Re: race condition between shutdown() and post_reset() resulting in deadlock

2018-01-16 Thread Hans de Goede
Hi, On 16-01-18 10:57, Oliver Neukum wrote: Hi, looking at your last patch I noticed something. I think it fails to work if we hit a peculiar race condition. It looks to me like we can get the following: task A pre_reset()   --> calling scsi_block_requests()

race condition between shutdown() and post_reset() resulting in deadlock

2018-01-16 Thread Oliver Neukum
Hi, looking at your last patch I noticed something. I think it fails to work if we hit a peculiar race condition. It looks to me like we can get the following: task A pre_reset()   --> calling scsi_block_requests() task B