oss_hdaudio broken with Linux 4.7.2 (Intel IOMMU conflicts with OSSv4) [semi-SOLVED]

[hash]

So "new" Kernel new problems:
oss_hdaudio cannot load normally, no devices in /dev/oss and this is what I see in dmesg:

Code: Select all

[ 5244.546779] DMAR: DRHD: handling fault status reg 2
[ 5244.546785] DMAR: [DMA Read] Request device [00:1b.0] fault addr 3c000 [fault reason 02] Present bit in context entry is clear
[ 5244.578779] oss_hdaudio: RIRB timeout (cad=0, nid=0, d=0, verb=f00, parm=0)
[ 5244.578792] oss_hdaudio: attach_codec: Codec #0 is not physically present
[ 5244.610962] oss_hdaudio: RIRB timeout (cad=0, nid=0, d=0, verb=f00, parm=0)
[ 5244.610970] oss_hdaudio: No hdaudio codecs were detected
[ 5244.611074] ------------[ cut here ]------------
[ 5244.611079] WARNING: CPU: 5 PID: 10971 at fs/proc/generic.c:575 remove_proc_entry+0x10d/0x150
[ 5244.611080] remove_proc_entry: removing non-empty directory 'irq/22', leaking at least 'oss_hdaudio0'
[ 5244.611081] Modules linked in: oss_hdaudio(O+) osscore(O) tun nvidia_drm(PO) nvidia_modeset(PO) nvidia(PO) x86_pkg_temp_thermal coretemp drm_kms_helper kvm_intel cfbfillrect syscopyarea cfbimgblt kvm sysfillrect sysimgblt $
[ 5244.611107] CPU: 5 PID: 10971 Comm: modprobe Tainted: P        W  O    4.7.2-gentoo #1
[ 5244.611110] Hardware name: System manufacturer System Product Name/P9X79-E WS, BIOS 1704 05/08/2015
[ 5244.611113]  0000000000000000 ffffffff813f0893 ffff88077446fab0 0000000000000000
[ 5244.611117]  ffffffff8106a88a ffff8800d000a0f8 ffff88077446fb00 ffff8807fc3d9c00
[ 5244.611121]  ffffffffa0052240 ffffffffa00524a0 ffff88076bd09780 ffffffff8106a8fa
[ 5244.611125] Call Trace:
[ 5244.611133]  [<ffffffff813f0893>] ? dump_stack+0x46/0x63
[ 5244.611139]  [<ffffffff8106a88a>] ? __warn+0xba/0xe0
[ 5244.611144]  [<ffffffff8106a8fa>] ? warn_slowpath_fmt+0x4a/0x50
[ 5244.611149]  [<ffffffff811916d5>] ? proc_entry_rundown+0x45/0x90
[ 5244.611152]  [<ffffffff811968bd>] ? remove_proc_entry+0x10d/0x150
[ 5244.611157]  [<ffffffff810ad217>] ? unregister_irq_proc+0xb7/0xc0
[ 5244.611161]  [<ffffffff810a72a6>] ? free_desc+0x26/0x70
[ 5244.611165]  [<ffffffff810a745d>] ? irq_free_descs+0x2d/0x70
[ 5244.611171]  [<ffffffff81035743>] ? mp_unmap_irq+0x53/0x60
[ 5244.611177]  [<ffffffff8102f335>] ? acpi_unregister_gsi_ioapic+0x25/0x40
[ 5244.611183]  [<ffffffff81472b54>] ? acpi_pci_irq_disable+0x6d/0x79
[ 5244.611187]  [<ffffffff8142748c>] ? do_pci_disable_device+0x4c/0x60
[ 5244.611191]  [<ffffffff814274d9>] ? pci_disable_device+0x39/0xa0
[ 5244.611198]  [<ffffffffa003916a>] ? osspci_probe+0xda/0x120 [oss_hdaudio]
[ 5244.611203]  [<ffffffff8142a140>] ? pci_device_probe+0x80/0xf0
[ 5244.611208]  [<ffffffff814c0aee>] ? driver_probe_device+0x17e/0x2b0
[ 5244.611212]  [<ffffffff814c0ca7>] ? __driver_attach+0x87/0x90
[ 5244.611216]  [<ffffffff814c0c20>] ? driver_probe_device+0x2b0/0x2b0
[ 5244.611220]  [<ffffffff814bee12>] ? bus_for_each_dev+0x52/0x80
[ 5244.611224]  [<ffffffff814c0183>] ? bus_add_driver+0x1a3/0x220
[ 5244.611227]  [<ffffffffa0068000>] ? 0xffffffffa0068000
[ 5244.611231]  [<ffffffff814c1246>] ? driver_register+0x56/0xd0
[ 5244.611234]  [<ffffffffa0068000>] ? 0xffffffffa0068000
[ 5244.611240]  [<ffffffffa006801a>] ? pcidrv_init+0x1a/0x2d [oss_hdaudio]
[ 5244.611244]  [<ffffffff810003d2>] ? do_one_initcall+0x32/0x140
[ 5244.611249]  [<ffffffff8113008e>] ? kmem_cache_alloc+0xce/0x160
[ 5244.611254]  [<ffffffff810ed726>] ? do_init_module+0x50/0x1b4
[ 5244.611258]  [<ffffffff810c9f98>] ? load_module+0x1bf8/0x1f80
[ 5244.611264]  [<ffffffff810c7670>] ? find_symbol_in_section+0xa0/0xa0
[ 5244.611269]  [<ffffffff81144756>] ? kernel_read+0x36/0x50
[ 5244.611273]  [<ffffffff810ca4a9>] ? SyS_finit_module+0x89/0x90
[ 5244.611279]  [<ffffffff817df59b>] ? entry_SYSCALL_64_fastpath+0x13/0x8f
[ 5244.611282] ---[ end trace 313e2129882d1192 ]---
[ 5244.611293] oss_hdaudio: probe of 0000:00:1b.0 failed with error -5

As far as I can see no other similar reports is here, so any way to fix this other then stay on 4.6 kernel?

[hash]

I tested oss_usb: no errors and empty /dev/oss/
oss_sblive: no errors on load, full set of files in /dev/oss/oss_sblive0/ but as soon as I try osstest I get simply spammed with these:

Code: Select all

[  329.629665] dmar_fault: 26991 callbacks suppressed
[  329.629668] DMAR: DRHD: handling fault status reg 702
[  329.629784] DMAR: [DMA Read] Request device [00:1e.0] fault addr 80000 [fault reason 06] PTE Read access is not set
[  329.629963] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.630142] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.630323] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
-----------------------------cut 1684 lines---------------------
[  329.970165] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.970347] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.970529] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.970711] DMAR: [DMA Read] Request device [00:1e.0] fault addr 80000 [fault reason 06] PTE Read access is not set
[  329.970892] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.971074] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.971254] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
-----------------------------cut 100 lines---------------------
[  329.991384] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.991569] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.991754] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.991937] DMAR: [DMA Read] Request device [00:1e.0] fault addr 80000 [fault reason 06] PTE Read access is not set
[  329.992121] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.992305] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  329.992489] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
-----------------------------cut 101 lines---------------------
[  330.012816] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.012999] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.013179] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.013362] DMAR: [DMA Read] Request device [00:1e.0] fault addr 80000 [fault reason 06] PTE Read access is not set
[  330.013543] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.013729] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.013911] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
-----------------------------cut 69 lines---------------------
[  330.028364] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.028554] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.028561] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.028578] DMAR: DRHD: handling fault status reg 500
[  330.028730] DMAR: DRHD: handling fault status reg 502
[  330.028734] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.028918] DMAR: DRHD: handling fault status reg 602
[  330.028923] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029082] DMAR: DRHD: handling fault status reg 702
[  330.029087] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029214] DMAR: DRHD: handling fault status reg 2
[  330.029218] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029221] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029224] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029227] DMAR: DRHD: handling fault status reg 300
[  330.029381] DMAR: DRHD: handling fault status reg 302
[  330.029384] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029387] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029390] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set
[  330.029393] DMAR: DRHD: handling fault status reg 600
[  330.029582] DMAR: DRHD: handling fault status reg 602
[  330.029587] DMAR: [DMA Read] Request device [00:1e.0] fault addr 7ffff000 [fault reason 06] PTE Read access is not set

[hash]

By the way:
gcc-4.8.5, glibc-2.23-r2, OSS 4.2 (b 2006/201608181820) (0x00040100)
Linux 4.7.2-gentoo #1 SMP x86_64 Intel(R) Core(TM) i7-3820 CPU @ 3.60GHz GenuineIntel GNU/Linux

[igorzwx]

By the way:
gcc-4.8.5, glibc-2.23-r2, OSS 4.2 (b 2006/201608181820) (0x00040100)
Linux 4.7.2-gentoo #1 SMP x86_64 Intel(R) Core(TM) i7-3820 CPU @ 3.60GHz GenuineIntel GNU/Linux

It seems that nobody else reported such problems.
Your problems might be unique, because you are using your own very special version of Gentoo.

Perhaps, you may find some useful info in the manual for Ubuntu newbies
_http://ossnext.trueinstruments.com/forum/viewtopic.php?f=3&t=5828

[hash]

Sadly no manual for Ubuntu newbies managed to help me on my problem but I found the culprit my self:
When I was "upgrading" from kernel 4.6.3 to 4.7.2, I decided to make some changes, among them was enabling IOMMU and, in particular, Intel-IOMMU.
After searching a more "generalized" error "spam"

Code: Select all

DMAR: [DMA Read] Request device fault addr [fault reason 06] PTE Read access is not set

I've came across this bug _https://bugzilla.kernel.org/show_bug.cgi?id=82761 , it is old, yes, but Intel's realization of IOMMU haven't changed since then at all.
Supplying kernel with parameter:

Code: Select all

intel_iommu=off

Fixes all problems, but Intel-IOMMU is some times needed for i915 and i965 drivers and as you might guess, on notebook, there isn't many options of not using that...

My theory is that OSSv4 do some thing incorrectly with DMA remapping, but I am far from any sort of expert in kernel programming or programming at all!

PS: I'm making thread as [semi-SOLVED] since some one might need Intel-IOMMU and might want OSSv4, so the problem definitely need to be look in if any one will ever look in any real problems with OSSv4 ever again.

[igorzwx]

Sadly no manual for Ubuntu newbies managed to help me on my problem but I found the culprit my self:
When I was "upgrading" from kernel 4.6.3 to 4.7.2, I decided to make some changes, among them was enabling IOMMU and, in particular, Intel-IOMMU.
After searching a more "generalized" error "spam"

Code: Select all

DMAR: [DMA Read] Request device fault addr [fault reason 06] PTE Read access is not set

I've came across this bug _https://bugzilla.kernel.org/show_bug.cgi?id=82761 , it is old, yes, but Intel's realization of IOMMU haven't changed since then at all.
Supplying kernel with parameter:

Code: Select all

intel_iommu=off

Fixes all problems, but Intel-IOMMU is some times needed for i915 and i965 drivers and as you might guess, on notebook, there isn't many options of not using that...

My theory is that OSSv4 do some thing incorrectly with DMA remapping, but I am far from any sort of expert in kernel programming or programming at all!

PS: I'm making thread as [semi-SOLVED] since some one might need Intel-IOMMU and might want OSSv4, so the problem definitely need to be look in if any one will ever look in any real problems with OSSv4 ever again.

It seems that this problem was already solved some years ago.

Re: Asus m5a99fx pro r2.0 - alc982 [Solved]
Post by donroc » Wed Mar 06, 2013 7:13 pm

Just an FYI, I passed the kernel parameter "iommu=pt" on the grub line and sound is working again.
_http://ossnext.trueinstruments.com/forum/viewtopic.php?f=3&t=5203&p=19272#p19272

Using Linux IOMMU Pass-Through to Run DPDK with Intel® VT-d
_http://dpdk.readthedocs.io/en/v16.04/linux_gsg/enable_func.html#using-linux-iommu-pass-through-to-run-dpdk-with-intel-vt-d

[hash]

Yes, I've also seen that topic.
I tried iommu=pt, no luck for me.

[igorzwx]

Yes, I've also seen that topic.
I tried iommu=pt, no luck for me.

It seems that it works for Ubuntu users:

Re: My HDAudio is no longer recognized
Postby keantoken » Wed Feb 19, 2014 10:10 pm

Well, using the iommu=pt kernel option works, so I don't have to disable IOMMU. Finally, I have sound.
_http://ossnext.trueinstruments.com/forum/viewtopic.php?f=3&t=5595&p=20442#p20442

If you do not like Ubuntu, you may try a sort of Arch Linux kernel with some patches from the OSS4 package for Arch Linux
_http://ossnext.trueinstruments.com/forum/viewtopic.php?f=3&t=5828#p21527

[hash]

Thank You Igor, but I prefer Gentoo.
Not Arch because I would start using AUR and when you do, your distribution starts slowly migrate to almost full AUR, and AUR is far too weak in comparison with Gentoo Ebuild system.
Not *buntu simply because it is release based, I ran from "Windows" world exactly because of that, new version, reinstall...., and all their attempts at making "live" Ubuntu are a bit pathetic, also the amount of simply crazy patches Ubuntu devs add to many packages, just Wine alone is enough to think many times not to go Ubuntu way.

[igorzwx]

Thank You Igor, but I prefer Gentoo.
Not Arch because I would start using AUR and when you do, your distribution starts slowly migrate to almost full AUR, and AUR is far too weak in comparison with Gentoo Ebuild system.
Not *buntu simply because it is release based, I ran from "Windows" world exactly because of that, new version, reinstall...., and all their attempts at making "live" Ubuntu are a bit pathetic, also the amount of simply crazy patches Ubuntu devs add to many packages, just Wine alone is enough to think many times not to go Ubuntu way.

There are Ubuntu kernels and Arch Linux kernels.
You may try to study them.
It may help you to compile your own kernel for Gentoo.

_______________________________

linux-git 4.8rc4.r0.g3eab887-1
_https://aur.archlinux.org/packages/linux-git/

It is not difficult to use Ubuntu kernels with Arch Linux, for example:
linux-lts-selinux 4.4.19-1
_https://aur.archlinux.org/packages/linux-lts-selinux/
_https://aur.archlinux.org/packages/?O=0&SeB=nd&K=linux-lts-selinux

It might be possible to utilize "foreign kernels" with Gentoo.