Sascha Wildner [Sat, 29 Aug 2020 12:40:18 +0000 (14:40 +0200)]
vkernel: Adjust vke(4) for the if_media.h changes (fixes vkernel build).
Sascha Wildner [Sat, 29 Aug 2020 10:13:58 +0000 (12:13 +0200)]
<net/if_media.h>: Adjust ethernet media types some more.
* IFM_1000_FX was cleaned up long ago in FreeBSD (see r95702) and
NetBSD.
* Add a useful comment from FreeBSD.
* Adjust some whitespace and indentation.
While here, clean up the ifmedia.4 manual page a bit and add a
IFM_10G_TWINAX description (from NetBSD).
Aaron LI [Sat, 29 Aug 2020 08:00:38 +0000 (16:00 +0800)]
x86_64/machdep.c: Fix two minor typos and indentation
Matthew Dillon [Sat, 29 Aug 2020 01:32:43 +0000 (18:32 -0700)]
kernel - Fix KVM implosion by enabling the IOAPIC
* We were disabling the IOAPIC universally when running under a VM,
because older VMs sometimes broke when it was enabled.
However, this can actually implode the virtual machine by causing
interrupt routing to go haywire.
Add a case statement and always enable the IOAPIC on bare hardware
and KVM guests.
* Change the default to also enable the IOAPIC on all other VMs.
Cases can be added for specific disablement if necessary.
* Fixes Google Cloud Environment booting.
Matthew Dillon [Sat, 29 Aug 2020 01:27:46 +0000 (18:27 -0700)]
kernel - Be more verbose about disabling the IOAPIC
* Be more verbose when disabling the IOAPIC.
* We current disable auto-EOI on the lapic. Add some code in
case we enable in the future to ensure that it remains disabled
for KVM guests.
Matthew Dillon [Sat, 29 Aug 2020 01:23:50 +0000 (18:23 -0700)]
kernel - Fix atomicy issue in APIC EOI and enable_icus
* Change ICU_INTREN() and ICU_INTRDIS() to only modify the ICU
whos bit is being changed, rather than always setting the mask
for both ICUs.
* If masking a level IRQ on the APIC, make sure the EOI
to the APIC is atomic with the masking operation.
* Make sure that ICU enablement is atomic with ICU masking
operations.
* Does not fix any known bugs
Tomohiro Kusumi [Fri, 28 Aug 2020 15:24:28 +0000 (00:24 +0900)]
sbin/hammer2: Introduce HAMMER2_SHOW_DEPTH to limit "show" depth
Useful to check HAMMER2 radix tree nodes not so deep in the tree.
Tomohiro Kusumi [Fri, 28 Aug 2020 15:23:46 +0000 (00:23 +0900)]
sbin/hammer2: Drop extra indentation in "show" when -q
Tomohiro Kusumi [Fri, 28 Aug 2020 15:39:59 +0000 (00:39 +0900)]
sbin/hammer2: Fix "show" and "volhdr" output format
Sascha Wildner [Fri, 28 Aug 2020 11:29:32 +0000 (13:29 +0200)]
libc/getdate: Add getdate_err to the Symbol.map too.
Seems lang/clisp picks up getdate().
Reported-by: zrj
Sascha Wildner [Fri, 28 Aug 2020 04:38:08 +0000 (06:38 +0200)]
kernel: Adjust wmesg.
Just like in vfs_rootmountalloc().
Pointed-out-by: aly
Sascha Wildner [Thu, 27 Aug 2020 15:41:04 +0000 (17:41 +0200)]
<net/if_media.h>: IFM_HPNA_1 != IFM_10G_LR, renumber as needed.
Reported-by: zrj
Tomohiro Kusumi [Thu, 27 Aug 2020 15:08:07 +0000 (00:08 +0900)]
sys/vfs/hammer2: Remove HAMMER2_MINIO* macros
HAMMER2_MINIO* were added in
1a7cfe5ae3c897f704a358fd3e556a55e430dcb1
in 2013 to switch HAMMER2_MINIO* values with ifdefs when ondisk format
wasn't stable.
This is no longer needed, and using both HAMMER2_MINIO* and HAMMER2_LBUF*
should be avoided. Users of HAMMER2_MINIO* exist in several userspace
programs which walk through blockrefs from top to bottom.
Sascha Wildner [Thu, 27 Aug 2020 15:17:52 +0000 (17:17 +0200)]
efibootmgr(8): Sync with FreeBSD.
Adds the -f (--fw-ui) and -F (--no-fw-ui) options to enter (or not) the
firmware's user interface upon next boot.
Taken-from: FreeBSD
Matthew Dillon [Thu, 27 Aug 2020 05:55:54 +0000 (22:55 -0700)]
tmpfs - Improve assertions
* Improve assertions to detect failed RB_INSERT() operations.
Matthew Dillon [Thu, 27 Aug 2020 05:41:05 +0000 (22:41 -0700)]
kernel - Deal with VOP_NRENAME races
* VOP_NRENAME() as implemented by the kernel can race any number of
ways, including deadlocking, allowing duplicate entries, and panicing
tmpfs. It typically requires a heavy test load to replicate this but
a dsynth build triggered the issue at least once.
Other recently reported tmpfs issues with log file handling might also
be effected.
* A per-mount (semi-global) lock is now obtained whenever a directory
is renamed. This helps deal with numerous MP races that can cause
lock order reversals.
Loosely taken from netbsd and linux (mjg brought me up to speed on
this). Renaming directories is fraught with issues and this fix,
while somewhat brutish, is fine. Directories are very rarely renamed
at a high rate.
* kern_rename() now proactively locks all four elements of a rename
operation (source_dir, source_file, dest_dir, dest_file) instead of
only two.
* The new locking function, cache_lock4_tondlocked(), takes no chances
on lock order reversals and will use a (currently brute-force)
non-blocking and lock cycling algorithm. Probably needs some work.
* Fix a bug in cache_nlookup() related to reusing DESTROYED entries
in the hash table. This algorithm tried to reuse the entries while
maintaining shared locks, since only the entries need to be manipulate
to reuse them. However, this resulted in lookup races which could
cause duplicate entries. The duplicate entries then triggered
assertions in TMPFS.
* nlookup now tries a little harder and will retry if the parent of an
element is flagged DESTROYED after its lock was released. DESTROYED
elements are not necessarily temporary events as an operation can wind
up running in a deleted directory and must properly fail under those
conditions.
* Use krateprintf() to reduce debug output related to rename race
reporting.
* Revamp nfsrv_rename() as well (requires more testing).
* Allow nfs_namei() to be called in a loop for retry purposes if
desired. It now detects that the nd structure is initialized
from a prior run and won't try to re-parse the mbuf (needs testing).
Reported-by: zrj, mjg
Matthew Dillon [Wed, 26 Aug 2020 18:56:52 +0000 (11:56 -0700)]
kernel - Document confusing physmap[1] initialization
* Document confusing physmap[1] initialization
Matthew Dillon [Wed, 26 Aug 2020 17:47:50 +0000 (10:47 -0700)]
tmpfs - Enhance assertion (not a bug fix)
* Enhance a reported assertion to provide more information.
Reported-by: zrj
Matthew Dillon [Wed, 26 Aug 2020 17:44:42 +0000 (10:44 -0700)]
kernel - Remove tty-related debug message
* Remove a debugging message that is no longer needed. The kprintf was
added to determine if a particular handled race could actually happen.
Answer is: yes it can.
Tomohiro Kusumi [Wed, 26 Aug 2020 15:15:40 +0000 (00:15 +0900)]
sys/vfs/hammer2: Remove comment on nonexistent MINALLOCSIZE
Removed in
1a7cfe5ae3c897f704a358fd3e556a55e430dcb1 in 2013.
Tomohiro Kusumi [Wed, 26 Aug 2020 15:15:18 +0000 (00:15 +0900)]
sbin/newfs_hammer2: Use ~HAMMER2_PBUFMASK64 to obtain pbuf offset
HAMMER2 in practice uses ~HAMMER2_PBUFMASK64 for this purpose
rather than HAMMER2_OFF_MASK_HI which is only used here.
Also more readable considering use of PBUF macros in HAMMER2.
Sascha Wildner [Wed, 26 Aug 2020 09:54:37 +0000 (11:54 +0200)]
kernel/{a,b}list: Use __func__ in panic messages.
This fixes some mistakes too (*_radix_free -> *_leaf_free).
Tomohiro Kusumi [Tue, 25 Aug 2020 17:16:15 +0000 (02:16 +0900)]
sys/vfs/hammer2: Fix whitespace to align indentation
Tomohiro Kusumi [Tue, 25 Aug 2020 17:10:19 +0000 (02:10 +0900)]
sys/vfs/hammer2: Remove unused HAMMER2_AV* macros
Appeared in
9061bde5ad5e46bcaee746ed03d598a75b38691f in 2013,
but never used.
The freemap layout when these were added is no longer compatible
with the current one, so these won't be used again either.
Sascha Wildner [Tue, 25 Aug 2020 04:43:26 +0000 (06:43 +0200)]
kernel/nata: Fix operator precedence issue.
Sascha Wildner [Mon, 24 Aug 2020 19:04:04 +0000 (21:04 +0200)]
zoneinfo: Update /etc/localtime when upgrading timezones.
* Add tzsetup(8) to the bootstrap-tools.
* Install the backward file to retain obsolete timezones as links
for backward compatibility. Remove all these timezones from
Makefile_upgrade.inc again. Also extend mtree/BSD.usr.dist for
this and remove the old 'dragonfly' file (which was only for UTC).
* After installing new timezones, run tzsetup -r to upgrade the user's
chosen timezone. Allow this to fail for quickworlders. It will start
working after the next full buildworld.
Taken-from: FreeBSD (with modifications)
Tomohiro Kusumi [Mon, 24 Aug 2020 16:20:11 +0000 (01:20 +0900)]
sys/vfs/hammer2: Remove unused HAMMER2_OFF_NOPREF
Appeared in
1a7cfe5ae3c897f704a358fd3e556a55e430dcb1 in 2013,
but never used.
Tomohiro Kusumi [Mon, 24 Aug 2020 16:14:36 +0000 (01:14 +0900)]
sys/vfs/hammer2: Remove unused HAMMER2_OFF_BAD and HAMMER2_MAX_COPIES
Appeared in
703720e4d599857d052f0a65f2840224ce36ec5a in 2012,
but never used.
Tomohiro Kusumi [Sun, 23 Aug 2020 15:44:38 +0000 (00:44 +0900)]
sys/vfs/hammer2: Remove unused HAMMER2_SPECTHREADS
No longer used since
2085215738c03d949e60de63843cb91e84836eb9 in 2016.
Sascha Wildner [Sun, 23 Aug 2020 09:50:00 +0000 (11:50 +0200)]
<sys/systm.h>: Fix muldivu64() comment.
Sascha Wildner [Sat, 22 Aug 2020 18:03:16 +0000 (20:03 +0200)]
kernel/pmap: Remove code under !defined(PMAP_ADVANCED).
We've been running with PMAP_ADVANCED by default since February 27.
Remove the old, inactive code.
Approved-by: dillon
Matthew Dillon [Fri, 21 Aug 2020 18:17:05 +0000 (11:17 -0700)]
kernel - Remove global cwd statistics counters
* These were only used for debugging purposes and interfere with
MP operation. Just remove them.
Sascha Wildner [Fri, 21 Aug 2020 17:36:48 +0000 (19:36 +0200)]
make(1): Fix default search path back to just /usr/share/mk.
This '...' hack was intended to fix pkgsrc problems at the time but
didn't work (see
2b01fcf91a45a96b and
3db89b525e9bf920). And it was
then left in in the hopes it might be useful for something.
However, in practice it can only lead to confusion when you're
building in another BSD's tree for testing purposes and make starts
looking for and using .mk files we don't have in the upper levels
of the foreign tree instead of properly reporting that it could not
find the file.
Roy Marples [Thu, 20 Aug 2020 19:53:00 +0000 (20:53 +0100)]
dhclient: teach the script about resolvconf(8)
Sascha Wildner [Thu, 20 Aug 2020 09:55:58 +0000 (11:55 +0200)]
kate.4: Sort SEE ALSO.
Daniel Fojt [Wed, 19 Aug 2020 18:14:00 +0000 (20:14 +0200)]
games/sail: adjust to updated ncurses
Reported by: Rimvydas Jasinskas
Daniel Fojt [Wed, 19 Aug 2020 18:12:11 +0000 (20:12 +0200)]
games/mille: adjust to updated ncurses
Reported by: Rimvydas Jasinskas
Tomohiro Kusumi [Wed, 19 Aug 2020 15:12:33 +0000 (00:12 +0900)]
sys/vfs/hammer2: Remove unused tailq structs
No longer used since
da6f36f44bb9c76612e31b56cdfe4d787e53d61f in 2014 and
ecfe89b8868b30c9ddfa9c86cb4b0a20365a248d in 2018.
Michael Neumann [Wed, 19 Aug 2020 14:11:34 +0000 (16:11 +0200)]
kernel - Remove machdep.hack_efifb_probe_early hack
* This hack was introduced as a temporary fix for bug #3167 (see commit
c2a57f42).
* Since commit faeb2db "kernel - Hack the DMAP size" the temporary hack is no longer required.
* Tested on TUXEDO InfinityBook Pro 14v4 laptop where this bug initially occured.
David Shao [Tue, 18 Aug 2020 11:18:36 +0000 (04:18 -0700)]
kernel: recognize newer Intel I219 ethernet controllers
Define additional Intel I219-LM and I219-V Ethernet PCI device numbers.
Taken from NetBSD current sys/dev/pci/pcidevs.h, verified on physical
hardware for I219-V11 0x0D4D.
Submitted via: https://bugs.dragonflybsd.org/issues/3244
Daniel Fojt [Fri, 14 Aug 2020 19:28:41 +0000 (21:28 +0200)]
Upgrade ncurses. 2/2
- update README.{DELETED,DRAGONFLY}
- adapt Makefiles
- regenerate header files
- import new generated include/comp_userdefs.c
- adjust Symbol.map
Daniel Fojt [Thu, 13 Aug 2020 16:27:16 +0000 (18:27 +0200)]
Upgrade ncurses. 1/2
Merge branch 'vendor/NCURSES' into master.
Daniel Fojt [Wed, 12 Aug 2020 07:35:19 +0000 (09:35 +0200)]
vendor/ncurses: upgrade from 6.0-
20160305 to 6.2-
20200212
The TERMINAL structure in <term.h> is now opaque. Doing that allowed
making the structure larger, to hold the extended numeric data.
When configured for wide-characters (i.e. ncursesw), the TERMINAL
structure is extended and is named TERMTYPE2. When configured for
8-bit (narrow) characters, the TERMTYPE2 structure is not used.
For detailed description of all other changes, enhancements, bugfixes
and performance improvements, see:
https://invisible-island.net/ncurses/announce-6.1.html
https://invisible-island.net/ncurses/announce-6.2.html
Sascha Wildner [Tue, 18 Aug 2020 21:13:52 +0000 (23:13 +0200)]
openpam: Add back includes removed in
7196b98de8a8d.
I should have just added <security/openpam.h>.
Sascha Wildner [Tue, 18 Aug 2020 19:07:17 +0000 (21:07 +0200)]
yacc(1): Add config.h and use it.
Sascha Wildner [Tue, 18 Aug 2020 18:28:00 +0000 (20:28 +0200)]
__realpath.2: Fix typo.
Daniel Fojt [Tue, 18 Aug 2020 17:15:38 +0000 (19:15 +0200)]
kdump(8): fix indentation in mkioctls
Tomohiro Kusumi [Tue, 18 Aug 2020 15:54:40 +0000 (00:54 +0900)]
sys/vfs/ext2fs: Don't do endian conversion twice
Doesn't affect anything on DragonFly.
from freebsd/freebsd@
86fbc13431233b277bc8739539a18019ae53e88f
Michael Neumann [Mon, 17 Aug 2020 21:19:51 +0000 (23:19 +0200)]
pthread - Respect RLIMIT_STACK for main thread's stack size
Prior to this commit, the stack size of the main thread was fixed
at 4 MB which was too low for certain applications.
This commit fixes bootstrapping the lang/crystal compiler (see [1]).
Reviewed-by: dillon
Fixes-issue: #3234
[1]: https://github.com/crystal-lang/crystal/pull/9178
Sascha Wildner [Mon, 17 Aug 2020 16:33:09 +0000 (18:33 +0200)]
Fix the synopses of some libarchive and openpam manual pages.
Sascha Wildner [Sat, 15 Aug 2020 15:06:15 +0000 (17:06 +0200)]
<sys/elf_common.h>: Add GNU note types and NT_GNU_PROPERTY_TYPE_0 bits.
Apparently some ports need those.
Reported-by: zrj
Taken-from: FreeBSD (r348628)
Sascha Wildner [Sat, 15 Aug 2020 06:22:03 +0000 (08:22 +0200)]
byacc: Add our local READMEs on master.
Sascha Wildner [Sat, 15 Aug 2020 06:21:09 +0000 (08:21 +0200)]
Merge branch 'vendor/BYACC'
Sascha Wildner [Sat, 15 Aug 2020 06:20:45 +0000 (08:20 +0200)]
vendor/BYACC: Remove our local READMEs from the vendor branch.
Sascha Wildner [Sat, 15 Aug 2020 06:10:52 +0000 (08:10 +0200)]
localedef(1): Set yydebug only when YYDEBUG is set.
Sascha Wildner [Sat, 15 Aug 2020 05:50:07 +0000 (07:50 +0200)]
Update the pciconf(8) database.
August 13, 2020 snapshot from https://pci-ids.ucw.cz
Sascha Wildner [Sat, 15 Aug 2020 05:48:09 +0000 (07:48 +0200)]
<sys/random.h>: Add missing __{BEGIN,END}_DECLS.
Reported-by: zrj
Tomohiro Kusumi [Thu, 13 Aug 2020 19:10:04 +0000 (04:10 +0900)]
usr.sbin/fstyp: Fix incorrect pfs_type test in ondisk inode
Fix the same bug fixed in
c2795ad60fbf0e88fa2f1c2dc29905d9c4f1bf92.
Sascha Wildner [Thu, 13 Aug 2020 18:14:00 +0000 (20:14 +0200)]
Enable -fno-common for world, too.
Sascha Wildner [Thu, 13 Aug 2020 18:13:47 +0000 (20:13 +0200)]
Stop mixing the userland specific <bsd.sys.mk> into the module build.
I don't know why it was there initially because we inherited it this way.
The only changes on the compile line are eliminating a -std=gcc99
(our userland -std setting) which was overridden by the kernel's
own -std=c99 anyway, and a second copy of our kernel warning flags.
By extension, it also makes the build log shorter by some percent.
Matthew Dillon [Thu, 13 Aug 2020 05:25:17 +0000 (22:25 -0700)]
dma - Update documentation
* Add some important missing information to the manual page.
Submitted-by: Doug Hardie
Sascha Wildner [Thu, 13 Aug 2020 01:39:00 +0000 (03:39 +0200)]
ifconfig(8): Fix typo.
Spotted-by: zrj
Matthew Dillon [Wed, 12 Aug 2020 22:37:07 +0000 (15:37 -0700)]
kernel - Fix multiple issues with if_tap and if_tun cloning (2)
* For tap and tun, detect whether created by ifconfig or
open("/dev/tap"...) etc and set MANUALMAKE flag appropriately.
* When closing a tap or tun interface and MANUALMAKE is not set,
destroy the device, softc, and interface via if_clone_destroy().
This fixes another issue where the various bitmaps could get out
of sync.
Matthew Dillon [Wed, 12 Aug 2020 20:47:43 +0000 (13:47 -0700)]
kernel - Fix multiple issues with if_tap and if_tun cloning
* tap and tun were being auto-destroyed when manually created via
ifconfig blah create, and not being auto-destroyed when auto-created
via open /dev/tap or open /dev/tun. Reverse the flag so it works
properly.
* The bitmap for the ifconfig based cloner and the bitmap maintained
internally by if_tap and if_tun could get out of sync with each
other due to if_clone_free_unit(ifp, ...) being issued after
the ifc_destroy() (the ifp is garbage at that point).
* Remove a KASSERT() that can be triggered trivially from usermode
(as root) in the interface cloning path.
Reported-by: zrj
Tomohiro Kusumi [Wed, 12 Aug 2020 17:16:01 +0000 (02:16 +0900)]
sbin/hammer2: Fix -Wsometimes-uninitialized from
18f4f120fa
Tomohiro Kusumi [Wed, 12 Aug 2020 15:51:28 +0000 (00:51 +0900)]
sys/vfs/hammer2: Remove misleading comment in readdir vop
since around
c847e8387ad749d611d395742d337213aefef3b9 in 2015.
Tomohiro Kusumi [Wed, 12 Aug 2020 15:15:11 +0000 (00:15 +0900)]
sbin/hammer2: Print iparent in "printinode" in hexadecimal
Sascha Wildner [Wed, 12 Aug 2020 14:45:56 +0000 (16:45 +0200)]
kdump(8): Gather more ioctl names to resolve.
Tomohiro Kusumi [Tue, 11 Aug 2020 22:45:52 +0000 (07:45 +0900)]
sbin/hammer2: ec -> ecode for consistency
Tomohiro Kusumi [Tue, 11 Aug 2020 18:04:02 +0000 (03:04 +0900)]
sbin/hammer2: Don't return 0 if "destroy*" failed
Sascha Wildner [Tue, 11 Aug 2020 22:14:47 +0000 (00:14 +0200)]
<sys/msgport.h>: For boolean_t, it is enough to include <sys/stdint.h>.
It was moved there from <sys/types.h> in
6b2b0a8b5c1c774a5c3cbb4c79bdc150.
Sascha Wildner [Tue, 11 Aug 2020 21:30:51 +0000 (23:30 +0200)]
kernel/iscsi: Remove iscsi's local definition of boolean_t.
It was locally defining boolean_t as int (4 bytes) for userland, but
boolean_t is _Bool (1 byte) in the kernel since quite some time.
This caused isc_opt_t's size being different in userland vs. kernel,
and ultimately caused the ioctl number of ISCSISETOPT, which includes
sizeof(isc_opt_t), to be different in userland vs. the kernel,
resulting in at first inexplicable ENOIOCTL issues when using
iscontrol(8).
Change all 'boolean_t' usage to 'bool' which is available for both
userland and kernel, defined as _Bool.
Thanks-to: Georg "megaT" Bege <georg@bege.email> for spotting the ioctl
number difference & testing
Tomohiro Kusumi [Tue, 11 Aug 2020 16:12:14 +0000 (01:12 +0900)]
sbin/hammer2: Cleanup "info"
Only volume, inode, indirect blockref (out of 9 types) needed for this.
Tomohiro Kusumi [Tue, 11 Aug 2020 15:47:08 +0000 (00:47 +0900)]
sbin/fsck_hammer2: Fix incorrect pfs_type test in ondisk inode
"ipdata.meta.pfs_type & HAMMER2_PFSTYPE_SUPROOT" happened to have
the same result (except HAMMER2_PFSTYPE_DUMMY could also match).
Daniel Fojt [Tue, 11 Aug 2020 08:53:20 +0000 (10:53 +0200)]
etc/Makefile: add a notice to "check-deprecated-files"
When checking dependencies for deprecated system libraries, kernel may
complain on "foreign" (non-DragonFly) ELF binaries, like:
"ELF interpreter /lib64/ld-linux-x86-64.so.2 not found"
Unfortunately, these complaints are printed directly to terminal, not
stdout or stderr, and there is no way to redirect it. So just print
a notice in advance, to avoid possible users' confusion.
Pointed out by: Pierre-Alain TORET
Discussed with: Sascha Wildner
Tomohiro Kusumi [Mon, 10 Aug 2020 17:11:19 +0000 (02:11 +0900)]
sbin/hammer2: Don't return 0 if "bulkfree" failed
Tomohiro Kusumi [Mon, 10 Aug 2020 17:08:36 +0000 (02:08 +0900)]
sbin/hammer2: Make "dumpchain" print strerror if not HAMMER2 inode
Tomohiro Kusumi [Mon, 10 Aug 2020 16:25:09 +0000 (01:25 +0900)]
sbin/hammer2: Minor cleanups
Daniel Fojt [Sun, 9 Aug 2020 14:13:29 +0000 (16:13 +0200)]
Upgrade dialog(1). 2/2
- update README.{DELETED,DRAGONFLY}
- adapt libdialog/Makefile
- regenerate dlg_config.h
Daniel Fojt [Mon, 10 Aug 2020 15:51:21 +0000 (17:51 +0200)]
Upgrade dialog(1). 1/2
Merge branch 'vendor/DIALOG' into master.
Daniel Fojt [Wed, 5 Aug 2020 06:08:20 +0000 (08:08 +0200)]
vendor/dialog: upgrade from 1.2-
20150920 to 1.3-
20200327
Summary of notable changes:
- add "--week-start" option for calendar widget
- modify editbox widget to add a trailing newline if the text has none
to ensure the last line is not ignored
- use Gregorian algorithm for leap year
- fix a regression from 2015/05/13 changes for escaping; it is necessary
to retain backslashes within quotes to make "\Z" escapes work
- change explicit checks for space character used for select or toggle
to make this rebindable to "TOGGLE"
- improve performance with very long command-lines
- interpret $DIALOGOPTS before expanding "--file", etc., to allow
the environment variable to turn on tracing in that process
- improve handling of SIGWINCH for several widgets
- add options --print-text-only, and --print-text-size for scripts that
adjust the widget size according to how the captions are formatted
- when trimming blanks, treat unconverted tabs the same as spaces
- allow for underline- and reverse-video flags in the ".rc" file
For detailed list of all changes, bugfixes and improvements, see CHANGES.
Sascha Wildner [Mon, 10 Aug 2020 14:19:43 +0000 (16:19 +0200)]
boot/efi: Fix boot1's compilation with -fno-common.
boot1 needs no local definitions of BS, IH, RS and ST because they are
defined in libefi that boot1 links to.
Tomohiro Kusumi [Mon, 10 Aug 2020 12:05:43 +0000 (21:05 +0900)]
sbin/hammer2: Use calloc(3)
Tomohiro Kusumi [Mon, 10 Aug 2020 10:15:09 +0000 (19:15 +0900)]
sbin/hammer2: Use nitems()
Tomohiro Kusumi [Sun, 9 Aug 2020 15:45:09 +0000 (00:45 +0900)]
sbin/hammer2: Fix no extern declaration errors
Sascha Wildner [Sun, 9 Aug 2020 13:48:43 +0000 (15:48 +0200)]
Fix GETATTR_LITE() build breakage with 'pseudo-device gzip' in the config.
Sascha Wildner [Sun, 9 Aug 2020 10:06:53 +0000 (12:06 +0200)]
boot/efi: Switch from HandleProtocol() to OpenProtocol().
HandleProtocol() is deprecated.
Bring in and use FreeBSD's OpenProtocolByHandle() helper function (which
is like the UEFI Spec's example) and generally reduce differences with
FreeBSD a bit more.
Taken-from: FreeBSD
Sascha Wildner [Sun, 9 Aug 2020 08:48:06 +0000 (10:48 +0200)]
boot/efi: Move the loader's efi_main() from libefi to the loader itself.
It will allow boot1 (that has its own efi_main()) to use libefi.
While here, reduce differences with FreeBSD a bit more.
Taken-from: FreeBSD
Sascha Wildner [Sun, 9 Aug 2020 07:49:25 +0000 (09:49 +0200)]
boot/efi: Make DevicePathNodeLength()'s result a size_t.
Taken-from: FreeBSD
Tomohiro Kusumi [Sat, 8 Aug 2020 19:59:59 +0000 (04:59 +0900)]
sbin/fsck_hammer2: Add HAMMER2_USE_OPENSSL
Sascha Wildner [Sat, 8 Aug 2020 20:07:59 +0000 (22:07 +0200)]
<sys/sysid.h>: Include <sys/types.h> for self-sufficiency.
Tomohiro Kusumi [Sat, 8 Aug 2020 18:43:39 +0000 (03:43 +0900)]
sbin/fsck_hammer2: Use RB_GENERATE() for better userspace portability
Matthew Dillon [Sat, 8 Aug 2020 04:39:28 +0000 (21:39 -0700)]
kernel - Refactor GETATTR_QUICK() -> GETATTR_LITE()
* Refactor GETATTR_QUICK() into GETATTR_LITE() and use struct
vattr_lite instead of struct vattr. The original GETATTR_QUICK()
just used a struct vattr.
This change ensures that users of this new VOP do not attempt to
access attr fields that are not populated.
Suggested-by: mjg
Sascha Wildner [Sat, 8 Aug 2020 10:48:58 +0000 (12:48 +0200)]
usb.4: Add urndis(4) reference in SEE ALSO.
Sascha Wildner [Sat, 8 Aug 2020 07:46:57 +0000 (09:46 +0200)]
kernel: Remove <machine/npx.h> from two headers that do not need it.
Sascha Wildner [Sat, 8 Aug 2020 07:26:39 +0000 (09:26 +0200)]
Remove unneded #includes in three header files.
Namely, <sys/ckpt.h>, <sys/posix4.h> and <sys/sfbuf.h>. None of them
needed anything from the headers they were including.
Sascha Wildner [Fri, 7 Aug 2020 18:38:44 +0000 (20:38 +0200)]
<machine/tss.h>: Oops, it should have been <sys/types.h>.
We want integer types too.
François Tigeot [Fri, 7 Aug 2020 16:42:14 +0000 (18:42 +0200)]
drm: Add linux/irqhandler.h
Sascha Wildner [Fri, 7 Aug 2020 06:24:51 +0000 (08:24 +0200)]
<machine/tss.h>: Include <sys/cdefs.h> for __packed.
Sascha Wildner [Fri, 7 Aug 2020 06:17:57 +0000 (08:17 +0200)]
<machine/segments.h>: Move the inclusion of <sys/tls.h> to the top.
Mainly to get at <sys/types.h> for u_int* and <sys/cdefs.h> for __packed.