Re: [PATCH 4/8] IPMI: Don't grab locks in run-to-completion mode

On Wed, 13 Feb 2008 10:27:49 -0600 Corey Minyard <[EMAIL PROTECTED]> wrote: > This patch prevents deadlocks in IPMI panic handler caused by msg_lock > in smi_info structure and waiting_msgs_lock in ipmi_smi structure. Again, what are the newly-added barrier()s for? Was cpu_relax() intended? If

[PATCH 4/8] IPMI: Don't grab locks in run-to-completion mode

From: Konstantin Baydarov <[EMAIL PROTECTED]> This patch prevents deadlocks in IPMI panic handler caused by msg_lock in smi_info structure and waiting_msgs_lock in ipmi_smi structure. Signed-off-by: Konstantin Baydarov <[EMAIL PROTECTED]> Signed-off-by: Corey Minyard <[EMAIL PROTECTED]> --- In