+0x0/0x1e0
> > hdb: set_drive_speed_status: status=0x40 { DriveReady }
> > ide: failed opcode was: unknown
> > Restarting tasks... done
>
> The above is an IDE problem.  Are you saying that the above only occurs
> when DRM is enabled?

It's dependent on STR only. STD works fine. DRM seems irrelevant to this 
problem.

>
> Please take more time and use more words when describing this problem,
> thanks.

Please bare with me, as I am not really verse in English.

>
> > Also, a 100HZ recompile in mainline causes a 400% delay doing a simple
> > lilo, which seems fixed in mm.  Is there a reason you can't move that to
> > mainline?
>
> Again, I don't understand what that means.  Are you saying that a kernel
> compiled with HZ=100 takes 4 times as long to run lilo?

Yes, in mainline a kernel compiled with HZ=100 takes 4 times as long to run 
lilo as it would take for a kernel compiled with HZ=1000.  This problem does 
not exist in mm.

Also, 

	hdparm -tT /dev/hda
	cat /dev/hda > /dev/null
	dd < /dev/hda > /dev/null

get ~30% improved throughput when setting

	echo 192 > /sys/block/hda/queue/max_sectors_kb
	echo 192 > /sys/block/hda/queue/read_ahead_kb

whereas anything outside 132-255 affects throughput negatively.
Is this a bug?

Thanks!

--
Al