Indan Zupancic
Indan Zupancic
> IIRC commitTime is called when we know there's enough budget to charge the SC without having to do any shuffling of sporadic replenishments or simulating a timer tick. chargeBudget...
I added a test for this, see https://github.com/seL4/sel4test/pull/76.
Great, that would solve this issue. Is it [this](https://github.com/seL4/seL4/compare/master...axel-h:kernel-debug-log) branch?
Another issue is that `ipiStallCoreCallback` calls `restore_user_context` on the non-IRQ path where the lock is attempted to be acquired via `NODE_LOCK_SYS`. However, currently `c_entry_hook` is called _after_ `NODE_LOCK_SYS`, so it...
This is fixed by #847. (No idea why mcsPreemptionPoint gets the current irq passed, probably a left-over that can be removed.)
Fix is merged.
I made a fork of your branch with my suggestions, see [here](https://github.com/seL4/seL4/compare/master...Indanz:seL4:corlewis/set-handler-params). While working on that I think I found a corner case that is currently not handled correctly: If...
I implemented your suggestion to check for null cap and force pushed to my branch.
> I haven't been able to work out a nice way of calling the invocations that makes use of them having optional parameters. From libsel4's point of view it's just...
> What I've been trying to handle is the case where deriveCap can return a null cap if for whatever reason certain non-endpoint caps are passed in. Ah, I didn't...