Install OPENSOUND results in KDL

[stargatefan]

Installing Opensound results in KDL on this machine. Listdev attached as well as picture of error. I attach the picture of the KDL "the pc freezes hard" Become unresponsive.

~> listdev
device Multimedia controller (Audio device) [4|3|0]
  vendor 1002: ATI Technologies Inc
  device aa58: Juniper HDMI Audio [Radeon HD 5700 Series]

device Multimedia controller (Audio device) [4|3|0]
  vendor 10de: nVidia Corporation
  device 0774: MCP72XE/MCP72P/MCP78U/MCP78S High Definition Audio

[stargatefan]

I get KDL at boot, I get KDL at desktop, It is breaking the media server.

I forgot the build and type

gcc4 Hybrid 39413

I could optionally test with a gcc2 or gcc2h.

In the meantime i am going to wipe the drive and reinstall as I don't know of a good way to clean the system.

[korli]

Please keep the png extension when attaching images. The current 2_2_3 extension prevents from viewing the image.

[mmu_man]

Please keep extensions at the end of filenames (and use the correct one, this seems to be a jpeg), else it screws up trac's preview.

[mmu_man]

Can you try to update to the latest package ? 4.2 is 1 year old and doesn't contain the fixes for pci device reservation.

[mmadia]

I haven't yet updated build/jam/OptionalPackages for the newer OpenSound, as it tends to also KDL on boot. luroh & I have tested it -- the native hda driver will work (properly) when oss is installed. however, if the native drivers are removed, then oss will KDL on boot.

Here's that newer build of OpenSound : ​http://www.haiku-files.org/files/optional-packages/opensound-232d0de71550-x86-gcc2-2010-11-20.zip

If you want, it's been added to haikuporter too as haikuporter -i opensound-232d0de71550

[stargatefan]

will fresh install and retry and report back.

[stargatefan]

Same error with gcc4h 39626.

could it be that its a gcc2 build and not a gcc4 build. could it also be that ingo added weak symbol support. Opensound used to work with gcc4 and gcc4h just fine.

[bonefish]

reuploaded image with usable name

[bonefish]

Replying to stargatefan:

Same error with gcc4h 39626.

could it be that its a gcc2 build and not a gcc4 build. could it also be that ingo added weak symbol support. Opensound used to work with gcc4 and gcc4h just fine.

The originally reported revision (hrev39413) predates the weak symbol branch merge (hrev39571).

[mmadia]

Replying to bonefish:

Replying to stargatefan:

Same error with gcc4h 39626.

could it be that its a gcc2 build and not a gcc4 build. could it also be that ingo added weak symbol support. Opensound used to work with gcc4 and gcc4h just fine.

The originally reported revision (hrev39413) predates the weak symbol branch merge (hrev39571).

For the heck of it, I rebuilt OpenSound with gcc4 on hrev39567. Once the native drivers are removed, OSS kicks in and KDL's

Could not load kernel add-on "/boot/system/add-ons/kernel/drivers/dev/audio/hmulti.zip": Not an executable
ps2: reset failed
ps2: devfs_publish_device input/mouse/ps2/0, status = 0xffffffff
oss_load_options(): oss_core
oss:init_osscore: 0x00000000
oss: matching pci 10de,03f0
oss_load_options(): oss_hdaudio
add_memory_type_range(3651, 0xffffffffdde78000, 0x4000, 0)
PANIC: Unexpected exception "General Protection Exception" occurred in kernel mode! Error code: 0x0

Welcome to Kernel Debugging Land...
Thread 92 "daemon" running on CPU 0
stack trace for thread 92 "daemon"
    kernel stack: 0x81861000 to 0x81865000
      user stack: 0x70000000 to 0x70040000
frame               caller     <image>:function + offset
 0 81864c08 (+  32) 80105dfe   <kernel_x86>:arch_debug_stack_trace + 0x0012
 1 81864c28 (+  16) 8007610f   <kernel_x86> stack_trace_trampoline(NULL) + 0x000b
 2 81864c38 (+  12) 8010b946   <kernel_x86>:arch_debug_call_with_fault_handler + 0x001b
 3 81864c44 (+  48) 80077bdb   <kernel_x86>:debug_call_with_fault_handler + 0x005b
 4 81864c74 (+  64) 80076333   <kernel_x86> kernel_debugger_loop(0x80147797 "PANIC: ", 0x80161600 "Unexpected exception "%s" occurred in kernel mode! Error code: 0x%lx
", 0x81864d20 "<binary>", int32: 0) + 0x021f
 5 81864cb4 (+  48) 80076694   <kernel_x86> kernel_debugger_internal(0x80147797 "PANIC: ", 0x80161600 "Unexpected exception "%s" occurred in kernel mode! Error code: 0x%lx
", 0x81864d20 "<binary>", int32: 0) + 0x0048
 6 81864ce4 (+  48) 80077f4c   <kernel_x86>:panic + 0x0024
 7 81864d14 (+  96) 8010779f   <kernel_x86> unexpected_exception(iframe*: 0x81864d80) + 0x0127
 8 81864d74 (+  12) 8010c8fd   <kernel_x86>:int_bottom + 0x003d
kernel iframe at 0x81864d80 (end = 0x81864dd0)
 eax 0xdde78000     ebx 0x81fb6208      ecx 0x206        edx 0xffffffff
 esi 0xdde78000     edi 0xffffffff      ebp 0x81864e0c   esp 0x81864db4
 eip 0x8010b88b  eflags 0x10006
 vector: 0xd, error code: 0x0
 9 81864d80 (+ 140) 8010b88b   <kernel_x86>:x86_write_msr + 0x000c
10 81864e0c (+  80) 81fb4bf5   </boot/system/add-ons/kernel/cpu/generic_x86>:generic_set_mtrrs + 0x00a1
11 81864e5c (+  48) 81fb4dd0   </boot/system/add-ons/kernel/cpu/generic_x86> amd_set_mtrrs(uint8: 0x6 (6), x86_mtrr_info*: 0x80183de0, uint32: 0x6 (6)) + 0x0034
12 81864e8c (+  48) 80103191   <kernel_x86> set_mtrrs(0x818a8688, int32: 0) + 0x0055
13 81864ebc (+  48) 8005e54a   <kernel_x86> process_pending_ici(int32: 0) + 0x0096
14 81864eec (+  48) 8005e669   <kernel_x86>:smp_intercpu_int_handler + 0x0015
15 81864f1c (+  32) 8010954c   <kernel_x86> i386_ici_interrupt(NULL) + 0x0014
16 81864f3c (+  48) 8004e2e2   <kernel_x86>:int_io_interrupt_handler + 0x005e
17 81864f6c (+  48) 80107ad6   <kernel_x86> hardware_interrupt(iframe*: 0x81864fa8) + 0x007a
18 81864f9c (+  12) 8010c966   <kernel_x86>:int_bottom_user + 0x005a
user iframe at 0x81864fa8 (end = 0x81865000)
 eax 0x1fa          ebx 0x58480c        ecx 0x180015a0   edx 0x1fb
 esi 0x180311f0     edi 0x180311f0      ebp 0x7003dcbc   esp 0x81864fdc
 eip 0x55c4c1    eflags 0x10297    user esp 0x7003dca4
 vector: 0xfd, error code: 0x0
19 81864fa8 (+   0) 0055c4c1   <libroot.so> BPrivate::superblock<0x180311f0>::isValid(0xffffffff) + 0x0045
20 7003dcbc (+  64) 0055705c   <libroot.so> BPrivate::hoardHeap<0x180015a0>::insertSuperblock(int32: 0, BPrivate::superblock*: 0x180311f0, BPrivate::processHeap*: 0x18000000) + 0x064c
21 7003dcfc (+  80) 0055ac3c   <libroot.so> BPrivate::threadHeap<0x180015a0>::malloc(uint32: 0x5 (5)) + 0x065c
22 7003dd4c (+  48) 0055b3d4   <libroot.so>:malloc + 0x0154
23 7003dd7c (+  48) 0055c06b   <libroot.so>:realloc + 0x0027
24 7003ddac (+  48) 002dd510   <libbe.so> BMessage<0x7003dee4>::_ResizeData(uint32: 0x0 (0), int32: 5) + 0x00d0
25 7003dddc (+  48) 002dd80e   <libbe.so> BMessage<0x7003dee4>::_AddField(0x2042b4 "from", uint32: 0x4c4f4e47, true, BMessage::field_header*: 0x7003de58) + 0x012e
26 7003de0c (+  80) 002dda1b   <libbe.so> BMessage<0x7003dee4>::AddData(0x2042b4 "from", uint32: 0x4c4f4e47, 0x7003deac, int32: 4, true, int32: 1) + 0x0087
27 7003de5c (+  64) 002df68e   <libbe.so> BMessage<0x7003dee4>::AddInt32(0x2042b4 "from", int32: 0) + 0x0036
28 7003de9c (+ 144) 00203c48   <_APP_>:init_syslog_output__FP12SyslogDaemon (nearest) + 0x0148
29 7003df2c (+8272) 0020331a   <_APP_> SyslogDaemon<0x7ffeee2c>::Daemon(0x0) + 0x00d2
30 7003ff7c (+  48) 0020336b   <_APP_> SyslogDaemon<0x7ffeee2c>::daemon_thread(NULL) + 0x001f
31 7003ffac (+  48) 004dd522   <libroot.so>:_get_next_team_info (nearest) + 0x0072
32 7003ffdc (+   0) 7003ffec   2609:daemon_92_stack@0x70000000 + 0x3ffec
kdebug> info
kernel build: Nov 22 2010 06:46:18 (gcc2 2.95.3-haiku-100818)
SVN revision: 39567

cpu count: 2, active times:
  [1] 4977682
  [2] 7366288
pages:		44284 (1048256 max)
sems:		972 (65536 max)
ports:		149 (4096 max)
threads:	130 (4096 max)
teams:		16 (2048 max)
kdebug> reboot

Considering this glaring issue, is it worth updating OptionalPackages? Should I attempt building this with a post-weak symbol merge?

[stargatefan]

I can confirm at gcc4H 39402 Opensound installs just fine and works. I tested this on my panasonic toughbook which has AUCH driver issues.

So its a recent problem somewhere. what changed between 39402 and 39413 ??? that would cuase this ? I hope this narrows it down enough to hunt around for it.

I geuss its a good thing that I only update that laptop very ocassionaly.

[bonefish]

Replying to stargatefan:

I can confirm at gcc4H 39402 Opensound installs just fine and works. I tested this on my panasonic toughbook which has AUCH driver issues.

So its a recent problem somewhere. what changed between 39402 and 39413 ??? that would cuase this ? I hope this narrows it down enough to hunt around for it.

Unfortunately that doesn't leave any reasonable candidates:

• hrev39403, hrev39408, hrev39411 are changes in the weak-symbols branch.
• hrev39404, hrev39405, hrev39406, hrev39409 changed documentation and a script for building a gcc 4 package in the buildtools.
• hrev39407 is a change in Debugger.
• hrev39410 is comment change.
• hrev39412 doesn't change semantics.
• hrev39413 adds POSIX locale macros, which shouldn't change anything kernel related.

The last one is really the only one that could possibly result in any difference between the images for the two revisions (save for the compiled-in revision number, that is).

[mmadia]

(In #6922) Setting this ticket as a blocker for #6849, as the current OpenSound package in build/jam/OptionalPackages (opensound-4.2-r1a2-x86-gcc2-2010-05-01.zip) does not contain the code to allow co-existence with native drivers.

[stargatefan]

Replying to bonefish:

Replying to stargatefan:

I can confirm at gcc4H 39402 Opensound installs just fine and works. I tested this on my panasonic toughbook which has AUCH driver issues.

So its a recent problem somewhere. what changed between 39402 and 39413 ??? that would cuase this ? I hope this narrows it down enough to hunt around for it.

Unfortunately that doesn't leave any reasonable candidates:

• hrev39403, hrev39408, hrev39411 are changes in the weak-symbols branch.
• hrev39404, hrev39405, hrev39406, hrev39409 changed documentation and a script for building a gcc 4 package in the buildtools.
• hrev39407 is a change in Debugger.
• hrev39410 is comment change.
• hrev39412 doesn't change semantics.
• hrev39413 adds POSIX locale macros, which shouldn't change anything kernel related.

The last one is really the only one that could possibly result in any difference between the images for the two revisions (save for the compiled-in revision number, that is).

Heres what I know. It worked at 39402, it didn;t work as early as 39413. One of the changes in between broke it. Its not a fluke. I pulled the 39402 cd and installed opensound on the machine that generated this ticket report. Installs just fine.

So it happened between 39402 and 39413. Maybe the posix support broke something else.

[stargatefan]

on gcc2H I am getting errors only with the envy24 and envy24ht driver attempting to aquire spinlock so far as I can tell. I will try to backtrace and screen shot the problem. I noticed similar behaviors favoring the envy drivers on the gcc4 builds as well.

could be that opensound does a poor job handling the envy chipset. wouldn't surprise me much.

[stargatefan]

Would anyone like the actual card cuasing the problem ? the via ice chipset maudio delta 1010lt card ?

If so I would be willing to ship it " on a good faith agremment it would be returned fairly promptly " I will pay shipping both ways.

Only way to get this fixed is to get the hardware to someone who can help.

[mmu_man]

The current 4.2_git-2 package has a fix for this (the KDL on 64bit). I thought I upstreamed it but it doesn't seem to be in their git tree yet.

Please test on a recent nightly.

[waddlesplash]

No reply in 2 years.

[nielx]

Remove milestone for tickets with status = closed and resolution != fixed