• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 98 subscribers
  • Views 640 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ti dsplink examples with linux 3.1.10

alexander reddy
Prodigy 250 points
Other Parts Discussed in Thread: DM3730, OMAP3530

Hi,

Regarding DSPLink with linux 3.1.10

While running the ti dsplink eamples ....first time after board boot and inserting the module  ....on DM3730 based boards....i am getting following warning...........

second times..onwards...this warning is not coming...........

Please let me know the solution.....

dvsdk_dm3730-evm_04_03_00_06_setuplinux ...................all the packages are taken from this

1.dspbios_5_41_03_17

2.dsplink_1_65_01_05_eng

linux version 3.1.10 used for building dsplink & lpm & cmem

/opt/ti_dsplink_examples # ./loopgpp loop.out 1000 1000
=============== Sample Application : LOOP ==========
==== Executing sample for DSP processor Id 0 ====
Entered LOOP_Create ()
Leaving LOOP_Create ()
Entered LOOP_Execute ()
[   62.549468]
[   62.549468] =============================================
[   62.556762] [ INFO: possible recursive locking detected ]
[   62.562438] 3.1.10-svn70 #1
[   62.565399] ---------------------------------------------
[   62.571075] DSPLINK_DPC_2/550 is trying to acquire lock:
[   62.576690]  (&obj->lock){+.+...}, at: [<bf011fcc>] SYNC_SpinLockStartEx+0x14/0x20 [dsplinkk]
[   62.585906]
[   62.585906] but task is already holding lock:
[   62.592041]  (&obj->lock){+.+...}, at: [<bf011fcc>] SYNC_SpinLockStartEx+0x14/0x20 [dsplinkk]
[   62.601196]
[   62.601196] other info that might help us debug this:
[   62.608093]  Possible unsafe locking scenario:
[   62.608093]
[   62.614349]        CPU0
[   62.616912]        ----
[   62.619476]   lock(&obj->lock);
[   62.622802]   lock(&obj->lock);
[   62.626129]
[   62.626129]  *** DEADLOCK ***
[   62.626129]
[   62.632385]  May be due to missing lock nesting notation
[   62.632385]
[   62.639556] 1 lock held by DSPLINK_DPC_2/550:
[   62.644134]  #0:  (&obj->lock){+.+...}, at: [<bf011fcc>] SYNC_SpinLockStartEx+0x14/0x20 [dsplinkk]
[   62.653717]
[   62.653717] stack backtrace:
[   62.658386] [<c001be9c>] (unwind_backtrace+0x0/0x130) from [<c0092e18>] (__lock_acquire+0x18f0/0x1a88)
[   62.668212] [<c0092e18>] (__lock_acquire+0x18f0/0x1a88) from [<c0093590>] (lock_acquire+0x98/0x120)
[   62.677764] [<c0093590>] (lock_acquire+0x98/0x120) from [<c046d2fc>] (mutex_lock_interruptible_nested+0x5c/0x38c)
[   62.688720] [<c046d2fc>] (mutex_lock_interruptible_nested+0x5c/0x38c) from [<bf011fcc>] (SYNC_SpinLockStartEx+0x14/0x20 [dsplinkk])
[   62.701385] [<bf011fcc>] (SYNC_SpinLockStartEx+0x14/0x20 [dsplinkk]) from [<bf004930>] (IPS_notify+0x88/0x1ac [dsplinkk])
[   62.713104] [<bf004930>] (IPS_notify+0x88/0x1ac [dsplinkk]) from [<bf00b0a8>] (ZCPYDATA_send+0x100/0x18c [dsplinkk])
[   62.724365] [<bf00b0a8>] (ZCPYDATA_send+0x100/0x18c [dsplinkk]) from [<bf00b474>] (ZCPYDATA_dpc+0x340/0x3a8 [dsplinkk])
[   62.735931] [<bf00b474>] (ZCPYDATA_dpc+0x340/0x3a8 [dsplinkk]) from [<bf011268>] (DPC_Callback+0xa0/0xec [dsplinkk])
[   62.747131] [<bf011268>] (DPC_Callback+0xa0/0xec [dsplinkk]) from [<c007b124>] (kthread+0x80/0x88)
[   62.756622] [<c007b124>] (kthread+0x80/0x88) from [<c00159d8>] (kernel_thread_exit+0x0/0x8)
Transferred 1000 buffers
Leaving LOOP_Execute ()
Entered LOOP_Delete ()
Leaving LOOP_Delete ()
====================================================

Awaitng for the positive response

Regards

Alex

  • 0
    TI__Mastermind 18260 points

    alexander reddy said:

    While running the ti dsplink eamples ....first time after board boot and inserting the module  ....on DM3730 based boards....i am getting following warning...........

    second times..onwards...this warning is not coming...........

    Please let me know the solution.....

    Alex,

    We're aware of this situation.  It is something we'd like to prevent from happening, but the unfortunate reality is that the fix within DSPLink is quite extensive.  Since it's just a warning I believe it can be ignored, although I'm sure you'd rather not see the warning at all, even though it occurs for just the first run.

    The "fix" to prevent the nested lock situation lies within DSPLink code, so even if we had a plan to eliminate the situation (which we don't), it's there in your product.

    The warning is about a *possible* deadlock, but our understanding of the situation is that it will not deadlock.  The kernel is being paranoid because it's been configured to be paranoid (see below).

    alexander reddy said:

    Regarding DSPLink with linux 3.1.10

    Upgrading to a newer Linux release has most likely caused this warning to appear, even though the situation being warned about exists when DSPLink is used with older kernels too.  This particular warning is output only when certain Linux kernel configuration elements are set.  I believe those elements include:
    CONFIG_LOCKDEP_SUPPORT
    CONFIG_LOCKDEP
    CONFIG_DEBUG_KERNEL
    and others, but I'm not very knowledgeable about kernel configuration.

    So, although I cannot tell you a way to avoid the situation, we believe it will not lead to deadlock, and you can research how to get rid of the warning based on what I've told you above.

    Regards,

    - Rob

     

  • 0 in reply to Robert Tivy
    Prodigy 250 points

    Hi,

    After successful loading the DSP Link 1.65.01.06[dsplinkk.ko] on Linux kernel 3.13 version on omap3530 processor based board, we ran the following example

    /opt/gpp/BIN/Linux/OMAP3530/DEBUG/messagegpp  /opt/dsp/BIN/DspBios/OMAP3530/OMAP3530_0/DEBUG/message.out 1 0

    after executing the messagegpp ,  Linux system is not responding to any command.

    we have set mem=100M at u-boot boot arguments

    Could you let me know any solution to fix this issue?