On Tue, Jul 14, 2020 at 02:08:02PM +0200, Joerg Roedel wrote: > The #VC entry code now tries to pretend that the #VC handler does not > use an IST stack by switching to the task stack if entered from > user-mode or the SYSCALL entry path. When it is entered from > kernel-mode it is doing its best to switch back to the interrupted > stack. This is only possible if it is entered from a known and safe > kernel stack (e.g. not the entry stack). If the previous stack is not > safe to use the #VC handler switches to a fall-back stack and calls a > special handler function which, as of now, just panics the system. For > now this is safe as #VC exceptions only happen at know places which > use a safe stack. > > The use of the fall-back stack is necessary so that the special > handler function can safely raise nested #VC exceptions, for > example to print a panic message. Can we get some more words -- preferably in actual code comments, on when exactly #VC happens? Because the only thing I remember is that #VC could happen on any memop, but I also have vague memories of that being a later extention.
