Re: [PATCH V3] nvme: fix multiple ctrl removal scheduling

2017-05-30 Thread Sagi Grimberg
Hi Rkesh, this looks reasonable, but we'll need to also adopt the non-PCI driver to the new state machine. I can give this a spin. At that point we probably want to move nvme_reset into common code somehow. Hi Guys, sorry for barging in late, I've been way too busy with internal stuff lately.

Re: [PATCH V3] nvme: fix multiple ctrl removal scheduling

2017-05-28 Thread Rakesh Pandit
On Sun, May 28, 2017 at 11:40:43AM +0200, Christoph Hellwig wrote: > Hi Rakesh, > > this looks reasonable, but we'll need to also adopt the non-PCI > driver to the new state machine. I can give this a spin. > .. Thanks for handling non-PCI driver part.

Re: [PATCH V3] nvme: fix multiple ctrl removal scheduling

2017-05-28 Thread Christoph Hellwig
Hi Rkesh, this looks reasonable, but we'll need to also adopt the non-PCI driver to the new state machine. I can give this a spin. At that point we probably want to move nvme_reset into common code somehow.

[PATCH V3] nvme: fix multiple ctrl removal scheduling

2017-05-26 Thread Rakesh Pandit
Commit c5f6ce97c1210 tries to address multiple resets but fails as work_busy doesn't involve any synchronization and can fail. This is reproducible easily as can be seen by WARNING below which is triggered with line: WARN_ON(dev->ctrl.state == NVME_CTRL_RESETTING) Allowing multiple resets can re