Hello Nicholas Bellinger,
The patch 2450acf26519: "target: Add support for COMPARE_AND_WRITE
emulation" from Aug 19, 2013, leads to the following static checker
warning: "drivers/target/target_core_sbc.c:508 sbc_compare_and_write()
warn: 'sem:&dev->caw_sem' is sometimes locked here and
sometimes unlocked."
drivers/target/target_core_sbc.c
502 /*
503 * Submit the READ first for COMPARE_AND_WRITE to perform the
504 * comparision using SGLs at cmd->t_bidi_data_sg..
505 */
506 rc = down_interruptible(&dev->caw_sem);
507 if ((rc != 0) || signal_pending(current))
508 return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
When we return here, we don't know if we are holding the lock or not.
I believe that we release this lock in compare_and_write_callback() so
this is actually a double unlock bug and not a deadlock.
509
510 ret = cmd->execute_rw(cmd, cmd->t_bidi_data_sg, cmd->t_bidi_data_nents,
511 DMA_FROM_DEVICE);
512 if (ret) {
513 up(&dev->caw_sem);
^^^^^^^^^^^^^^^^^^
This may be a double unlock bug as well.
514 return ret;
515 }
516 /*
517 * Unlock of dev->caw_sem to occur in compare_and_write_callback()
518 * upon MISCOMPARE, or in compare_and_write_done() upon completion
519 * of WRITE instance user-data.
520 */
521 return TCM_NO_SENSE;
regards,
dan carpenter
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
