Re: 答复: rwl manipulation fails when

First, read this:
https://www.rtai.org/userfiles/documentation/magma/html/api/whatis_lxrt.html

For each real-time thread in user-space, RTAI creates a "buddy" thread in
kernel space (that's what rt_task_init_schmod does). It's the buddy thread
that does the actual locking and unlocking of the RWL, and that's why your
thread needs to have done rt_task_init_schmod before. Without it, the buddy
thread doesn't exist and hence the kernel oops: "BUG: unable to handle
kernel NULL pointer dereference". When you allocate an RWL, you are just
allocating stuff, and you don't need a special kernel-space thread (your
buddy) to do it on your behalf; it's a non real-time task.

The buddy system is not ideal, and it sure does make things complicated
(some things can only be done in a real-time thread, something can only be
done in non real-time threads etc). However, it does give you some
interesting options, for example you can share an RWL between user-space
and kernel-space!

If you understand the buddy thread, things become a lot clearer!

On Thu, Dec 1, 2016 at 8:24 PM, Yamato Kira <[email protected]> wrote:

> Dear Shahbaz:
>
>         How wise you are!  The RWL can be locked and unlocked now~
>
>         However, I found that I can init a semaphore before
> rt_task_init_schmod and signal/wait the semaphore successfully after
> rt_task_init_schmod and rt_make_hard_real_time. Is that unsafe too?
>
>         What's more important , would you tell me how to fix the reason
> that cause the problem?
>
>
>         3Q~
>
>
>
>
>
>
>                                                       yours: C.P.
>
>
>
>                                                      2016-12-02
>
>
> 发送自 Outlook <http://aka.ms/weboutlook>
>
>
> ------------------------------
> *发件人:* Shahbaz Youssefi <[email protected]>
> *发送时间:* 2016年12月1日 23:38
> *收件人:* Yamato Kira
> *抄送:* rt ai
> *主题:* Re: [Rtai] rwl manipulation fails when
>
> Is your thread a realtime thread? You can't use the lock unless your
> thread is realtime (not necessarily hard realtime). You would rt_task_init*
> functions are your friends.
>
> On Thu, Dec 1, 2016 at 3:29 AM, Yamato Kira <[email protected]> wrote:
>
>> Hi, everyone
>>
>>
>> I am trying to use the RWL to achieve a synchronization of data
>> processing on LinuxCNC which has an RTAI core of 3.9.
>>
>>
>> The RWL I used can be initialized correctly in a thread, but it caused a
>> kernel error when I try to lock it or unlock it. Have you ever meet this
>> problem before? Or can you help me to fix it? 3Q~
>>
>>
>> The RWL codes I used read as:
>>
>>
>> **************************************************
>>
>> RWL * rt_RWL_01;
>>
>> ......
>>
>> rt_RWL_01 = rt_rwl_init(nam2num("RWL1"));
>>     if(0==rt_RWL_01){
>>         printf("RWL Init Fails %d\n",rt_RWL_01);
>>     }else{
>>         printf("RWL Init Succeeds %d\n",rt_RWL_01);
>>     }
>> ......
>> while(1){
>> ......
>>         printf("Before RWL\n");
>>         rt_rwl_wrlock(rt_RWL_01);
>>         //rt_rwl_unlock(rt_RWL_01);
>>         printf("After RWL\n");
>> }
>> **************************************************
>>
>> and the dmesg command gives:
>>
>> kern  :alert : [ 2855.372788] BUG: unable to handle kernel NULL pointer
>> dereference at 0000007c
>> kern  :alert : [ 2855.372793] IP: [<f86f2760>] rt_rwl_wrlock+0x30/0x200
>> [rtai_sched]
>> kern  :warn  : [ 2855.372798] *pde = 00000000
>> kern  :warn  : [ 2855.372801] Oops: 0000 [#2] PREEMPT SMP
>> kern  :warn  : [ 2855.372804] Modules linked in: rtai_sched(O)
>> rtai_hal(O) parport_pc ppdev lp parport rfcomm bnep uinput nfsd nfs nfs_acl
>> auth_rpcgss fscache lockd sunrpc loop snd_hda_codec_realtek btusb bluetooth
>> uvcvideo snd_hda_intel snd_hda_codec videobuf2_vmalloc videobuf2_memops
>> videobuf2_core videodev media ideapad_laptop i2c_core sparse_keymap
>> snd_hwdep snd_pcm coretemp rfkill snd_page_alloc evdev serio_raw snd_seq
>> snd_seq_device snd_timer snd crc32c_intel aesni_intel aes_i586 aes_generic
>> battery ac cryptd power_supply wmi soundcore video button ext4 crc16 jbd2
>> mbcache usbhid hid sg sd_mod crc_t10dif microcode xhci_hcd ahci libahci
>> r8169 mii libata scsi_mod ehci_hcd fan thermal_sys usbcore usb_common [last
>> unloaded: psmouse]
>> kern  :warn  : [ 2855.372859]
>> kern  :warn  : [ 2855.372862] Pid: 4345, comm: RTAI1 Tainted: G      D
>> O 3.4-9-rtai-686-pae #1 LENOVO 80HR/ZHAOYANG E40-80
>> kern  :warn  : [ 2855.372866] EIP: 0060:[<f86f2760>] EFLAGS: 00213002
>> CPU: 1
>> kern  :warn  : [ 2855.372870] EIP is at rt_rwl_wrlock+0x30/0x200
>> [rtai_sched]
>> kern  :warn  : [ 2855.372872] EAX: 00000200 EBX: 00000000 ECX: f8702c00
>> EDX: 00000000
>> kern  :warn  : [ 2855.372874] ESI: 00000030 EDI: 00000201 EBP: f2b19f70
>> ESP: f2b19ee4
>> kern  :warn  : [ 2855.372876]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS:
>> 0068
>> kern  :warn  : [ 2855.372877] CR0: 80050033 CR2: 0000007c CR3: 36a9f000
>> CR4: 001407d0
>> kern  :warn  : [ 2855.372879] DR0: 00000000 DR1: 00000000 DR2: 00000000
>> DR3: 00000000
>> kern  :warn  : [ 2855.372881] DR6: fffe0ff0 DR7: 00000400
>> kern  :emerg : [ 2855.372883] Process RTAI1 (pid: 4345, ti=f2b18000
>> task=f2b4ad80 task.ti=f2b18000)
>> kern  :emerg : [ 2855.372885] I-pipe domain Linux
>> kern  :emerg : [ 2855.372886] Stack:
>> kern  :warn  : [ 2855.372887]  f77c8e00 d17a140a 00000298 1dcd6500
>> f8766220 f86f2730 b74ea324 f86e4375
>> kern  :warn  : [ 2855.372894]  00000000 b76514e0 00000000 f8766220
>> 003d0f00 b74ea388 08049ab9 00000000
>> kern  :warn  : [ 2855.372901]  00000001 c1053a4d f77c8e00 f77c8e00
>> f77c8e00 f2b4ad80 f2b4af40 00000001
>> kern  :emerg : [ 2855.372908] Call Trace:
>> kern  :warn  : [ 2855.372913]  [<f86f2730>] ? rt_rwl_rdlock+0x210/0x210
>> [rtai_sched]
>> kern  :warn  : [ 2855.372917]  [<f86e4375>] ?
>> handle_lxrt_request+0xa5/0xaf0 [rtai_sched]
>> kern  :warn  : [ 2855.372922]  [<c1053a4d>] ?
>> pick_next_task_fair+0x4d/0x110
>> kern  :warn  : [ 2855.372925]  [<c1362a62>] ? __schedule+0x132/0x400
>> kern  :warn  : [ 2855.372930]  [<f86e4e18>] ? rtai_lxrt_invoke+0x58/0x180
>> [rtai_sched]
>> kern  :warn  : [ 2855.372934]  [<f86e3945>] ?
>> lxrt_intercept_syscall+0x75/0x100 [rtai_sched]
>> kern  :warn  : [ 2855.372938]  [<c10e851b>] ? sys_write+0x5b/0x70
>> kern  :warn  : [ 2855.372941]  [<c10979ec>] ?
>> __ipipe_notify_kevent+0xc/0x10
>> kern  :warn  : [ 2855.372944]  [<c101af46>] ?
>> __ipipe_syscall_root+0x36/0xd0
>> kern  :warn  : [ 2855.372947]  [<c1368bff>] ? sysenter_past_esp+0x5c/0x73
>> kern  :emerg : [ 2855.372948] Code: 10 8b 5c 24 20 9c 58 fa 25 00 02 00
>> 00 64 8b 15 10 e7 4f c1 f0 0f ab 15 90 d1 72 f8 19 d2 85 d2 0f 84 ed 00 00
>> 00 89 c7 8d 73 30 <8b> 43 7c 85 c0 0f 84 85 00 00 00 89 34 24 e8 4d e9 ff
>> ff 3d 00
>> kern  :emerg : [ 2855.372991] EIP: [<f86f2760>] rt_rwl_wrlock+0x30/0x200
>> [rtai_sched] SS:ESP 0068:f2b19ee4
>> kern  :warn  : [ 2855.372996] CR2: 000000000000007c
>> kern  :warn  : [ 2855.372999] ---[ end trace 61a57933bf2bc264 ]---
>> kern  :err   : [ 2855.373002] BUG: Unhandled exception over domain RTAI
>> at 0xf86f2760 - switching to ROOT
>> kern  :warn  : [ 2855.373004] Pid: 4345, comm: RTAI1 Tainted: G      D
>> O 3.4-9-rtai-686-pae #1
>> kern  :warn  : [ 2855.373005] Call Trace:
>> kern  :warn  : [ 2855.373007]  [<c101adb0>] ?
>> __ipipe_handle_exception+0x220/0x230
>> kern  :warn  : [ 2855.373009]  [<f86f2760>] ? rt_rwl_wrlock+0x30/0x200
>> [rtai_sched]
>> kern  :warn  : [ 2855.373011]  [<c13647f0>] ? error_code+0x6c/0x7c
>> kern  :warn  : [ 2855.373013]  [<c104007b>] ?
>> __alloc_workqueue_key+0x36b/0x3e0
>> kern  :warn  : [ 2855.373014]  [<f86f2760>] ? rt_rwl_wrlock+0x30/0x200
>> [rtai_sched]
>> kern  :warn  : [ 2855.373016]  [<f86f2730>] ? rt_rwl_rdlock+0x210/0x210
>> [rtai_sched]
>> kern  :warn  : [ 2855.373018]  [<f86e4375>] ?
>> handle_lxrt_request+0xa5/0xaf0 [rtai_sched]
>> kern  :warn  : [ 2855.373020]  [<c1053a4d>] ?
>> pick_next_task_fair+0x4d/0x110
>> kern  :warn  : [ 2855.373022]  [<c1362a62>] ? __schedule+0x132/0x400
>> kern  :warn  : [ 2855.373024]  [<f86e4e18>] ? rtai_lxrt_invoke+0x58/0x180
>> [rtai_sched]
>> kern  :warn  : [ 2855.373026]  [<f86e3945>] ?
>> lxrt_intercept_syscall+0x75/0x100 [rtai_sched]
>> kern  :warn  : [ 2855.373028]  [<c10e851b>] ? sys_write+0x5b/0x70
>> kern  :warn  : [ 2855.373030]  [<c10979ec>] ?
>> __ipipe_notify_kevent+0xc/0x10
>> kern  :warn  : [ 2855.373032]  [<c101af46>] ?
>> __ipipe_syscall_root+0x36/0xd0
>> kern  :warn  : [ 2855.373033]  [<c1368bff>] ? sysenter_past_esp+0x5c/0x73
>> kern  :info  : [ 2855.373036] note: RTAI1[4345] exited with preempt_count
>> 2
>> kern  :warn  : [ 2855.373097] LXRT releases PID 4345 (ID: RTAI1).
>> kern  :warn  : [ 3065.563237] LXRT releases PID 4346 (ID: RTAI1).
>> kern  :warn  : [ 3065.572713] LXRT releases PID 4344 (ID: RTAI1).
>>
>> **************************************************
>>
>>
>> _______________________________________________
>> Rtai mailing list
>> [email protected]
>> https://mail.rtai.org/cgi-bin/mailman/listinfo/rtai
>>
>
>

_______________________________________________
Rtai mailing list
[email protected]
https://mail.rtai.org/cgi-bin/mailman/listinfo/rtai