Skip to content

Conversation

@midnightveil
Copy link

@midnightveil midnightveil commented Sep 24, 2025

Freya and others added 30 commits January 10, 2025 17:17
…nel elfs that had to be compiled, happy to be asked about anything
Running qemu with -d guest_errors would previously print out

    LDR|INFO: Setting all interrupts to Group 1
    LDR|INFO: GICv2 ITLinesNumber: 0x00000000
    gic_cpu_write: Bad offset 80
    gicv2m_write: Bad offset 4

The first write is supposed to be into the GIC_DIST region, then the
second into the GIC_CPU region. Cross referencing to QEMU [1], the GIC
addresses in loader.c are incorrect. Now microkit prints:

    LDR|INFO: Setting all interrupts to Group 1
    LDR|INFO: GICv2 ITLinesNumber: 0x00000008

The generated seL4 devices_gen.h from the DTS is correct, this is purely
a microkit issue.

[1]: https://github.com/qemu/qemu/blame/de278e54/hw/arm/virt.c#L164-L165

Signed-off-by: julia <[email protected]>
* setup the exception handling at the start of the loader
* setup exception handler for EL2 when we enable the MMU
  instead of just EL1.

Signed-off-by: Ivan-Velickovic <[email protected]>
Signed-off-by: julia <[email protected]>
Signed-off-by: julia <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Both of these match sel4bench, yay

Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
this is prelim for a cross core version of this.

name,mean,stddev,min,max,runs,sum,sum_squared
signal low to high same core,980.27163,17.20088158040001,945,1745,100000,98027163,96122833891
signal high to low same core,503.23385,23.43306646978741,493,1224,100000,50323385,25379341639
signal high to low cross core,483.34956,19.77788582751853,480,1251,100000,48334956,23401796192
signal 2way low to high same core,1960.52388,32.47431061232068,1900,2758,100000,196052388,384470846490

1960.52 is almost perfectly 2x 980.27, wow.

Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
... mostly. can't figure out what's going on with the
CTLR bits check in seL4

Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
this fixes translation faults due to us overwriting
uboot page tables.

Signed-off-by: Julia Vassiliki <[email protected]>
stops double interrupt

Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Signed-off-by: Julia Vassiliki <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants