On Mon, Oct 31, 2022 at 07:45:29AM +, Miod Vallat wrote:
> > +.if (${COMPILER_VERSION:L} != "clang" && ! exists(/usr/local/bin/eg++))
> > +regress:
> > + @echo 'Run "pkg_add g++" to run unittests on GCC architectures'
> > + @echo SKIPPED
>
> Or the C++ test could be downgraded to C++98 so
size(1) output is not padded and uses a single tab between each column
which often looks ugly:
$ size bsd /usr/lib/libc.so.96.4 `which echo cc`
textdatabss dec hex
896234 31208 54664 982106 efc5a /usr/lib/libc.so.96.4
106152 11200 12640
Makefile bits adapted from regress/lib/libssl/interop/botan/Makefile.
required for stderr
required for exit(3)
required for memcmp(3)
g++-8.4.0p12 on sparc64 prints coloured warnings/errors, but that is
enother issue.
On sparc64, `make CXX=clang++' without patch works as well.
A follow-up
On Fri, Oct 28, 2022 at 11:08:55AM +0200, Martin Pieuchot wrote:
> On 20/10/22(Thu) 16:17, Martin Pieuchot wrote:
> > On 11/09/22(Sun) 12:26, Martin Pieuchot wrote:
> > > Diff below adds a minimalist set of assertions to ensure proper locks
> > > are held in uvm_mapanon() and uvm_unmap_remove()
On Sun, Oct 30, 2022 at 01:58:40PM +0100, Martin Pieuchot wrote:
> On 30/10/22(Sun) 12:45, Klemens Nanni wrote:
> > On Sun, Oct 30, 2022 at 12:40:02PM +, Klemens Nanni wrote:
> > > regress on i386/GENERIC.MP+WITNESS with this diff shows
> >
> > Another one; Th
On Sun, Oct 30, 2022 at 12:40:02PM +, Klemens Nanni wrote:
> regress on i386/GENERIC.MP+WITNESS with this diff shows
Another one; This machine has three read-only NFS mounts, but none of
them are used during builds or regress.
This one is most certainly from the NFS regress tests themsel
On Fri, Oct 28, 2022 at 11:08:55AM +0200, Martin Pieuchot wrote:
> On 20/10/22(Thu) 16:17, Martin Pieuchot wrote:
> > On 11/09/22(Sun) 12:26, Martin Pieuchot wrote:
> > > Diff below adds a minimalist set of assertions to ensure proper locks
> > > are held in uvm_mapanon() and uvm_unmap_remove()
Only five legacy half-duplex hardware drivers require this function to
change between playing and recording:
i386: ess(4), gus(4), pas(4), sb(4)
luna88k: nec86(4)
If defined, it is always called early in audio_open(), so just move the
call from audio(4) to each hardware driver's
On Sat, Oct 29, 2022 at 09:49:56PM +, Klemens Nanni wrote:
> Physical interfaces may at/detach at runtime but ifconfig(8) cannot
> create them like virtual ones, yet netstart(8) always tries to do so:
>
> # sh /etc/netstart -n em0
> { ifconfig em0 || ifco
Wrong diff, failed to juggle the netstart hunks, new one soon.
Physical interfaces may at/detach at runtime but ifconfig(8) cannot
create them like virtual ones, yet netstart(8) always tries to do so:
# sh /etc/netstart -n em0
{ ifconfig em0 || ifconfig em0 create; }
ifconfig em0 up
vifscreate() is used to create all virtual
After converting all drivers to play/record/duplex checks in their
open() handler, those bits are no longer used and can go.
OK once the other diffs are in?
---
share/man/man9/audio.9 | 5 -
sys/dev/audio.c| 9 -
sys/dev/audio_if.h | 7 ---
3 files changed, 21
On Thu, Oct 27, 2022 at 03:51:45PM +0200, Alexandre Ratchov wrote:
> IMO, for play-only drivers, the check should be:
>
> if (flags & FREAD)
> return ENXIO;
>
> which will reject record-only and full-duplex.
My mistake when typing things down: those are drivers not
On Thu, Oct 27, 2022 at 03:51:05PM +0200, Alexandre Ratchov wrote:
> On Thu, Oct 27, 2022 at 01:08:57PM +0000, Klemens Nanni wrote:
> > @@ -1040,6 +1041,9 @@ ad1848_open(void *addr, int flags)
> >
> > DPRINTF(("ad1848_open: sc=%p\n", sc));
> >
> &g
On Thu, Oct 27, 2022 at 03:49:55PM +0200, Alexandre Ratchov wrote:
> On Thu, Oct 27, 2022 at 01:09:57PM +0000, Klemens Nanni wrote:
> > @@ -1859,6 +1857,9 @@ utvfu_audio_open(void *v, int flags)
> > if (usbd_is_dying(sc->sc_udev))
> > return (EIO);
27 Oct 2022 17:31:45 Miod Vallat :
> Wait, this is wrong. All your
>
> (flags & (FWRITE | FREAD))
>
> comparisons must be
>
> ((flags & (FWRITE | FREAD)) == (FWRITE | FREAD))
>
> in order to truly catch full-duplex opens. Otherwise the condition will
> always be satisfied since one
Make drivers which do *not* adverise AUDIO_PROP_FULLDPLEX return ENXIO
in their open() if full-duplex mode was requested.
This way, sys/dev/audio.c:audio_open() will fail immediately rather than
later through the to-be-removed get_props() check.
These are all drivers which simply don't support
Make drivers which do *not* adverise AUDIO_PROP_FULLDPLEX return ENXIO
in their open() if full-duplex mode was requested.
This way, sys/dev/audio.c:audio_open() will fail immediately rather than
later through the to-be-removed get_props() check.
utvfu(4) seems to be the only driver that
Make drivers which do *not* adverise AUDIO_PROP_FULLDPLEX return ENXIO
in their open() if full-duplex mode was requested.
This way, sys/dev/audio.c:audio_open() will fail immediately rather than
later through the to-be-removed get_props() check.
This is the first round for drivers with logic in
The kernel lock was added back when there was a single sofnet task queue:
sys/net/nd6.c revision 1.213
date: 2017/08/06 12:53:30; author: mpi; state: Exp; lines: +15 -21;
Reduce contention on the NET_LOCK() by moving the nd6 address expiration
task to the
On Wed, Oct 26, 2022 at 11:30:20AM +0200, Uwe Werler wrote:
> Hi all,
>
> when running ksh in vi-mode an empty line could be stored in history:
>
> Steps to reproduce in command mode:
>
> 1. go history back e.g.: k
> 2. start editor with: v
> 3. delete the whole line and :wq
> 4. check history
There is no point in waiting when only printing would-be commands.
OK?
I'll promptly update this diff or the "netstart: improve shell style"
one in case either lands since they both touch the same line.
Index: netstart
===
RCS
On Mon, Oct 24, 2022 at 07:01:02PM +, Klemens Nanni wrote:
> IPv6LL no longer uses SOII (ifconfig(8) `soii' description got fixed).
Relevant commit:
sys/netinet6/in6_ifattach.c revision 1.114
date: 2019/08/21 15:32:18; author: florian; state: Exp; lines: +2
On Wed, Oct 05, 2022 at 10:31:38AM +, Klemens Nanni wrote:
> On Sat, Sep 10, 2022 at 02:10:22AM +0000, Klemens Nanni wrote:
> > It does not have the prettiest signature, but nicely folds identical
> > copies into MI softraid.c, which then allows us to
> > - avoid furt
The property bits of audio(9) are obsolete and ought to be removed
completely.
sys/dev/audio.c:audio_open() currently uses get_props() to bail out if
read *and* write was requested on a driver non-duplex driver.
Drivers that currently support playing but not recording need a little
adjustment
On Tue, Oct 25, 2022 at 11:07:33AM +0100, Jason McIntyre wrote:
> i'm ok with this. however it is really *a* command script, rather than *the*
> command script.
Sure, i fixed my local diff.
The mixed use of upper and lower case variables is neither obvious nor
consistent.
PRINT_ONLY is local to netstart.
ip6kernel is local to netstart.
multicast gets sources from rc.subr(8).
1. uppercase ip6kernel as is common for global variables in base scripts
2. use the simpler true/false idiom
- just call it (a sh(1)) script, in line with MAKEDEV(8) and rc.d(8)
- use only .Nm thereafter instead of .Nm/the .Nm script/...
- zap the additional rc.conf(8) bits for they can be found in there
- zap unhelpful "(or can be)"
Feedback? Objection? OK?
Index: netstart.8
If there is no default route but some interface has AUTOCONF, printing
what would be done still waits for... nothing to happen.
OK?
Index: netstart
===
RCS file: /cvs/src/etc/netstart,v
retrieving revision 1.220
diff -u -p -r1.220
IPv6LL no longer uses SOII (ifconfig(8) `soii' description got fixed).
Don't ignore this step in dry-runs. In analogy to how we print
{ ifconfig trunk0 || ifconfig trunk0 create; }
the file check could be deferred and printed as well, i.e.
[[ -f /etc/soii.key ]] && sysctl ...
On Fri, Oct 21, 2022 at 06:58:52PM +, Klemens Nanni wrote:
> On Fri, Oct 21, 2022 at 06:39:59PM +0100, Jason McIntyre wrote:
> > i was going to incliude this in my answer, but eventually thought it
> > looks like a seperate diff:
> >
> > the text currently read
On Wed, Oct 12, 2022 at 05:55:10PM +, Klemens Nanni wrote:
> sets up CLEANFILES, build/install targets and more, so no
> need to roll them ourselves if we just set PROG and SRCS.
>
> STRIPFLAG is defined but not used and in INSTALL_STRIP is
> the magic word, so fix that.
>
On Fri, Oct 21, 2022 at 03:50:41PM +0200, Solene Rapenne wrote:
> hi
>
> I found a weird behavior trying to use timeout(1) in a script with
> interactive commands. My use case is to ask for unlocking /home
> from /etc/rc.local but continue if I don't type it in time.
>
> this is working:
>
>
On Sat, Oct 22, 2022 at 02:09:40PM -0600, Todd C. Miller wrote:
> The cat_opts and decomp_opts fields in struct compressor are unused
> so just remove them. Also mark functions unused outside of main.c
> as static.
OK kn
vmt(4) reads MAC and IP addresses periodically through a task on the
kernel locked system task queue `systq'.
vmt_nicinfo_task() grabs the exclusive net lock which protects the
interface list and addresses, but it only reads and thus may grab the
shared net lock.
Again, not a big difference due
Logging the presence of a keydisk the same way offline data chunks are
reported seems unjustified.
Noting offline chunks serves as warning, so the user can bring them
online and update bootblocks.
Keydisks, however, are not to be touched by installboot(8) in any case
and whether they're
On Sat, Oct 22, 2022 at 04:11:25PM +0200, Theo Buehler wrote:
> > I don't think we have another script in base that is meant to be run
> > manually, so netstart(8) stands out.
>
> MAKEDEV
Right, what I meant was "to be run manually **as argument to sh(1)**".
I keep doing `cd /dev; sh MAKEDEV
On Sat, Oct 22, 2022 at 07:24:21AM +0100, Jason McIntyre wrote:
> On Fri, Oct 21, 2022 at 06:58:52PM +0000, Klemens Nanni wrote:
> > But now that I see your diff, I think we should actually include `sh'
> > in the synopsis since /etc/netstart is not executable, so the documente
On Fri, Oct 21, 2022 at 10:55:57PM +0200, Mark Kettenis wrote:
> Be careful. By moving more stuff into .rodata, you may overflow the
> .text/.rodata block. Make sure you build and test the kernels and
> also test an actual bsd.rd.
Yes, I keep testing bsd and bsd.rd/miniroot on amd64 and
On Fri, Oct 21, 2022 at 11:29:32AM -0900, Philip Guenther wrote:
> On Fri, Oct 21, 2022 at 11:04 AM Klemens Nanni wrote:
>
> > --- a/sys/arch/sparc64/sparc64/db_disasm.c
> > +++ b/sys/arch/sparc64/sparc64/db_disasm.c
> >
> ...
>
> > @@ -877,7
This is the second biggest read-write global in a sparc64 bsd.mp:
01c091b8 g O .data 2cb8 sparc_i
ddb(4) does not ever write to it and happily decodes instructions after
moving it to .rodata:
ddb{3}> x /i %pc
db_inst_branch+0x10:be,pn
On Fri, Oct 21, 2022 at 06:39:59PM +0100, Jason McIntyre wrote:
> i was going to incliude this in my answer, but eventually thought it
> looks like a seperate diff:
>
> the text currently reads as though only a singular interface is
> specified, and -n is kind of hidden. so i reworked the text. i
As usual, a global that's only ever read during match and attach.
The device type pointer gets stored in the softc, so that member must be
const as well.
Builds on amd64 but I don't have Apple notebooks to run-test:
820dd550 l O .rodata0410 ubcmtp_devices
Tests?
These are the compile-time constant strings behind vmstat(1)'s "malloc",
the kern.malloc.kmemnames sysctl(8) and panic strings.
Boots and displays unchanged on amd64:
82068690 g O .rodata0490 memname
OK?
Index: kern/kern_malloc.c
Not that big, but quirks are likely to grow in the feature and there is
no reason to keep them writable:
822c0750 g O .data 0160 uvideo_devs
Builds fine on amd64 but I don't have uvideo(4) attaching anywhere.
OK?
diff --git a/sys/dev/usb/uvideo.c b/sys/dev/usb/uvideo.c
On Fri, Oct 21, 2022 at 03:02:33PM +0300, Vitaliy Makkoveev wrote:
> netlock could be completely dropped here.
We could probably drop the net lock around ifnetlist wherever the kernel
lock is currently held, but that would just make it harder to reason
about ifnetlist locking, imho.
Eventually,
All interface ioctls always run with the kernel lock anyway, so this
doesn't make a difference, except that it reflects how ifnetlist is not
modified.
OK?
diff --git a/sys/net/if_loop.c b/sys/net/if_loop.c
index a15820420ef..16006c4c102 100644
--- a/sys/net/if_loop.c
+++ b/sys/net/if_loop.c
@@
Using -n does not require an interface:
# sh /etc/netstart -n ; echo $?
ifconfig lo0 inet 127.0.0.1/8
route -qn add -inet6 fe80:: -prefixlen 10 ::1 -reject
...
0
OK?
Index: etc/netstart
===
Another global that can go to .rodata:
8227fec0 g O .data 0420 acpisbs_battery_checks
Builds fine on amd64, but I don't have access to this hardware.
Tests? OK?
diff --git a/sys/dev/acpi/acpisbs.c b/sys/dev/acpi/acpisbs.c
index 7fd85860f79..252588da591 100644
---
On Thu, Oct 20, 2022 at 09:05:42AM -0700, Heppler, J. Scott wrote:
> Using the miniroot img from https://cdn.openbsd.org/pub/OpenBSD/7.2, the
> set retrival still points to snapshots.
Can't confirm, default install points at 7.2/ for me:
$ ftp
Looking at users of the global interface list `ifnetlist', drivers like
this read out an IP or MAC address to report to the hypervisor.
None of this modifies any of the fields in the global list, they
iterate over and read from it.
net/if_var.h documents the `if_list' member as kernel lock
All consumers now use C99 struct init and none of them sets `.setfd'.
My X230 still plays audio just fine:
azalia0 at pci0 dev 27 function 0 "Intel 7 Series HD Audio" rev 0x04: msi
azalia0: codecs: Realtek ALC269, Intel/0x2806, using Realtek ALC269
audio0 at azalia0
Feedback? OK?
diff --git
Those .data symbols are already treated as const, so move to .rodata:
82290610 g O .data 05d0 ahc_pci_ident_table
82290be0 g O .data 0198 ahd_pci_ident_table
650 const struct ahc_pci_identity *
651 ahc_find_pci_device(pcireg_t id, pcireg_t subid,
Looking at symbols in /bsd .data again for .rodata candidates:
82281b70 l O .data 1688 asmc_prods
amsc(4) never writes, but a const table requires a const softc member:
263 for (i = 0; asmc_prods[i].pr_name && !sc->sc_prod; i++)
264 if
On Tue, Oct 11, 2022 at 05:43:21PM +, Klemens Nanni wrote:
> sparc64 RAMDISK already has NFSCLIENT enabled and miniroot's FSSIZE
> still fits after linking mount_nfs:
>
> /dev/vnd0a6312 5592 72089% 3 59
> 5% /mnt
>
> Th
On Wed, Aug 31, 2022 at 05:16:27PM +0200, Uwe Werler wrote:
> Hi folks,
>
> I'm just wondering why there's a test for an empty line and args to start the
> editor for line editing. In bash one can start vi immediately with an empty
> line by ^v. Might be there's another reason I don't understand
Used only for lookups; builds on i386.
OK?
Index: usb/uaudio.c
===
RCS file: /cvs/src/sys/dev/usb/uaudio.c,v
retrieving revision 1.169
diff -u -p -r1.169 uaudio.c
--- usb/uaudio.c21 Mar 2022 19:22:42 - 1.169
+++
This is the only "implementation" of setfd() throughout the tree.
With this gone and the audio_hw_if C99 struct init for most drivers,
we can safely remove setfd() code from audio(9).
OK?
diff --git a/sys/arch/luna88k/cbus/nec86.c b/sys/arch/luna88k/cbus/nec86.c
index 2d2d965d1bc..ea1e8b8f6df
On Tue, Oct 18, 2022 at 02:17:54PM +, Klemens Nanni wrote:
> audio(9) cleanup demands removing a member from this struct, which is
> cumbersome in our current tree as drivers initialise it inconsistently,
> i.e. a simple removal of ".member_name = driver_func," is no
The only driver with a non-const struct.
Seen while looking at it for audio(9).
OK?
diff --git a/sys/arch/arm64/dev/aplmca.c b/sys/arch/arm64/dev/aplmca.c
index 07a53103be4..03ade86fee7 100644
--- a/sys/arch/arm64/dev/aplmca.c
+++ b/sys/arch/arm64/dev/aplmca.c
@@ -135,7 +135,7 @@ int
Each only used once for a printf() call in *_attach().
Seen while tweaking their *_hw_if struct.
OK?
Index: dev/isa/ess.c
===
RCS file: /cvs/src/sys/dev/isa/ess.c,v
retrieving revision 1.28
diff -u -p -r1.28 ess.c
--- dev/isa/ess.c
audio(9) cleanup demands removing a member from this struct, which is
cumbersome in our current tree as drivers initialise it inconsistently,
i.e. a simple removal of ".member_name = driver_func," is not always
possible.
Most drivers call their driver_*() functions like the member name,
- some
On Tue, Oct 18, 2022 at 09:32:27AM +0200, Alexandre Ratchov wrote:
> On Mon, Oct 17, 2022 at 07:45:33PM +0000, Klemens Nanni wrote:
> > I was looking at one particular driver and needed to know what those
> > flags are, turns out AUDIO_PROP_FULLDUPLEX is the only au
On Wed, Oct 05, 2022 at 12:10:26AM +, Klemens Nanni wrote:
> On Sun, Sep 25, 2022 at 12:59:27AM +0000, Klemens Nanni wrote:
> > On a sparc64 T4-2 I boot OpenBSD from hardware RAID 1E using the
> > built-in controller:
> >
> > mpii0 at pci15 dev 0 function 0
On Mon, Oct 17, 2022 at 09:53:26PM +0200, Mark Kettenis wrote:
> > Date: Mon, 17 Oct 2022 19:45:33 +
> > From: Klemens Nanni
> >
> > I was looking at one particular driver and needed to know what those
> > flags are, turns out AUDIO_PROP_FULLDUPLEX
I was looking at one particular driver and needed to know what those
flags are, turns out AUDIO_PROP_FULLDUPLEX is the only audio(9) property
that is in use.
The other two are only ever set in *_get_props(); their handling
disappeared from sys/dev/audio.c in
commit
On Sat, Oct 08, 2022 at 07:25:26PM +, Klemens Nanni wrote:
> If /usr is mounted read-only, kernel relinking fails silently without
> any log trace.
>
> # ksh /usr/libexec/reorder_kernel
> /usr/libexec/reorder_kernel[35]: cannot create
> /usr/share/relink
The installboot regress tests use -p on octeon which works fine.
I don't have serial console access to octen machines at the moment, but
this should just fly as installboot already has -p support on octeon.
It's also easy to test with snapshot media as one can easily apply this
patch in the
On Fri, Oct 14, 2022 at 02:11:19PM -0400, Okan Demirmen wrote:
> the below seems to work: Walter (thanks!) found the spot i missed on
> replacing height with ascent+descent.
>
> please let me know.
Works just fine for me, no more visual glitches, thanks!
On Thu, Oct 06, 2022 at 05:22:17PM +, Klemens Nanni wrote:
> On Tue, Sep 06, 2022 at 03:11:45PM +0000, Klemens Nanni wrote:
> > On rare occasions, I need 'disable xxx' in /etc/bsd.re-config to be able to
> > boot a system, e.g. to ignore quirky devices crashing driver
Two more, then all daemons in my accumulated `rcctl ls on' output should
be covered.
OK?
Index: smtpd
===
RCS file: /cvs/src/etc/rc.d/smtpd,v
retrieving revision 1.7
diff -u -p -r1.7 smtpd
--- smtpd 11 Jan 2018 19:52:12 -
Two more easy cases.
rad(8) requires a config file and there is no default /etc/rad.conf, so
configtest will fail unless a valid config (behind -f in rad_flags) exists.
dhcpleased(8) is happy without config at all, so configtest is always happy
unless a bogus config exists.
OK?
Index:
On Thu, Oct 13, 2022 at 04:39:04PM +0200, Omar Polo wrote:
> On 2022/10/13 13:00:34 +0000, Klemens Nanni wrote:
> > On Thu, Oct 13, 2022 at 08:28:50AM -0400, Okan Demirmen wrote:
> > > And I keep missing it! I can't reproduce this - can you share the font
> &
On Thu, Oct 13, 2022 at 08:28:50AM -0400, Okan Demirmen wrote:
> And I keep missing it! I can't reproduce this - can you share the font
> you're using maybe?
Whatever is the default, I never fiddled with fonts in X, no xorg.conf,
`cwm -c/dev/null' shows the glitch for me on a ThinkPad X230.
On Thu, Oct 13, 2022 at 10:11:39PM +1000, Jonathan Matthew wrote:
> While working on ypconnect(2), Theo suggested that ypldap(8) should
> not default to plaintext LDAP connections, since the data it's dealing with
> is pretty important to the security of the system. Here's a straightforward
>
This has annoyed me for a long time...
1. have at least two windows
2. open the window menu with M-slash (alt+/)
3. show all windows with C-A (ctrl+a)
4. move cursor over list from top to bottom,
observe no glitch
5. move cursor over list from bottom to top,
observe "[" and "]" chars in
Then there's also bootblk.text: not required by bootblk, not installed
and only used with Sun's fakeboot in a target that cannot work:
$ make -n fake
([ -h machine ] || ln -s
/home/kn/src/sys/arch/sparc64/stand/bootblk/../../include machine)
../fakeboot/fakeboot -elf32
sets up CLEANFILES, build/install targets and more, so no
need to roll them ourselves if we just set PROG and SRCS.
STRIPFLAG is defined but not used and in INSTALL_STRIP is
the magic word, so fix that.
Don't replace "bootblk" wth "${PROG}" to avoid churn.
No object change, `make' does
On Sat, Nov 13, 2021 at 10:09:27AM +, Klemens Nanni wrote:
> On Sat, Dec 26, 2020 at 05:19:55PM +0100, Klemens Nanni wrote:
> > Our make(1) is behind NetBSD's and FreeBSD's make(1) on at least the
> > rules of variable substitution.
> >
> > Our DESCRIPION says
&g
There since the powerpc -> macppc move/rename.
KERN_AS usage disappeared in
commit 5b7db11d478192c5908038bb1345e7d51cc35c8e
Author: rahnds
Date: Mon May 5 16:47:15 1997 +
only build one version of the libraries.
REAL_VIRT usage disappeared in
.../ofdev.c:130:1: warning: non-void function does not return a value
[-Wreturn-type]
OK?
Index: sys/arch/macppc/stand/ofdev.c
===
RCS file: /cvs/src/sys/arch/macppc/stand/ofdev.c,v
retrieving revision 1.27
diff -u -p -r1.27
12 Oct 2022 02:22:15 Todd C. Miller :
> This problem is not confined to sparc64 is it? At least macppc
> seems to have the same issue.
Looks like it, I have not yet tested other architectures.
> Some, like amd64 and i386, cheat and
> just pull in the 4 source files from the libz directory.
Testing the libz size fix in ofwboot.net showed that rebuilding ofwboot
after updating libz sources did not relink ofwboot as it should, imho:
$ cd sys/arch/sparc64/stand
$ make
...
$ touch ../../../lib/libz/crc32.c
$ make
===> bootblk
===>
sparc64 RAMDISK already has NFSCLIENT enabled and miniroot's FSSIZE
still fits after linking mount_nfs:
/dev/vnd0a6312 5592 72089% 3 59
5% /mnt
This boots fine on T4-2 guest domains and I can upgrade via NFS:
Location of sets? (disk http
The ERR trap is supposed to be run once, regardless of errexit:
$ ksh -c 'trap "echo ERR" ERR; false'
ERR
$ ksh -c 'trap "echo ERR" ERR; false' -e
ERR
ERR
This duplication is a regression of the following bin/ksh/main.c commit:
revision 1.52
On Sat, Oct 08, 2022 at 11:22:59AM -0600, Todd C. Miller wrote:
> On Sat, 08 Oct 2022 16:01:18 -0000, Klemens Nanni wrote:
>
> > Account for this to make permanent redirection failure trigger the ERR
> > trap:
> >
> > $ ./obj/ksh -c 'trap "echo ERR" E
No functional changes, just breaking the long line.
OK?
Index: install.sub
===
RCS file: /cvs/src/distrib/miniroot/install.sub,v
retrieving revision 1.1210
diff -u -p -r1.1210 install.sub
--- install.sub 5 Oct 2022 19:30:47 -
On Tue, Oct 04, 2022 at 08:11:46PM +, Klemens Nanni wrote:
> This function's style is a bit off: it wraps the body in a subshell to
> discard all stdout/err at once, but a still uses return inside it.
>
> 1. A command list (using {}) would be enough here as it groups like a
&
On Wed, Oct 05, 2022 at 04:56:57PM +0200, Alexander Hall wrote:
> While I dislike the ">/dev/null 2>&1" sledgehammer, this is in the right
> direction.
Agreed.
We should be fine silencing only the test condition which produces legit
output and warnings.
All else produces no output and should
On Sun, Oct 02, 2022 at 12:29:34PM +, Klemens Nanni wrote:
> dhcpd(8) has
> -n Only test configuration, do not run dhcpd.
>
> rc_configtest() taken from other rc.d scripts.
>
> Feedback? OK?
P
If /usr is mounted read-only, kernel relinking fails silently without
any log trace.
# ksh /usr/libexec/reorder_kernel
/usr/libexec/reorder_kernel[35]: cannot create
/usr/share/relink/kernel/GENERIC.MP/relink.log: Read-only file system
This stderr line does not show up during
Failure to permanently redirect I/O for the shell itself, i.e. when no
command is executed, must trigger the ERR trap like any other failure,
but that is currently not happening:
(non-zero exit code is trigger, no redirection)
$ ksh -c 'trap "echo ERR" ERR ; false'
ERR
(failed
On Sat, Oct 08, 2022 at 07:33:38AM +0200, Florian Obser wrote:
> Being up2date doesn't feel like an error to me, what am I missing?
I concur.
When upgrading to releases, the installer fills rc.firsttime(8) with
a syspatch(8) snippet possibly displaying available patches.
That snippet itself checks for a release version as well as an existent
installurl(5) file as a precondition for syspatch, see the diff below.
syspatch, however, has
On Tue, Sep 06, 2022 at 03:11:45PM +, Klemens Nanni wrote:
> On rare occasions, I need 'disable xxx' in /etc/bsd.re-config to be able to
> boot a system, e.g. to ignore quirky devices crashing drivers during attach.
>
> bsd.re-config(5) currently applies to GENERIC(.MP
The Oracle OpenBoot 4.x Administration Guide[0] documents a few useful
options for network boot.
Basically do either
{ok} setenv network-boot-arguments tftp-retries=0
{ok} boot net ...
or
{ok} boot net:tftp-retries=0 ...
Newer machines like the T4-2 using OpenBoot 4.38.16
On Sat, Sep 10, 2022 at 02:10:22AM +, Klemens Nanni wrote:
> It does not have the prettiest signature, but nicely folds identical
> copies into MI softraid.c, which then allows us to
> - avoid further diverging MD code
> - implement the keydisk fix on tech@ once instead of thri
On Sun, Sep 25, 2022 at 12:59:27AM +, Klemens Nanni wrote:
> On a sparc64 T4-2 I boot OpenBSD from hardware RAID 1E using the
> built-in controller:
>
> mpii0 at pci15 dev 0 function 0 "Symbios Logic SAS2008" rev 0x03: msi
> mpii0: Solana On-Board, fir
This is only required for the single fchmod(2) ensuring default permissions
which only happens in the -c code path; "fattr" without "cpath" makes no
sense here.
OK?
Index: tftpd.c
===
RCS file: /cvs/src/usr.sbin/tftpd/tftpd.c,v
There is no problem to fix, but every boot I read "/clearing /tmp" and
know it is a useless step since my /tmp live on volatile RAM anyway.
Other steps in rc(8) also check and print/log conditionally, so this
can do as well, saving yet another line.
There is also the unconditional
echo
301 - 400 of 1834 matches
Mail list logo