I had noted a problem with choppy audio after the pci.c update of a week
ago; this turns out to be more general. I've also lost firewire and
the memory-stick slot (3rd usb controller) completely:
---
Copyright (c) 1992-2003 The FreeBSD
I had noted a problem with choppy audio after the pci.c update of a week
ago; this turns out to be more general. I've also lost firewire and
the memory-stick slot (3rd usb controller) completely:
---
Following up to my own note: With no visible change to dmesg, the
Mine is ICH3 with the same problem; this depends on a change sometime
between last Friday (worked fine) and this morning (broke) PST...
It looks as if someone made a change to interrupt routing sometime since
the changes to the fsstat struct (I had recompiled then and the sound
problem was NOT
ASUS P5A or P5AB (one of each now) - won't boot without setting
hw.ata.ata_dma=0. GEOM recognizes the drive itself but then the reads
of the disklabels fail with a message that I can't find in the source,
about a DMA failure. System then prompts to mount root and nothing I
type works. I see
Well, I misspoke on the timer; the one that works is booting in non-acpi
mode by beasties choice; the older one still fails if I comment the acpi
ignore timer statement. Actually that one appears to run slow by about half.
Both fail equally in ATAng so acpi isn't making any difference (nor does
Someone added -std=gnu89 -pedantic to libcrypto's make; given that
stdlib.h and several other places contain long long, this results in
several thousands of warnings about C89. We need to either go modern
or not but not try both at once :-)
Also usr.bin/window generates LOTS of warnings; I
This one stops the build:
Same cvsup (of about an hour ago) as for my previous note.
-- Pete
-
=== libexec/tcpd
cc -O -pipe -march=pentium3 -DREAL_DAEMON_DIR=\/usr/libexec\ -DSEVERITY=LOG_INFO
-DRFC931_TIMEOUT=10 -DHOSTS_DENY=\/etc/hosts.deny\
pedantic and Werror together cause problems again... I presume we really
need the quad type here. (or is this one due to a compiler upgrade?)
-- Pete
---
=== lib/csu/i386-elf
rm -f .depend
mkdep -f .depend -a
This may be just my infamous vaio acting up again, but since the
recent commit to fxp driver (Monday?) I get a panic on device probe
(page fault in kernel mode).
That and the way the pccbb act up (always return 0 for event and
status register reads, and don't reset pending interrupt on event reg
I don't know how system-specific this problem is, but:
Sony VAIO R505ES
Sound is Intel ICH3 + Yamaha.
This or something closely related has been happening for weeks.
Several times earlier this week and last week sound panic'd, and
also sometimes there was a panic (several different kinds) on
This is a Vaio R505ES. Looks as if acpi is both
totally necessary
and doesn't work right.
...
psm doesn't work (fails probe too). Complains about
unable to
allocate irq.
...
You might try this. I have a Sony Vaio GRX-670 and
the touch pad didn't work. Took me a while to track
This is a Vaio R505ES. Looks as if acpi is both totally necessary
and doesn't work right.
In my last 2 compiles (since about 12/20) I get a lock-order-reversal
panic during boot, after init starts. continue in ddb makes
things appear to work right. If this was fixed since 12/23 I
apologize for
During a make world, from a cvsup of about an
hour ago, this happened:
/usr/obj is symlinked to /d/obj-c:
cc -fpic -DPIC -O -pipe -march=pentium3 -D_IEEE_LIBM -D_ARCH_INDIRECT=i387_ -c
/usr/src/lib/msun/src/s_sinf.c -o s_sinf.So
cc -fpic -DPIC -O -pipe -march=pentium3 -D_IEEE_LIBM
Pardon sending this to both stable and current, but this is
relevant to both and the code paths may be different...
I just acquired a usb mass device which partly identifies.
This is a Dazzle multi-card reader. The CF slot works fine;
identifies as da0, and though it is slower than the cf pcmcia
At the moment, NO_WERROR has no effect on a kernel build; there is
a warning in cardbus.c:954 (easily fixed, but annoying...)
On another note, the VAIO R505ES problem is a little better (DVD/CDR is now
recognized and actually works). However, it panics if I try to mount
a DVD with UDF directory;
This is, for a change, not relative to my Vaio (which still won't boot
current, in spite of various improvements apparent in boot -v).
Hardware is a Supermicro P3TDDE (SMP P3-1G, via chipset).
2 60g drives, ad0 has a complete -stable on it, ad1 current.
First, GEOM:
On trying to do boot0cfg -s1
Replying to my own message:
---
2 60g drives, ad0 has a complete -stable on it, ad1 current.
First, GEOM:
On trying to do boot0cfg -s1 /dev/ad0 to get back to booting stable,
to try to recover from the next problem, I get operation not permitted.
This didn't happen
CVSUP as of last night about 10PM PST:
Well, phk's stuff now appears to work right (even finds the memory stick
most of the time) and the bad malloc in acpi is fixed. This seems to
leave irq issues (still!!)
PSM0 still can't allocate irq even though 12 isn't used anywhere else.
on the init
Well, now the bad free (actually appeared to me to be a bad malloc or a bad
call to malloc) is fixed; now I get a panic from GEOM (actually from
witness). There are 3 or 4 geom debug messages and then a panic
indicating Giant not owned. I'll follow up with an actual trace from a
serial
Following up on my own message:
Well, now the bad free (actually appeared to me to be a bad malloc or a bad
call to malloc) is fixed; now I get a panic from GEOM (actually from
witness). There are 3 or 4 geom debug messages and then a panic
indicating Giant not owned. I'll follow up with
Cross builds will be necessary for a goodly while; as on hardware like mine
(new VAIO laptop) where current won't boot at all without lots of fiddling,
for example...
I'm going to put an empty stdint.h in stable for now, just to see
what happens -- it delays the failure to the compile of
There is a 't' modifier to a format in bin/sh that just crept in;
it prevents a cross buildworld under stable without NO_WERROR.
Perhaps the compiler+libc needs to be built first? (and does sh
need to be a build tool; I'd hope the make scripts stuck to a
fairly least-common-denominator shell
With current as of today, I can finally get the system to
boot without acpi enabled, but not otherwise. However,
with all combinations of hw.pcic.intr_path={0,1,2},
hw.pcic.irq=0, and hw.pcic.init_routing={0,1} the TI
is not happy and I can't use the wireless.
Since this system is acpi-only
This showed up a few days ago; a tool-dependency problem
in a cross-build.
I build current under stable since this (acpi-only) system
won't yet boot current (mostly appears to be the problem with
TI pcic/cardbus chip interrupt routing)
The build of aicasm in the kernel mkdep works right; that
Well, now it boots (presumably thanks to IRQ fixes for TI bridges...)
Now it doesn't spot *any* devices on pc-card or the built-in wi0;
also the mouse gets lost (psm0: unable to allocate IRQ).
The Memory-stick does work (with a manual camcontrol rescan).
(there are plenty of complaints from
I got around the lack of IRQ on pcic1 by means of a trick involving
hw sets in loader.conf (thanks to the mobile contributors); now it
won't mount root.
I know what the problem is but not why; the loader brings the kernel in
fine, and probes work fine; I'd presume if the partition were too
This may relate to the X-window reboots - my machine shows that too, if
I boot without acpi; with acpi X never gets going.
Supermicro P3TDDE, 2x1ghz P3:
Via Apollo Pro 266 or 266T chipset, generic Nvidia TNT2 display.
(there is a conflict in the manual which says 266T but 8633 north bridge).
I
I'm doing a cross-build of current on a system running stable, with a
separate drive for current mounted as /current (with all sub-mounts correct).
Separate problem causing me to cross-build: current doesn't like to boot
lately - it hangs immediately after loading acpi.ko (on the SMP Supermicro;
Supermicro P3TDDE (I think; may be a TDDA, but it does have the promise
chip).
This board is Via-based and not Intel. Could easily be my problem :-(
Don't know about the APIC.
Stable runs just fine on the same motherboard (dual-boot, stable on
ad0 and current on ad1) including reboot (see
I have a system using a fairly new Supermicro MB, with 2 P3-1GHZ, and 512mb
ram. Running stable works fine at least a day or so with LOTS of activity.
Running current it hangs (with no output of any kind, and apparently all
interrupts disabled) so DDB does me no good... This requires a fair
I don't think this is just my system, and no binutils changes have been made
recently, (and the disk is NOT full, anywhere near), and noone else has
complained about a related build problem either:
--
cd
The new acpi version apparently fixed my panic (I didn't change any
other configs, and things now boot, apparently correctly, on the A7V
board.) (and it keeps time right :-)
--
Copyright (c) 1992-2001 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991,
Is there a way to set a loader env from a file? (I presume that is part
of what prompted the rather funny quasi-flame-war about loader interpreter
base. Lisp indeed :-) Actually I remember Jordan (and at least one more
who is now in the fbsd group; who?) getting into the forth loader
On both of my -current systems, I can't remotely display X apps back
to my (non-current) laptop. I don't know if this is related to the
upgrade in ssh (my suspicion) or some other (likely library) issue.
One of them is running X 4.1.0 downloaded from xfree86.org; the other
3.3.6, so the
Terry Lambert wrote:
unknown: PNP0400 can't assign resources
unknown: PNP0400 at port 0x378-0x37f on isa0
unknown: PNP0501 can't assign resources
unknown: PNP0501 at port 0x3f8-0x3ff on isa0
unknown: PNP0501 can't assign resources
unknown: PNP0501 at port 0x2f8-0x2ff on isa0
K6-2-450, bus running at 95mhz, Acer 1541 (A? B?)
All works fine with the new ACPI _except_ the clock; the time of day
advances about twice as fast as it should, and I get LOTS of
calcru negative time and time went backwards messages.
NTP is not capable of correcting this gross rate error :-(
I have a panic that I haven't seen in the notes so far; if I allow load
of the acpi module, it gets a can't allocate memory very early in the
boot process.
Hopefully the following is complete - KDE2 cut-and-paste didn't work
as I expected for a big region.
Motherboard is ASUS A7V with 1.2G
I got a panic today on a fresh kernel...
Compiled with netgraph but non of the netgraph modules.
Immediately after the memory probe, a message about sequencers 0-15,
then:
Panic: spinlock ng_worklist not in order list
Before these was WARNING: size of kinfo_proc (648) should be 644!!!
-- Pete
This may relate to a commit about noon (PST) today fixing a different
problem. I'm just waiting it out :-)
Welcome to "current" where (especially lately) about half the time things
don't 'make'... (I'm trying to recompile my kernel after recovering
from the libc circus, trying to prevent
There is also a new access point (either just available or "RSN") from
Zyxel (316); it is a combination of a 310 (cable modem/bridged DSL/PPPOE
router) and single-card bridged access point.
I'm using one at work (overkill since I'm not using the router) as a
bridged access point; it works just
Immediate problem is fixed by including machine/param.h in netstat/if.c.
ifmcstat, rip6query, rtadvd/dump.c, i4b/isdnd/rc_config.c too...
Those appear to be all.
I don't know the "canonical" solution;
maybe including machine/param.h in if_var.h? (or was removing it
for "cleanliness" the
Probably not many -current users use msdosfs; I use it only for the
occasional floppy (extremely rare now) and the flash cards from my
electronic cameras. On trying to mount one of these today, I ran into
a problem (apparently in the msdosfs_vnops change of Sept 2, since a system
made the
cvsup this morning; there was a patch out for this a week ago?
---
cc -c -O -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes
-Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -fformat-extensions -ansi
Somehow in the internat merge the kerb4 versions of rshd and rlogind
disappeared (rshd -k; rlogind -ek and -k). They are often used for
scripting in ways that kerberized telnet can't handle (yes, I can
use expect but what a kluge :-), or I could adapt ssh for this and
probably will, but that
We should switch to using just libdescrypt and being allowed to switch
crypt formats easily between md5 and des. My proposed solution using
login.conf is at http://people.FreeBSD.org/~green/crypt_switching.patch,
and it's going to be put into production usage relatively soon (that is,
There are many cases (e.g. apache with external sources doing crypt)
where it is necessary to have libscrypt the default crypt for logins,
etc, and libdescrypt available for linking (e.g. apache). Since the
realignment of crypt sources to the internat versions (which I generally
like), this is
I notice that though ifconfig does a kld as appropriate, pccardd doesn't
at least as of a week ago. Still looks like it isn't in the source.
1. Is this in the works from one of the normal maintainers? If not I
might take a look at fixing it up over the next week or so.
-- Pete
To
47 matches
Mail list logo