Messages in this thread | | | From | Linus Torvalds <> | Date | Sun, 15 Apr 2012 11:59:19 -0700 | Subject | Re: [PATCH] PCI: Fix regression in pci_restore_state() |
| |
On Sun, Apr 15, 2012 at 11:47 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote: > > mdelay(10) doesn't really look good either to me in this case, though.
Oh, I agree. What kind of ass-backwards device actually needs that kind of crazy delays? It is almost certainly buggy.
With retries, 10ms delays are totally unacceptable. There's something wrong.
A single ms *may* be ok.
Anyway, can you also split the actual "write _one_ register with retry" into a function of its own? The code looks like crap with those multiple levels of looping, with conditionals inside them etc. With a simple helper function, you could change the break into return, and it would look much better, I bet.
Linus
| |