talk to warner (bsdimp) about the right way to do the kernel config stuff. the XXX_BASE stuff was me abstracting out config files in the atheros mips world before it got too bad. It's not the "right" way. Ask him what the right way is!
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Apr 20 2016
This uses the read/write macros in a novel way (register ids, not register addresses) but I think it's fine.
Update diffs. The only change now is changing '#include <fbsd-*>' to '/include/ "fbsd-*"' in the base SoC dtsi files.
Apr 19 2016
OK. I'll do this tomorrow.
I mean, can't we have the ofw probe routine check the normal device list, rather than compat data?
In D5993#127869, @adrian wrote:surely the flash DTS stuff could be modified to probe the flash_ident array, based on the names?
surely the flash DTS stuff could be modified to probe the flash_ident array, based on the names?
Choose correct repo
Apr 18 2016
Apr 17 2016
Yeah, I think we should break this out into something in sys/kern/ that we can optionally include where appropriate.
Apr 16 2016
Apr 15 2016
ooo!
Apr 13 2016
Apr 12 2016
Apr 11 2016
Two commits is fine.
It's probably better that I do this as 2 separate commits:
- just for the change to sys/mips/include/resource.h
- for sys/mips/mediatek/mtk_pcie.[ch]
remove std.mediatek from diff
Remove a leftover printf.
Apr 8 2016
Thanks!
Nothing in the docs (at least the ones I have seen) unfortunately... I hope something will pop up in the future and we'll be able to fix this...
I don't like the magic numbers in the PCI phy setup, but that can be fixed later. Do you have values for them anywhere?
I'll test with smaller delays in the future. This one seems to work for now and is as seen in "other places" :-)
yup, great!
Heh, that delay is pretty hilarious. :-P
Add proper controller initialization.
Order diff a bit better.