Hello,

I'm using 2.6.39.1 version of the Linux kernel. In my block device
driver, I'm queuing a struct work_struct from the bottom half of the
interrupt handler, i.e., in the ->bi_end_io function of a previously
submitted BIO. The work queue handler function then submits the next
BIO to be processed. The system becomes unresponsive when we do this.
We believe that calling submit_bio() from our work queue handler
function in this setup is causing the problem. Can someone explain
what the problem is in this scenario?

Thanks,
Pallav