Ecosyste.ms: Timeline
Browse the timeline of events for every public repo on GitHub. Data updated hourly from GH Archive.
r1mikey created a comment on a pull request on richlowe/illumos-gate
> It fails to detect the block device; this is the relevant boot log diff: > > ```diff > -Block device: blkdev@0,0, blkdev0 > -blkdev0 is /emmc2bus/mmc@7e340000/blkdev@0,0 > -/emmc2bus/mmc@7e...
richlowe pushed 1 commit to arm64/rpi4-pci-regspec-hack richlowe/illumos-gate
- bcm2711-emmc2: add many panic calls to mmc_init aef8850
richlowe created a comment on a pull request on richlowe/illumos-gate
the rpi4 drivers need to be written in terms of the DDI but aren't. That said, they also shouldn't break. I'm going to throw some appropriate `panic()` calls into `mmc_init` and update that test ...
hadfl created a comment on a pull request on richlowe/illumos-gate
I've booted this on the rpi 4 with an additional commit on top (thanks @richlowe ) that disables PCI. It fails to detect the block device; this is the relevant boot log diff: ```diff -Block de...
richlowe created a comment on a pull request on richlowe/illumos-gate
> > Did you deal with the PROM-based "dma-ranges" stuff? I've been meaning to, but forgetting. I think that's the last part we use the PROM but shouldn't. > Should we add this to the parent priv...
richlowe created a comment on a pull request on richlowe/illumos-gate
> > It was indeed shoddy DTB hacking, so I think it's probably fine but we'll find out :) > > What was the issue? I'd managed to screw it up so it thought we had both kinds of GIC at once som...
r1mikey created a comment on a pull request on richlowe/illumos-gate
> It was indeed shoddy DTB hacking, so I think it's probably fine but we'll find out :) What was the issue?
r1mikey created a review comment on a pull request on richlowe/illumos-gate
Do I? If it fails _or_ if `FIXED` is not in types, then panic? Or do you mean the text itself?
r1mikey created a review comment on a pull request on richlowe/illumos-gate
I need to dig into this to be sure, but using `oneshot_timer_reprogram` (from `i86pc/io/pcplusmp/apic_timer.c`) as a reference, I think we might need to tweak this. I do see this as us as potential...
richlowe pushed 1 commit to arm64/rpi4-pci-regspec-hack richlowe/illumos-gate
- stop delivering PCI bits, so rpi 4 can boot and try the rest 5843878
richlowe pushed 1 commit to arm64/rpi4-pci-regspec-hack richlowe/illumos-gate
- stop delivering PCI bits, so rpi 4 can boot and try the rest f2eead0
richlowe created a comment on a pull request on richlowe/illumos-gate
It was indeed shoddy DTB hacking, so I think it's probably fine but we'll find out :)
richlowe created a comment on a pull request on richlowe/illumos-gate
So I pulled this and #110 into the same branch, and tried to run with the hacked dtb that puts the console serial on the simple-bus, and we're crashing because we can't map the gic (v2). The addre...
richlowe created a branch on richlowe/illumos-gate
arm64/rpi4-pci-regspec-hack - Local changes to the Illumos OS/Net consolidation.
richlowe created a comment on a pull request on richlowe/illumos-gate
Oh, but this will crash on the pi because of the PCI bits, won't it :\
richlowe created a comment on a pull request on richlowe/illumos-gate
@hadfl can you give this a run? Or I can arrange the simple-bus-with-devices-under-it hack again, if I ever find out where I put the dtb part. Did you deal with the PROM-based "dma-ranges" stuf...
richlowe closed a pull request on richlowe/illumos-gate
gic: Move from `setlvl` and `setlvlx` to `intr_enter` and `intr_exit`
These names are at least very likely much more intelligible, and are inspired by the names of the PSM operations which underly them in the intel implementation. Unlike on intel, I have changed the...
richlowe created a review comment on a pull request on richlowe/illumos-gate
You have your error message backwards :)
richlowe created a review comment on a pull request on richlowe/illumos-gate
If this were to happen right now, would we drop the event? That seems bad, but I think also not actually different from before
r1mikey created a comment on a pull request on richlowe/illumos-gate
Retested on qemu-virt, this is good to go if somebody can test it on a pi4.
r1mikey created a review comment on a pull request on richlowe/illumos-gate
I'll leave it in for now - it's something I've found fairly useful.
r1mikey created a comment on a pull request on richlowe/illumos-gate
`cbe` and `gic` conversion to DDI drivers here: #110
r1mikey opened a draft pull request on richlowe/illumos-gate
Move cyclic backend to be a DDI driver, move GIC implementations to be DDI drivers
The cyclic backend and the GIC drivers are currently intimately aware of FDT via the raw prom interfaces. Move the cyclic backend to be a module, using the Arm Generic Timer device tree node to ...
r1mikey created a comment on a pull request on richlowe/illumos-gate
Oh, and I can only test on Rpi4 in about two weeks when I get back to the UK. Before integrating we could ask one of the other interested folks to test on Pi4.
r1mikey created a comment on a pull request on richlowe/illumos-gate
I'll clean up the assertions today and update this. I'm also rebasing some GIC DDI-usage changes - basically make cbe instantiate a proper generic timer driver to implement the cyclic backend, whi...
r1mikey created a review comment on a pull request on richlowe/illumos-gate
Yeah, I think that's a better assertion in the DDI implementation.