Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 348653 - [TRACKER] Make Gentoo/MIPS less bad
Summary: [TRACKER] Make Gentoo/MIPS less bad
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: New packages (show other bugs)
Hardware: MIPS Linux
: High major (vote)
Assignee: MIPS Porters
Keywords: Tracker
Depends on: mips-loongson 348647 348649 348650
  Show dependency tree
Reported: 2010-12-13 20:58 UTC by Matt Turner
Modified: 2022-02-19 01:15 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Note You need to log in before you can comment on or make changes to this bug.
Description Matt Turner gentoo-dev 2010-12-13 20:58:05 UTC
We've got some work to do.
Comment 1 Alexander 2011-12-16 15:13:50 UTC
any news here?
Comment 2 Matt Turner gentoo-dev 2011-12-16 16:28:25 UTC
(In reply to comment #1)
> any news here?

Yes, check the tracker bugs. New stages are built. Now we need to build new install media. If you'd like to help or test, please join #gentoo-mips on
Comment 3 Joshua Kinard gentoo-dev 2015-09-08 07:09:26 UTC
Things I could use some help on to make creating new install media easier:

- Semi-recent multilib stage3 for mips3 or mips4 big-endian.  I last had this working in 2013, but one of the major package moves killed the thing dead, and it's probably bitrotted too far to be salvageable in catalyst now.  I'll take any seed stage3 from late 2014 on up to base off of, if anyone's got one.

- uclibc-based stage3, mips1 through mips4, big-endian.  This would be for the netboots.  I already using musl to base a netboot off of, but musl doesn't play well with differing values for PAGE_SIZE.  I.e., if I build the netboot on a machine running PAGE_SIZE=64K, then 4K and 16K machines will have problems.  I suspect musl is hardcoding the PAGE_SIZE value somewhere.  I'd like to keep the netboots small, under ~15MB, which uclibc can help with.  If push comes to shove, though, I may just shoehorn the thing w/ glibc and go overboard on compression.

With those two puzzle pieces out of the way, and with working netboots, creating LiveCD media should be pretty easy...though arcload's multi-machine functions haven't been tested in a LONG time.

And then:

- Anyone with working IP27 (Origin 200/2000/Onyx2) hardware or IP28 (Indigo2 Impact) hardware to test kernels with.  Would like some confirmation that IP27's lockup bug isn't affecting only me, and for IP28, to get an idea of how long that machine can stay online before the spec exec issue drops it, and what the last action was before it died.