On 7/3/23 09:49, Linus Torvalds wrote:
On Mon, 3 Jul 2023 at 00:08, Helge Deller deller@gmx.de wrote:
Great, that patch fixes it!
Yeah, I was pretty sure this was it, but it's good to have it confirmed. Committed.
FWIW, my qemu boot tests didn't find any problems with other architectures.
Guenter
I wonder if you want to #define VM_STACK_EARLY VM_GROWSDOWN even for the case where the stack grows down too (instead of 0), just to make clear that in both cases the stack goes downwards initially.
No, that wouldn't work for the simple reason that the special bits in VM_STACK_INCOMPLETE_SETUP are always cleared after the stack setup is done.
So if we added VM_GROWSDOWN to those early bits in general, the bit would then be cleared even when that wasn't the intent.
Yes, yes, we could change the VM_STACK_INCOMPLETE_SETUP logic to only clear some of the bits in the end, but the end result would be practically the same: we'd still have to do different things for grows-up vs grows-down cases, so the difference might as well be here in the VM_STACK_EARLY bit.
Linus