Re: ohci1394 problem (MMIO broken) (was 2.6.25-rc6-git6: Reported regressions from 2.6.24)

[Thomas Meyer <thomas@m3y3r.de>]

Stefan Richter schrieb:
> Rafael J. Wysocki wrote:
>> Bug-Entry    : http://bugzilla.kernel.org/show_bug.cgi?id=10080
>> Subject        : 2.6.25-rc2: ohci1394 problem (MMIO broken)
>> Submitter    : Thomas Meyer <thomas@m3y3r.de>
>> Date        : 2008-02-20 08:47 (31 days old)
>> References    : http://lkml.org/lkml/2008/2/20/58
>> Handled-By    : Stefan Richter <stefanr@s5r6.in-berlin.de>
>
> This bug is alas an orphan.  It is /not/ handled by me, because it is 
> not an IEEE 1394 subsystem bug and I have no idea what to do about it.
>
> Note the following:
>
>   - Several or all of ohci1394's MMIO reads return ~0 (all bits set
>     to one) --- or 0 --- where different values are expected.
>
>   - In http://lkml.org/lkml/2008/2/23/244 we get to see a
>     WARNING: at arch/x86/mm/ioremap.c:137 __ioremap+0xa7/0x16a()
>     which is "WARN_ON_ONCE(page_is_ram(pfn));".
>     After that, the failures start.
>     But before that, "Unknown symbol" messages pop up when ohci1394
>     is loaded.  These symbols are implemented by ieee1394 on which
>     ohci1394 depends.
>
> More context from http://lkml.org/lkml/2008/2/23/244:
>
>> [  199.908169] ath_pci: wifi0: Atheros 5424/2424: mem=0x94300000, irq=17
>
> ath_hal taints the kernel.
I removed the ath driver, see bug 
http://bugzilla.kernel.org/show_bug.cgi?id=10080.
>
>> [  847.318678] ohci1394: Unknown symbol hpsb_iso_wake
>> [  847.318791] ohci1394: Unknown symbol hpsb_resume_host
> ...
>
> Some unexplained build problem.
I moved ohci1394.ko + ieee1394.ko file manually to avoid automatic 
loading. then i used insmod on those files and forgot to insmod ieee1394 
first. That's all. This is not related to the bug.
>
>> [  856.789954] ACPI: PCI Interrupt 0000:0c:03.0[A] -> GSI 19 (level, 
>> low) -> IRQ 19
>> [  856.790040] ------------[ cut here ]------------
>> [  856.790044] WARNING: at arch/x86/mm/ioremap.c:137 
>> __ioremap+0xa7/0x16a()
>> [  856.790048] Modules linked in: ohci1394(+) ieee1394 wlan_wep 
>> wlan_scan_sta ath_rate_sample ath_pci wlan ath_hal(P) firmware_class 
>> fuse snd_seq snd_seq_device nls_iso8859_15 nls_cp850 vfat fat usbhid 
>> appletouch applesmc input_polldev led_class dummy binfmt_misc tun 
>> pktcdvd loop msr cpuid coretemp hwmon eeprom cpufreq_powersave 
>> cpufreq_conservative acpi_cpufreq thermal ehci_hcd tpm_infineon 
>> i2c_i801 i2c_core tpm uhci_hcd usbcore tpm_bios processor ac battery 
>> sr_mod rng_core iTCO_wdt button firewire_ohci firewire_core sg 
>> snd_hda_intel snd_pcm snd_timer snd soundcore snd_page_alloc evdev 
>> intel_agp cdrom [last unloaded: microcode]
>> [  856.790119] Pid: 7140, comm: insmod Tainted: P         2.6.25-rc2 
>> #115
>> [  856.790125]  [<c011e5ab>] warn_on_slowpath+0x40/0x4f
>> [  856.790143]  [<c0118045>] __wake_up+0x29/0x39
>> [  856.790154]  [<c0277305>] netlink_broadcast+0x26e/0x2af
>> [  856.790169]  [<c01cc36b>] kobject_uevent_env+0x33d/0x361
>> [  856.790178]  [<c025ef2b>] pci_mmcfg_write+0xc4/0xd5
>> [  856.790187]  [<c0260157>] raw_pci_write+0x3e/0x46
>> [  856.790200]  [<c01148f8>] __ioremap+0xa7/0x16a
>> [  856.790210]  [<f8b14425>] ohci1394_pci_probe+0x20c/0x5a6 [ohci1394]
>> [  856.790226]  [<c01d63f1>] pci_device_probe+0x36/0x55
>> [  856.790236]  [<c02226cb>] driver_probe_device+0x9d/0x114
> ...
>> [  856.790598] ---[ end trace 5b0384c17c339107 ]---
>> [  856.943807] ohci1394: fw-host0: Get PHY Reg timeout 
>> [0x00008400/0x00000000/100]
>
> MMIO read expected 0x00008400, got 0x00000000.
>
> ...
>> [  921.826439] ohci1394: fw-host0: Unhandled interrupt(s) 0xfc7cfe0c
>
> MMIO read got random bits in the interrupt event register.
>
> ...
>> [  994.471724] ohci1394: fw-host0: OHCI-1394 0.35 (PCI): IRQ=[19]  
>> MMIO=[100000000-1000007ff]  Max Packet=[65536]  IR/IT contexts=[32/32]
>
> The values for Max Packet and IR/IT contexts came from bogus MMIO reads.
>
> Thomas, you wrote in http://lkml.org/lkml/2008/3/17/316 that the 
> problem resurfaced.
>   - Are the "Unknown symbol"s still there?  These are not supposed to
>     happen.
No. (See explanation above).
>   - Is the "WARNING: at arch/x86/mm/ioremap.c" still there?
No. I couldn't reproduce this warning, yet.

>   - Can you reproduce it without the atheros driver?
Yes.

See http://bugzilla.kernel.org/show_bug.cgi?id=10080.

>
> Rafael, no matter how it looks, this is not "handled-by: me". :-)
> I don't know nothing about kbuild nor about arch/x86/mm nor about WLAN.
> The IEEE 1394 bits in the bug of which I know something about are 
> purely accidental.
Seems so.