Re: [PATCH] scsi: storvsc: Fix a panic in the hibernation procedure

[Paul E. McKenney]

On Wed, Apr 22, 2020 at 05:23:51PM +0800, Ming Lei wrote:
> On Wed, Apr 22, 2020 at 04:58:14AM +0000, Dexuan Cui wrote:
> > > From: Ming Lei <ming.lei@xxxxxxxxxx>
> > > Sent: Tuesday, April 21, 2020 9:16 PM
> > > ...
> > > > > > When we're in storvsc_suspend(), all the userspace processes have been
> > > > > > frozen and all the file systems have been flushed, and there should not
> > > > > > be too much I/O from the kernel space, so IMO scsi_host_block() should
> > > be
> > > > > > pretty fast here.
> > > > >
> > > > > I guess it depends on RCU's implementation, so CC RCU guys.
> > > > >
> > > > > Hello Paul & Josh,
> > > > >
> > > > > Could you clarify that if sysnchronize_rcu becomes quickly during
> > > > > system suspend?
> > > >
> > > > Once you have all but one CPU offlined, it becomes extremely fast, as
> > > > in roughly a no-op (which is an idea of Josh's from back in the day).
> > > > But if there is more than one CPU online, then synchronize_rcu() still
> > > > takes on the order of several to several tens of jiffies.
> > > >
> > > > So, yes, in some portions of system suspend, synchronize_rcu() becomes
> > > > very fast indeed.
> > > 
> > > Hi Paul,
> > > 
> > > Thanks for your clarification.
> > > 
> > > In system suspend path, device is suspended before
> > > suspend_disable_secondary_cpus(),
> > > so I guess synchronize_rcu() is not quick enough even though user space
> > > processes and some kernel threads are frozen.
> > > 
> > > Thanks,
> > > Ming
> > 
> > storvsc_suspend() -> scsi_host_block() is only called in the hibernation
> > path, which is not a hot path at all, so IMHO we don't really care if it
> > takes 10ms or 100ms or even 1s. :-)  BTW, in my test, typically the
> 
> Are you sure the 'we' can cover all users?
> 
> > scsi_host_block() here takes about 3ms in my 40-vCPU VM.
> 
> If more LUNs are added, the time should be increased proportionallly,
> that is why I think scsi_host_block() is bad.

If the caller must wait until the grace period ends, then the traditional
approach is to use a single synchronize_rcu() to cover all LUNs.  This of
course can require some reworking of the code.

If the caller does not need to wait, then either call_rcu() or kfree_rcu()
can work quite well.

							Thanx, Paul