Update arm-gic to 0.7

[korken89]

No description provided.

[korken89]

Ops, broke something. Looking into it.

[korken89]

I gave their CHANGELOG a look and also went through their example - and could not really find anything that's changed which would make interrupts stop working.

I'll look a bit more tonight, but if someone sees the obvious error feel free to ping me. :)

[jonathanpallant]

uuuugh why isn't 0.7 tagged in the repo

I will go complain on their Discord

[jonathanpallant]

• test.sh MISMATCH: Binary gic-map for target armv8r-none-eabihf mismatched
• test.sh MISMATCH: Binary gic-static-section-irq for target armv8r-none-eabihf mismatched
• test.sh MISMATCH: Binary gic-unified-irq for target armv8r-none-eabihf mismatched

[jonathanpallant]

These tests pass locally for me, using QEMU 10.1.

No, due to #59 I wasn't running the failing tests. Not a QEMU issue.

[jonathanpallant]

I was able to replicate by using QEMU 9 on x86-64 (in a VM, on my Mac).

Edit: and locally on my Mac using QEMU 10.

[jonathanpallant]

The examples need this code adding, to replace some set-up that 0.6.1 did and 0.7.0 does not:

    gic.distributor().configure_default_settings();
    gic.distributor().enable_affinity_routing_non_secure(true);
    gic.distributor().enable_group1_non_secure(true);

However, this is impossible due to a typo in the definition of the fn distributor() function in arm-gic.

They wrote:

pub fn distributor(&mut self) -> &'a mut GicDistributor<'_> {

They meant:

pub fn distributor(&mut self) -> &mut GicDistributor<'a> {

Because 'a is the lifetime annotation for the references inside the GicDistibutor structure.

[korken89]

When the new arm-gic is released I'll squash my cleanup stuff and provide a singular commit.

[korken89]

Alright, arm-gic 0.7.1 is released and this should now be ready! 🎉