Dear* Sombat Tragolgosol, *

*My name is Gary and I noticed your post on ozlabs.org about running the libspe2 testsuite under modern kernels and getting some failures. I was wondering if you ever found a solution for this issue. I run a custom Linux on my PS3 (basically the latest ppc64 fedora with a PS3 compatible kernel). I'm currently running on kernel 6.6.142. I've built a ppu-gcc and a spu-gcc from gcc 8.3.0 and now I'm trying to package a working libspe2-2.3.0.*

*On testing, the first test failure I see is test_single_gang_context.elf which causes a segfault. If you could help me analyze the trace and point me in the right direction, I would appreciate it so much.*

*dmesg returns:*

*[10814.901983] BUG: Kernel NULL pointer dereference on read at 0x00000038
[10814.902009] Faulting instruction address: 0xc000000000299bb8
[10814.902031] Oops: Kernel access of bad area, sig: 11 [#4]
[10814.902048] BE PAGE_SIZE=4K MMU=Hash SMP NR_CPUS=2 NUMA PS3
[10814.902068] Modules linked in: snd_ps3 snd_seq snd_seq_device snd_pcm snd_timer snd soundcore ps3vram [10814.902173] CPU: 0 PID: 1206 Comm: scratch Tainted: G      D W          6.6.142-10.PS3_1.ppc64 #1
[10814.902196] Hardware name: SonyPS3 Cell Broadband Engine 0x701000 PS3
[10814.902212] NIP:  c000000000299bb8 LR: c000000000299b84 CTR: c000000000295750 [10814.902228] REGS: c0000000035b39d0 TRAP: 0300   Tainted: G   D W           (6.6.142-10.PS3_1.ppc64) [10814.902247] MSR:  8000000000008032 <SF,EE,IR,DR,RI> CR: 22000822  XER: 20000000
[10814.902335] DAR: 0000000000000038 DSISR: 40000000 IRQMASK: 0
               GPR00: c000000000299b84 c0000000035b3c70 c000000000b72000 c00000000220b0c0                GPR04: c000000001972900 c000000001fc7798 0000000000000008 0000000000000000                GPR08: 0000000000000000 0000000000000000 0000000000000010 c000000001bb5a80                GPR12: c000000008551900 c000000000d6c000 00000001064cb110 00000001063b9710
               GPR16: 00000001063b9714
[10814.910166] systemd-journald[93]: /dev/kmsg buffer overrun, some messages lost.
[10814.913614] 00000001063b9788 0000000106351408 000000004822228a
               GPR20: 00000001063bd558 00000001064f6700 00000001063e0210 00000001063baf84                GPR24: 0000000000000000 0000000000000000 00003fffe4e70798 c0000000032f0290                GPR28: c00000000220b0c0 c000000003516e20 00000000000a801d c000000001fc7700
[10815.241151] Call Trace:
[10815.241169] [c0000000035b3c70] [c000000000299b84] 0xc000000000299b84 (unreliable)
[10815.241207] [c0000000035b3d10] [c000000000295790] 0xc000000000295790
[10815.241231] [c0000000035b3d90] [c00000000001ca94] 0xc00000000001ca94
[10815.241257] [c0000000035b3e10] [c00000000000b354] 0xc00000000000b354
[10815.241284] --- interrupt: c00 at 0x00003fff9f0dd3d8
[10815.241299] NIP:  00003fff9f0dd3d8 LR: 00003fff9f0dd3a4 CTR: 0000000000000000 [10815.241317] REGS: c0000000035b3e80 TRAP: 0c00   Tainted: G   D W           (6.6.142-10.PS3_1.ppc64) [10815.241338] MSR:  800000000200e032 <SF,VEC,EE,PR,FP,IR,DR,RI>  CR: 22000822  XER: 00000000
[10815.241452] IRQMASK: 0
               GPR00: 0000000000000006 00003fffe4e6ffd0 00003fff9f109f00 0000000000000003                GPR04: 0000000010021c20 0000000000000004 000000000000001f 00003fff9f196b98                GPR08: 0000000000000002 0000000000000000 0000000000000000 0000000000000000                GPR12: 0000000000000000 00003fff9f19e200 00000001064cb110 00000001063b9710                GPR16: 00000001063b9714 00000001063b9788 0000000106351408 000000004822228a                GPR20: 00000001063bd558 00000001064f6700 00000001063e0210 00000001063baf84                GPR24: 0000000000000000 0000000000000000 00003fffe4e70798 0000000000000001                GPR28: 0000000010002948 0000000010002510 0000000000000000 0000000000000003
[10815.314808] --- interrupt: c00
[10815.317969] Code: 813f0014 39400008 71280002 41820194 5528044a 55080163 40820054 75290200 38bf0098 408202a4 e8650008 e9030030 <e9280038> e9290440 2fa90000 419e0030
[10815.321344] ---[ end trace 0000000000000000 ]---*

*Unfortunately I have much to learn when it comes to this level of debugging. I do know that de-referencing a NULL pointer is a big time catastrophe though.*

*I've been writing little dummy programs just to test out functions like spe_context_create and spe_gang_context_create but haven't been able to reproduce the segfault yet...*

*Anyway hope you are well. Thanks for your consideration!*

*Gary*

*(https://ps3linux.net admin)*

Reply via email to