[RFC/PATCH 4/16] Abstract MSI suspend

[Eric W. Biederman]

Michael Ellerman <michael at ellerman.id.au> writes:

> That was my suspicion as well, I was hoping someone who knew the code
> better than me would pipe up and let me know why it was a special case.
> Have the original MSI authors vanished without a trace?

I have never gotten any useable feedback from that direction anyway.

> It seems to date from the initial MSI submission, and has only ever been
> called from pci_free_resources(). The rest of the pci hotunplug code
> paths are not clear to me though, so I don't know whether we can rely on
> pci_disable_msi() already being called for us.

Good question.  I do know I have always been a little suspicions of
the suspend/resume path, because I haven't read them.

There may be a point in the hot-unplug where the code actually makes
sense in the don't touch the hardware kind of way.

Eric