Age | Commit message (Collapse) | Author |
|
ok deraadt@ danj@
|
|
files during upgrade".
Mistakenly removed during dhclient.conf cleanup of r1.1050.
|
|
files.
Usual man page help & ok jmc@
|
|
hostname.if files. dhclient(8) does that itself. Part of the
"further script optimizations" promised in r1.200 of netstart.
Tested & ok gnezdo@
|
|
return before the interface address and route are set in the kernel.
The auto installer script runs a ftp(1) command immediatelty after
that, it fails without a local address. Before calling ftp, wait
until the address appears in ifconfig(8) output and dhclient leases
file. This makes autoinstall more reliable.
discussed with patrick@; input and OK krw@
|
|
|
|
to perform network configuration (for example, "!route source -ifp em0")
Split mygate and myname manual pages (how did anyone ever believe these
are related), and perform hostname configuration much earlier in rc.
discussed with benno, claudio, jmc, etc etc, last version of !command
parser by tb
|
|
multiple root disks, do the upgrade on the one with auto_upgrade.conf
present.
use case pointed out by, test & OK kettenis
|
|
auto_upgrade.conf in bsd.rd instead of sysupgrade(8) itself.
ajacoutot@ fixed this, but had to introduce a bunch of compat code in
bsd.rd.
It is time to retire this. Only 6.5-current contained this flaw,
sysupgrade(8)ing from there to -current will no longer work but we
never made any promises that it would anyway.
OK deraadt, ajacoutot
|
|
backslash at the end for line continuation
Breaking long lines into multiple ones must still be possible and does
require to treat the backslash as an escape character.
Breakage reported by Mark Patruck <mark at wrapped dot cx >, thanks!
---
distrib/miniroot/install.sub revision 1.1151
etc/netstart revision 1.203
date: 2020/05/21 11:54:41; author: kn; state: Exp; lines: +2 -2;
Do not treat backslashe as an escape character in hostname.if(5) lines
ifstart() should always pass such lines unaltered, especially if they
contain "nwid" or "description" lines with arbitrary strings.
<bsdlisten at gmail dot com> reported SSIDs such as "Mike's" during
installation end as broken; this was because the installer escaped
the single quote using backslashes which ended up being treated as
escape characters much later during hostname.if parsing in netstart(8).
Ok deraadt
|
|
Neither netstart's nor install.sub's (subtly different) implementations
remove trailing comments on lines not starting as a comment, e.g.,
lines like "up #not down" go through unaltered and without "#not down"
being removed.
Only lines *beginning* with the comment sign ("#") are stripped.
No functional change, just updating function descriptions.
|
|
ifstart() should always pass such lines unaltered, especially if they
contain "nwid" or "description" lines with arbitrary strings.
<bsdlisten at gmail dot com> reported SSIDs such as "Mike's" during
installation end as broken; this was because the installer escaped
the single quote using backslashes which ended up being treated as
escape characters much later during hostname.if parsing in netstart(8).
Ok deraadt
|
|
create audio(4) and midi(4) device nodes.
|
|
luna88k and sgi.
ok otto@ deraadt@ (for diff with same effect)
|
|
|
|
MDFSOPT and add a missing prototype.
|
|
/dev/stdin to fake a dhclient.conf file during install. Simplifies and
shortens install.sub code. Allows further restrictions to be applied
to '-c' specified files.
|
|
during installation. This is the default value.
|
|
Do it the same way in all three copies of the code.
Based on a patch from Chris Waddey.
with & ok ajacoutot
|
|
tight. At upgrade time (just before new sets hit the disk) we can prune
some potentially unused clang, gcc, and perl pieces. This seems to help.
(This is an improvement on a previous strategy)
discussed with sthen and tb, tested by kmos, in snaps for 2 days
This strategy ins't perfect, and people should consider fresh installs once
in a while, or more strict pruning especially of /usr/lib/lib*.so
|
|
responsible for creating auto_upgrade.conf.
No behavior change intended.
While here, add missing `-x' to usage().
tested by and ok florian@, ok deraadt@
|
|
between fetch&verify)
ok sthen@
|
|
Instead of ignoring SSIDs containing whitespaces, slightly adjust the
commands to take everything in between 'nwid ' and ' chan' as SSID; if it
has double quotes at start *and* end, simply remove those.
This enables users to select networks such as "Unitymedia WifiSpot" and
"FRITZ!Box 7490" for example which are common among the quoted ones at
least in germany.
The only SSIDs known to break this are those containing " chan " as this
substring is used as delimiter. Picking "some chan 4 me" would therefore
result in _nwid being assigned '"some' (literal double quote), but that
seems reasonably acceptable compared to the current behaviour.
Input rpe
"Yes please" stsp
|
|
The previous mechanism used a single timeout for the entire upgrade which
was kept when introducing the per-set watchdog.
Half an hour now seems more sensible to safely catch the biggest sets on
slow hardware, so avoid needlessly stalling (failed) upgrades for too long.
OK sthen deraadt
|
|
* Remove syspatch files from the installed system and not the ramdisk.
* Directly try to remove the old files without superfluous check.
* Use an extended glob to match /usr/lib/gcc-lib contents that don't
belong to this OS version.
* For /usr/lib/clang, set an array to the expansion of the glob, which is
an alphabetically sorted list, drop the last element, and rm the rest.
ok tb@
|
|
Instead of having a single timeout for the entire upgrade, start anew right
before extracting each set's tarball.
Do so by replacing the background job with a proper co-process that gets
killed and restarted (reset), analogue to the dmesg listener in interactive
mode.
Prompted by and OK sthen
OK deraadt
|
|
OK sthen
|
|
on my anoncvs server it takes about 20m already to untar sets, which is
a bit close to the limit for my liking (we only want to trigger the
watchdog if things are *really* stuck; if it's still proceeding then
rebooting mid-upgrade can leave things in a worse state).
we can probably do something smarter (e.g. reset the watchdog after
untarring each set) but this is an easier-to-test initial improvement.
ok florian deraadt kn
|
|
possible, and remove the failed previous attempts at sharing
in ./miniroot and ./ramdisk. maybe now that differences are
eliminated we can start a new sharing effort? i dunno..
|
|
there is no downside for others.
|
|
|
|
espie@. Need to look deeper.
|
|
moving progress bar during auto upgrade/install and a clean log
afterwards. ok deraadt@
|
|
Adapt to recent sysupgrade(8) change
ok florian@ kmos@ kn@
|
|
Should fix autoinstalls, suppress pointless questions, allow section
of 'none' via a number, etc.
Issues first reported by reyk@
Testing & ok florian
|
|
It didn't always print the actual question so it was hard to debug
problems in the autoinstall(8) script.
OK florian@
|
|
Allow 'dhcp' even if the install media lacks /sbin/dhclient.
Great simplification of code.
ok florian@
|
|
an install. Use the info as default answers in
relevant questions.
|
|
via [v4|v6]_info() after calling ifconfig.
As a side effect 'inet' lines in hostname.if will always use the
hex format for the netmask, i.e. as displayed by ifconfig(8).
|
|
verification by the installer:
Verification is triggered by the presence of SHA256.sig.
In sysupgrade, remove SHA256.sig once the signature has been verified.
Compare SHA256 against a stored copy from the previous install/upgrade.
In the installer, stash away a copy of SHA256 and move that code
into install_files() where the sets are actually processed.
Confirm in the response file that we want to proceed without
verifying the sets.
ok florian@ tj@
|
|
invalid address or netmask/prefixlen values are entered.
Keeps the presented default values sane in the face of
cats or otherunintended keyboard events.
|
|
more like each other.
|
|
completely clear existing configurations after restarting an install.
|
|
when restarting an install. Even if the original dhcp
attempt did not successfully get a lease and configure
the interface.
ok beck@ florian@
|
|
relevant interfaces at once, and ask for a user decision right after
we ask for the IPv4 default route.
ok florian@
|
|
tweak & OK kmos, OK semarie
|
|
'autoconf' if install restarts. Be more
careful to keep things clean. Remove unnecessary
checks left over from rtsol days.
Started by claudio@ noticing 'autoconf' is not
remembered.
ok florian@ kn@ (mostly)
|
|
Such media are not entirely accurate -- that's fine. However one case
is very bad -- the TZ tree is entirely skipped, resulting in artificially
smaller media and confusion. Just assume DESTDIR=/ for that case...
|
|
It seems like sysupgrade(8) could use this to figure out if a new
snapshot is available.
|
|
Eliminate many more differences. When it makes sense, build bsd.rd
and miniroot/ramdisk.fs and cd.iso in the same directory. More steps
coming after this..
|