Details of sys-kernel/mips-sources:

${m_name} Support has been marked as needing testing in this
ebuild revision.  This usually means that any patches to support
${m_name} have been forward ported and maybe even compile-tested,
but not yet booted on real hardware, possibly due to a lack of access
to such hardware.  If you happen to boot this kernel and have no
problems at all, then please inform the maintainer.  Otherwise, if
experience a bug, an oops/panic, or some other oddity, then please
file a bug at bugs.gentoo.org, and assign it to the mips team.
IP27 Origin 2k/Onyx2 systems may be prone to sudden hard lockups.
The exact trigger is unknown at this time.
Things that might work, but have problems, or are unknown:
\t- Serial support on the Octane uses a very basic UART driver that drives
\t\040\040the 16550A chip on the IOC3 directly.  It does not use interrupts,
\t\040\040only a polling routine on a timer, which makes it slow and CPU-
\t\040\040intensive.  The baud rate is limited to no more than 38.4kbps on
\t\040\040this driver.  Patches for getting the Altix IOC3 serial driver to
\t\040\040work (which uses DMA and supports faster baud rates) are welcome.
\t- MENET boards appear to have the four ethernet ports detected, however
\t\040\040the six serial ports don't appear to get picked up by the IOC3
\t\040\040UART driver.  The NIC part number is also not read correctly
\t\040\040from the four Number-In-a-Cans.
\t- Other XIO-based devices, like various Impact addons, remain untested
\t\040\040and are not guaranteed to work.
\t- Upon rebooting, Octane may not return to the PROM.  Holding down the
\t\040\040reset button for up to ten seconds, releasing, then pressing it
\t\040\040again for one or two seconds may trigger the system to return
\t\040\040to the PROM.  Please wait at least five seconds after each attempt
\t\040\040to see if the machine responds.  As a last resort, unplugging the
\t\040\040power cord for 30 seconds will also work, but be aware the power
\t\040\040supplies on these systems will not appreciate this due to their
\t\040\040relative old age.

${m_name} Support has been marked as needing testing in this
ebuild revision.  This usually means that any patches to support
${m_name} have been forward ported and maybe even compile-tested,
but not yet booted on real hardware, possibly due to a lack of access
to such hardware.  If you happen to boot this kernel and have no
problems at all, then please inform the maintainer.  Otherwise, if
experience a bug, an oops/panic, or some other oddity, then please
file a bug at bugs.gentoo.org, and assign it to the mips team.
IP27 Origin 2k/Onyx2 systems may be prone to sudden hard lockups.
The exact trigger is unknown at this time.
Things that might work, but have problems, or are unknown:
\t- CONFIG_TRANSPARENT_HUGEPAGE should work now, but there may still be
\t\040\040intermittent issues.  Additionally, CONFIG_HUGETLBFS must also be
\t\040\040selected for hugepages to work.  If use of this feature continues
\t\040\040to trigger random Instruction Bus Errors (IBEs), then it is best to
\t\040\040disable the functionality and perform a cold reset of the machine
\t\040\040after powering it down for at least 30 seconds.
\t- Serial support on the Octane uses a very basic UART driver that drives
\t\040\040the 16550A chip on the IOC3 directly.  It does not use interrupts,
\t\040\040only a polling routine on a timer, which makes it slow and CPU-
\t\040\040intensive.  The baud rate is limited to no more than 38.4kbps on
\t\040\040this driver.  Patches for getting the Altix IOC3 serial driver to
\t\040\040work (which uses DMA and supports faster baud rates) are welcome.
\t- UHCI Cards are known to have issues, but should still function.
\t\040\040This issue primarily manifests itself when using pl2303 USB->Serial
\t\040\040adapters.
\t- MENET boards appear to have the four ethernet ports detected, however
\t\040\040the six serial ports don't appear to get picked up by the IOC3
\t\040\040UART driver.  The NIC part number is also not read correctly
\t\040\040from the four Number-In-a-Cans.  Additional testing would be
\t\040\040appreciated and patches welcome.
\t- Other XIO-based devices, like various Impact addons, remain untested
\t\040\040and are not guaranteed to work.  This applies to various digital
\t\040\040video conversion boards as well.