Rev 1.56 of if_dc.c removed calls to mii_pollstat() from the dc_tick()
routine. dc_tick() is called regularly to detect link up and link down
status, especially when autonegotiating.
The expectation was that mii_tick() (which is still called from dc_tick())
would update status information automatically in all cases where it would
be sensible to do so.
Unfortunately, with authentic 21143 chips this is not the case, and
the driver never successfully autonegotiates. This is because (despite
what it says in the 21143 manual) the chip always claims that link is not
present while the autonegotiation enable bit is set. Autonegotation takes
place and succeeds, but the driver tests the link bits before it switches
off the autonegotiation enable bit, and success is not recognised.
The simplest solution is to call dcphy_status() more often for MII_TICK
calls by dropping out of the switch statement instead of exiting when
we are autonegotiating and link appears to not be present. When
autonegotiation succeeds, dcphy_status() will note the speed and fdx/hdx
state and turn off the autonegotiation enable bit. The next call to
dcphy_status() will notice that link is present, and the dc driver code
will be notified.
Macronix chips also use this code, but implement link detection as
described in the manual, and hence don't need this patch. However, tests
on a Macronix 98715AEC-C show that it does not adversely affect them.
This could be done better but is the minimal effective change, and most
closely mimics what was happening prior to rev 1.56 of if_dc.c. (Actually
I also deleted a small amount of unnecessary code while I was in the area.)
Reviewed by: wpaul
- ifdefs around vendor sccsid were left reversed after switching to
__FBSDID().
- vertical whitespace after some of the non-FALLTHROUGH cases was lost.
- too much vertical whitespace before prototypes.
- don't use unusual indentation for 39 lines of declarations when only 2
of the lines benefit from it.
- don't use __DECONST(). This was the one use of it in the tree, and it
was just wrong. It was used to hide the warning about tgetnum() having
the wrong prototype (missing a `const') due to libncurses being
misconfigured. libncurses has been fixed, so the original code now
compiles cleanly with WARNS=4.
Added NOOBJ if anyone even attempts to "make obj" here.
Revert to installing files with mode 644 except README.
Make this overall look like a BSD-style Makefile rather
than roll-your-own (this is not a bug).
For the record. Previous revision also fixed the breakage
introduced by the sys.mk,v 1.60 commit: bsd.own.mk is no
longer automatically included from sys.mk.
Reported by: jhay
bsd.own.mk as of share/mk/sys.mk,v 1.60.
I did not notice this because I tested with DESTDIR=/foo/5.0,
and the "exists(/foo/5.0)" test apparently succeeded.
Reported by: fenner
of makefiles: bsd.own.mk was included before ../Makefile.inc,
effectively hiding the (wrong) LIBCOMPATDIR assignment here.
share/mk/sys.mk,v 1.60 (and assorted share/mk fixes) fixed
this order, revealed this bug, and broke "make release" and
"make installworld" with either of -DCOMPAT1X or -DCOMPAT2?.
Reported by: jhay
option (try to break at word bounaries) for SUSv3 conformance.
Partially based on the NetBSD version, with the following changes:
- style(9)
- break on <blank>s, not spaces, per POSIX (and GNU)
- when looking for last space on line, search backwards instead of
forwards; less comparisons needed this way.
- use LINE_MAX macro instead of a magic number and a comment saying it is
LINE_MAX.
PR: 36245
Reviewed by: mike
Obtained from: NetBSD (partially)
due to conditions that suggest the possible need for stack growth.
This has two beneficial effects: (1) we can
now remove calls to vm_map_growstack() from the MD trap handlers and (2)
simple page faults are faster because we no longer unnecessarily perform
vm_map_growstack() on every page fault.
o Remove vm_map_growstack() from the i386's trap_pfault().
o Remove the acquisition and release of Giant from i386's trap_pfault().
(vm_fault() still acquires it.)
- get rid of the 4th level headings
- seperate the actions required to prepare the installation (getting
the CD, settings up the various daemons, ...) from the actual
installation
Translation teams: This patch moves some paragraphs around and changes
a lot of <sectX> headers, but does not contain any content changes in
the strict sense of the word (i.e. I did not rewrite paragraphs yet).
There were no changes to the indentation, either.
Reviewed by: bmah, -sparc, -doc
to be emitted as per the C preprocessor. It updates the manual page in
regards to standards accordingly.
PR: standards/36075
Submitted by: tjr
Reviewed by: mike
MFC after: 1 week
loop is inversly proportional to hz.
This makes things more sane for configurations with hz != 100.
Cosmetic: Make the loops look similar to the loops in digi.c
loop is inversly proportional to hz.
This makes things more sane for configurations with hz > 100.
Submitted by: Peter Jeremy <peter.jeremy@alcatel.com.au>
machine_checks.
This fixes pci config reads for non existing devices on secondary
pci busses.
Thanks to Andrew Gallatin for pointing me to the register
Reviewed by: gallatin
Approved by: gallatin
to use ``.if defined()'' inside bsd.own.mk to test for defines
in individual makefiles. For example, setting DEBUG_FLAGS in
Makefile didn't take the desired effect on the STRIP assignment.
Added bsd.init.mk (like in NetBSD) that handles the inclusion
of ../Makefile.inc and bsd.own.mk from all bsd.*.mk files that
"build something".
Back out bsd.own.mk,v 1.15: moved OBJFORMAT initialization back
to sys.mk (several source tree makefiles want to check it early)
and removed MACHINE_ARCH initialization (it's hard to see from
looking at the commitlogs what the problem was at the time, but
now it serves no purpose).
Prohibit the direct inclusion of bsd.man.mk and bsd.libnames.mk.
Protect bsd.obj.mk from repetitive inclusion. Prohibiting the
direct inclusion of bsd.obj.mk might be a good idea too.
