ralink2860 crashes in combat_read

[czeidler]

anybody has a clue whats going wrong here? KDL after the first attempt to access the network.

This is hrev43739 but I also saw it before the recent bsd8 driver update.

frame               caller     <image>:function + offset
 0 cd648ba8 (+  32) 801245ca   <kernel_x86>:arch_debug_stack_trace + 0x0012
 1 cd648bcp_info 0x0
[net/ralink2860/0] [00:27:0d:e3:2d:08] reassoc succes 0x000b
 2 cd648bd8 (+  12) 8012998e   <kernel_x86>:arch_debug_call_with_fault_handler + 0x001b
 3 cd648be4 (+  48) 8009313e   <kernel_x86>:debug_call_with_fault_handler + 0x0056
 4 cd648c14 (ASSOC -> RUN arg 48
[net/ralink2860/0] sta_newstate: ASSOC -> RU717[0m [36m"PANIC: "[0m, [34m0x801809a0[0m [36m"object_cache: object %p is already freed"[0m, [34m0xcd648cc0[0m [36m""[0m, int32: [34m0[0m) + 0x0213
 5 cd648c54 (+  48) 80091c3b   <kernel_x86> kernel_debugger_internal([34m0x8016d717[0m [36m"PANIC: "[0m, [34m0x801809a0[0m [36m"object_cache: object %p is already freed"[0m, [34m0xcd648cc0[0m [36m""[0m, int32: x820f3400<00:27:0d:e3:2d:08> refcnt 3
[net/ralink2860/0] beacon :panic + 0x0024
 7 cd648cb4 (+  64) 800fde7a   <kernel_x86>:object_cache_free + 0x0062
 8 cd648cf4 (+  48) cd5badbd   </boot/system/add-ons/kernel/drivers/dev/net/ralink2860>:mb_free_ext + 0x0081
 9 cd648d24 (+  48) cd5bae20   </boot/system/add-ons/kernel/drivers/dev/net/ralink2860>:m_free + 0x0034
10 cd648d54 (+  48) cd5bad28   </boot/system/add-ons/kernel/drivers/dev/net/ralink2860>:m_freem + 0x0028
11 cd648d84 (+  64) cd5b0a10   </boot/system/add-ons/kernel/drivers/dev/net/ralink2860>:compat_read + 0x0190
12 cd648dc4 (+  64) 800a34b2   <kernel_x86> BPrivate::AbstractModuleDevice<[32m0xce583d00[0m>::Read([34m0x8227ed00[0m, int64: [34m8617[0m, int64: [34m-3646632513095964392[0m, [34m0xcd648e64[0m) + 0x002a
13 cd648e04 (+  80) 800a875e   <kernel_x86> devfs_read(fs_volume*: [34m0x82199e78[0m, fs_vnode*: [34m0xce7ad3c8[0m, [34m0xce7a58e0[0m, int64: [34m8617[0m, [34m0xce866918[0m, [34m0xcd648f0c[0m) + 0x0126
14 cd648e54 (+  64) 800dee5f   <kernel_x86> file_read(file_descriptor*: [34m0xce5f40b0[0m, int64: [34m8617[0m, [34m0xce866918[0m, [34m0xcd648f0c[0m) + 0x0067
15 cd648e94 (+  96) 800cb7e9   <kernel_x86>:_kern_read + 0x0149
16 cd648ef4 (+  64) 801407eb   <kernel_x86>:read + 0x001f
17 cd648f34 (+  64) cd626b8b   </boot/system/add-onde STA state RUN
[net/ralink2860/0] ieee80211_ref_node (ieee8021ice*: [34m0x823202a0[0m, net_buffer*: [34m0xcd648fb0[0m) + 0t/ralink2860/0] beacon miss, mode STA state RUN
[net/ralink2860/l/network/stack> device_reader_thread([34m0xce5edd58[0m) + 0x003b
19 cd648fb4 (+  48) 80079980   <kernel_x86> common_thread_entry([34m0xcd648ff0[0m) + 0x0080
kdebug> reboot

[korli]

Clemens,

the driver isn't integrated in FreeBSD, thus not updated periodically. FreeBSD compat layer was updated at least two times for the network layer and one time for the wireless layer.

You could :

• check if anything relevant was fixed in ​http://repo.or.cz/w/ralink_drivers/rt2860_fbsd8.git/tree since initial import in Haiku.
• check what has changed in the compat layer and require a change in the driver.

BTW if the driver is now unstable in trunk, we should probably not integrate it in the image anymore.