Re: [Qemu-devel] [PATCH 2/2] scsi: add block job opblockers for scsi-block

[Kevin Wolf]

Am 12.02.2018 um 15:32 hat Paolo Bonzini geschrieben:
> On 12/02/2018 15:30, Kevin Wolf wrote:
> >>> We shouldn't be adding new instances of BLOCK_OP_* at all. I couldn't
> >>> find the time yet to remove the existing ones, but any new protections
> >>> should be using the permission system.
> >> I agree.  But does this include not fixing bugs wherever clients are
> >> using the old op blockers?
> > I'm not saying that we shouldn't fix the bug, just that we should fix it
> > properly with the best infrastructure we have.
> > 
> > The old op blockers are "fixing" the problem at the symptom level, and
> > you have to check for each high-level operation if it does something
> > problematic internally. You have to repeat this analysis every time you
> > add a new operation or modifiy an existing one (which noone ever does).
> > The risk that this breaks sooner or later is pretty high.
> > 
> > The new permission system, on the other hand, directly addresses the
> > root cause, and any new feature that uses dirty bitmaps will then
> > automatically get the protection, too.
> > 
> > So in fact, I would say that the bug isn't really fixed (but at best
> > papered over) until we add a proper fix on the permission level.
> 
> Okay, we are in agreement about this and you expressed very well why I
> (at the gut feeling level) didn't like the old op blockers.  But you
> bypassed the real question, which is: should I send a pull request for
> these two patches or not? :)

I didn't spell it out that explicitly, but this is essentially a NACK.
I'd very much prefer if you could replace it with the proper solution.
Of course, we can always make exceptions when there is a good reason,
but with 2.12 still two months away, I doubt we have one.

Kevin