hasn't published a list of its affected products, but it appears that Xen
was given this information since they have a specific detection code.
We could just unconditionally enable eager; but on x86_32 eager may have
a greater performance cost than lazy, and we don't want to lose
performance on unaffected (and ~old) CPUs running NetBSD/i386.
So use the same code as Xen: take Family 6, and whitelist certain models.
only if the fpu already has a state. We're fine with getting a DNA, what
we're not fine with is if the DNA is received while the FPU is busy.
I believe (even though I couldn't trigger it) that the panic would
otherwise fire if PT_SETFPREGS is used. And also ACPI sleep/wakeup,
probably.
- Get all of the drivers onto the new match quality constants.
- Introduce a new helper function, iic_use_direct_match(), that has
all of the logic for direct-config matching. If it returns true,
the driver returns the match result (which may be 0). If it returns
false, the driver does indirect-config matching.
- iic_compat_match() now returns a weighted match quality; matches to
lower-indexed "compatible" device property are more-specific matches,
and return a better match quality accordingly.
In addition to the above:
- Add support for direct-config matching this driver based on
"compatible" properties.
- Address-only matching is now done based on the specific addresses
the requested model supports.
- "compatible" property can specify to the driver which model is to
be used, so that using config "flags" directives aren't required
in the direct-config case.
XXX More changes coming that require re-factoring some other code.
- Get all of the drivers onto the new match quality constants.
- Introduce a new helper function, iic_use_direct_match(), that has
all of the logic for direct-config matching. If it returns true,
the driver returns the match result (which may be 0). If it returns
false, the driver does indirect-config matching.
- iic_compat_match() now returns a weighted match quality; matches to
lower-indexed "compatible" device property are more-specific matches,
and return a better match quality accordingly.
XXX This driver is an odd-ball with respect to the hardware device.
See comments in the match routine. Unclear how best to handle it.
- Get all of the drivers onto the new match quality constants.
- Introduce a new helper function, iic_use_direct_match(), that has
all of the logic for direct-config matching. If it returns true,
the driver returns the match result (which may be 0). If it returns
false, the driver does indirect-config matching.
- iic_compat_match() now returns a weighted match quality; matches to
lower-indexed "compatible" device property are more-specific matches,
and return a better match quality accordingly.
Replace unportable manual calculation of alignof() that causes UB, with
a GCC extension __alignof__.
This fixes a problem reported by UBSan with the MKSANITIZER distribution.
rk_getpwuid_r() returns a pointer pwd->pw_dir to a buffer pwbuf[].
It's not safe to store another a copy of pwd->pw_dir in outter scope and
use it out of the scope where there exists pwbuf[].
This fixes a problem reported by ASan under MKSANITIZER.
for several special PHY conditions, particularly:
- Properly re-initialise the PHY upon resume
- Store next page in the Link Partner Next Page register for compatibility
with 802.3ab on 88E3016 PHYs. Fixes some autonegotiation problems on msk(4)
- Make 88E3016 actually work
- Make sure page 0 is selected when we initialize the PHY. Fixes problems
with the eephy(4) that attaches to nfe(4) on machines like the Sun Ultra 40.
(we had condition for this, now the page 0 is selected for any PHY type)
- Disable fiber/copper auto-selection on the 88E1111 if it is in RGMII mode, to
work around the fact that the onboard PHYs attached to nfe(4) on the Sun
X4100 M2 have fiber/copper auto-selection enabled even though the interfaces
are clearly copper-only
make sure to also add appropriate licenses, since basically nothing
really significant remains out of previous code
use FreeBSD <dev/mii/e1000reg.h> for register definitions as a base instead
of OpenBSD <dev/mii/eephyreg.h>, since it has some extra definitions for some
3016 bits, but add the several extra bits from OpenBSD needed by the code;
removed no longed used <dev/mii/makphyreg.h>
tested with PHY 88E1111, there no particular change observed - the
link status works as it did before, just now it does media nego
even before the interface is up
should however fix 88E3016 support and hence PR kern/49270 and PR kern/53301
started. Otherwise it is possible that the first context switch on one of
the cpus will restore an invalid fpu state in the new lwp, if that lwp
had its fpu state stored on another cpu that didn't have time to do an
fpu save since eager-fpu was enabled.
Use barriers and all the related crap. The point is that we want to
ensure that no context switch occurs between [each fpu is cleared] and
[x86_fpu_eager is set to 'true'].
Also add KASSERTs.
equivalents; where possible left the original code intact to make it
easier to compare against FreeBSD original
made non-working stubs for some functions which don't have direct
equivalent yet - this includes all of <sys/buf_ring.h>, m_getjcl(),
m_append, and m_collapse(); these need to be adressed
left XXX/TODO in if_enavar.h, colocated with the non-working stubs
