Sascha Wildner [Fri, 5 Jan 2018 01:57:01 +0000 (02:57 +0100)]
Clean up a bit for symlinked headers.
<errno.h> as well as <poll.h> are symlinked to their <sys/...>
counterparts, so there is no need to include both.
Also fix a comment that suggests that there is a difference.
Peeter Must [Thu, 4 Jan 2018 12:34:57 +0000 (14:34 +0200)]
kernel/evdev: Expose headers to userland.
zrj [Thu, 4 Jan 2018 10:02:59 +0000 (12:02 +0200)]
libc/nls: Sync with FreeBSD.
Move out catentry structure destroy code to its own function.
Set refcount = 0 in SAVEFAIL() macro.
Taken-from: FreeBSD
zrj [Wed, 3 Jan 2018 07:25:04 +0000 (09:25 +0200)]
Remove src/share/nls.
It was never hooked into the build after initial Citrus import in:
32a7b49af5025caa26dc703952f47b39d046c9eb
zrj [Wed, 3 Jan 2018 07:20:52 +0000 (09:20 +0200)]
Simplify legacy share/nls handling.
Remove src/nls.alias and create legacy compat links directly.
The nls.alias method is unsafe and will break with our 'make upgrade'.
Eitan Adler [Thu, 4 Jan 2018 05:19:08 +0000 (21:19 -0800)]
morse(6): add static where appropriate
Sepherosa Ziehau [Wed, 3 Jan 2018 05:41:00 +0000 (13:41 +0800)]
ipfw.8: Fix 'not [iface]' and 'not [iface:net]' support
Sepherosa Ziehau [Sat, 9 Dec 2017 04:10:25 +0000 (12:10 +0800)]
em/emx: Sync with Intel's em 7.6.2
- Fix RX filter settings when exiting promiscous mode.
- Add workaround for I219 resetting.
- Fix some settings for 82574.
Tested-by: dillon@ (I219)
François Tigeot [Tue, 2 Jan 2018 21:25:54 +0000 (22:25 +0100)]
drm/i915: Add Coffee Lake support
* This commit mostly adds PCI IDs
* Handle all Coffee Lake GPUs as Kaby Lake ones for now
zrj [Tue, 2 Jan 2018 07:03:55 +0000 (09:03 +0200)]
cal(1): Rename to use de_AT.ISO8859-15 calendar.
zrj [Tue, 2 Jan 2018 06:02:42 +0000 (08:02 +0200)]
locales: Add "is_IS.ISO8859-1" locale.
Mainly for testing, we already have -15 and UTF-8 ones.
zrj [Tue, 2 Jan 2018 05:59:15 +0000 (07:59 +0200)]
cldr2def: Fix few typos in tools.
While there, do not hardcode /usr/src as a source directory.
zrj [Mon, 1 Jan 2018 08:51:48 +0000 (10:51 +0200)]
kernel/nata: Add loader hint to disable nata attach.
For cases where atapci devices (including cdroms) would fail to attach or have
other issues and are not needed for booting(like setup uses CAM, NVME, VIRTIO).
We include nata(4) in a X86_64_GENERIC kernel config, probing of ata bus can be
disabled by setting: "atapciN.disabled=1" (where N is ata pci unit).
Sascha Wildner [Mon, 1 Jan 2018 00:00:00 +0000 (01:00 +0100)]
Bump the copyrights.
Happy new 2018 to everyone and thanks for all the help and support! :)
zrj [Mon, 27 Nov 2017 08:28:03 +0000 (10:28 +0200)]
kernel/nata: Add more INTEL chipset IDs.
Refactor the ata_intel_new_setmode().
Taken-from: FreeBSD
zrj [Mon, 27 Nov 2017 07:57:17 +0000 (09:57 +0200)]
kernel/nata: Add bunch of chipset IDs.
Minor fixes here and there. Expertimental.
Taken-from: FreeBSD
zrj [Mon, 27 Nov 2017 06:05:41 +0000 (08:05 +0200)]
kernel/nata: Handle soft resets in ahci better.
This brings back the ATA_AHCI_GHC_AE|ATA_AHCI_GHC_HR.
Could not make ata_ahci_issue_cmd() to work reliably on my Intel 63XXESB2
SATA300 controller, for now this version will do. Experimental.
It is better to focus on implementing the CAM_ATA and prune ata-ahci.c.
zrj [Sun, 26 Nov 2017 15:53:16 +0000 (17:53 +0200)]
kernel/nata: Deal with ATA_DEV() and atadev->unit.
Hopefully I got all places correctly.
While there, some misc cleanup.
zrj [Sun, 26 Nov 2017 13:04:22 +0000 (15:04 +0200)]
kernel/nata: Fix HT1000 DMA engine 64k transfer issue.
Some chips seem to need 16bit writes.
Taken-from: FreeBSD
Also noticed interesting change back in:
5b527db0a83182c576889a67555810d780759bb7
zrj [Sun, 26 Nov 2017 12:13:19 +0000 (14:13 +0200)]
kernel/nata: Extract ata_tf_read() and ata_tf_write().
While there swap check order ata_intel_31244_command() and
adjust few places to use request->parent.
No functional change.
zrj [Sun, 26 Nov 2017 11:02:45 +0000 (13:02 +0200)]
kernel/nata: Extract ad_get_geometry().
While there, extract ata_unit2str() too.
zrj [Sat, 25 Nov 2017 19:07:45 +0000 (21:07 +0200)]
kernel/nata: Add Adaptec 1420 support.
It depends on Marwell so group it together.
While there, split out Cenatek, Micron generic chipset support to their own
drivers. Also fix up few drivers that cannot reliably do 64k transfers and
remove few unused ATA_AHCI_P_CMD_* defines.
Taken-from: FreeBSD
zrj [Sat, 25 Nov 2017 16:21:00 +0000 (18:21 +0200)]
kernel/nata: Extract reset functions in chipset drivers.
Explicitly do not import any port multipliers support for now.
Use only higher half of device signature, some devices return
incorrect values in lower part and causes issues in detection.
All of this is a bit messy in ata-ahci.c.
While there, include several later bugfixes.
Taken-from: FreeBSD
zrj [Sat, 25 Nov 2017 09:01:41 +0000 (11:01 +0200)]
kernel/nata: Misc cleanup in non chipset codes.
* Move some stuff around.
* Add local implementations of biofinish() and g_io_deliver().
* Add prints for READ_NATIVE_MAX_ADDRESS.
* Use >= in comparisons for devclass_get_maxunit()
No functional change.
zrj [Fri, 24 Nov 2017 15:17:16 +0000 (17:17 +0200)]
kernel/nata: Return more data for natacontrol(8).
* include info about backing subdisks
* use last 16 bytes of serial number in meta (as MatrixRAID does)
* add optional automatc spindown/spinup support (dmesg noisy)
* various cleanups
* natacontrol(8) additions + cleanup
Taken-from: FreeBSD
zrj [Fri, 24 Nov 2017 13:08:13 +0000 (15:08 +0200)]
kernel/nata: Pass interrupt function to ata_setup_interrupt().
There is one strange change ATA_AHCI_GHC_HR 0x80000001 -> 0x00000001.
It is used for AHCI reset, all other bits are already in place so going
with the flow on this one.
Futher reduce differences with FreeBSD r183724:
* add ata_netcell_setmode() instead of relying on ata_generic_chipinit()
* add ata_promise_mio_setprd() for promise
* prepare siliconimage for port multipliers
* spray some KASSERT() here and there
* fix typo for ATA_READ_NATIVE_MAX_ADDRESS
Taken-from: FreeBSD
zrj [Thu, 23 Nov 2017 20:24:45 +0000 (22:24 +0200)]
kernel/nata: Constification.
Attempt at making nata a bit more safe:
* const attributes where possible
* add and use ata_set_desc() helper
* add early returns (for future split modules support)
* fix a bug in ata-sis.c where it was patching ids table.
As a bonus, now CCVER=gcc47 no longer warn on -Warray-bounds in ata-ite.c,
finally allowing to buildkernel w/o NO_WERROR (ata_mode2idx() in ata-pci.c).
No functional change intended, except for bugfix in ata-sis.c.
zrj [Thu, 23 Nov 2017 17:34:27 +0000 (19:34 +0200)]
kernel/nata: Separate sata bits.
Do some cleanup of defines and headers.
Rename few ATI controllers.
While there, remove no longer used PC98 defines.
zrj [Thu, 23 Nov 2017 12:19:53 +0000 (14:19 +0200)]
kernel/nata: Allow disabling complicated chipset drivers.
There is no point for ATA_NO_ATI option (ati chipset setup relies on sii).
Using ATA_NO_AHCI will automatically disable these ata chipset drivers:
intel, marvel, nvidia, promise, siliconimage, sis, via
Only for development until nata(4) gets updated.
zrj [Thu, 23 Nov 2017 10:10:10 +0000 (12:10 +0200)]
kernel/nata: Move-out public chipset setup related defines.
Taken-from: FreeBSD
zrj [Thu, 23 Nov 2017 07:33:31 +0000 (09:33 +0200)]
kernel/nata: Use reg == 0x63 for NVIDIA chipsets.
It is not clear whether it was modularization artifact or actual bugfix.
For now changing to match FreeBSD ata variant. No hardware to test it.
This might need to be reverted if chipset initialization is wrong.
zrj [Thu, 23 Nov 2017 05:50:35 +0000 (07:50 +0200)]
kernel/nata: Use cfg1 on AMD and NVIDIA chipsets.
It is now clear that only one config state is needed.
Mainly to match the FreeBSD variant for less diffs.
zrj [Wed, 22 Nov 2017 19:04:14 +0000 (21:04 +0200)]
kernel/nata: Remove common ata_via_family_setmode().
Just extract and use vendor specific versions.
Note: there is some confusion about reg == 0x53 vs 0x63 for some chipsets.
For now leave the previous behaviour, thus no functional change.
zrj [Wed, 22 Nov 2017 13:54:19 +0000 (15:54 +0200)]
kernel/nata: Split out huge ata-chipsets.c
In preparations for pseudo-modularization.
If we decide to modularizate this one, it would be just wrapping each of the
chipsets sources in its own kernel module. Current variant just simply includes
everything back together while still allowing to work with individual chipset.
Thus no functional change.
Sascha Wildner [Sun, 31 Dec 2017 16:49:00 +0000 (17:49 +0100)]
ftpd(8): Revert a Makefile change that was not meant to be pushed.
Sascha Wildner [Sun, 31 Dec 2017 16:42:07 +0000 (17:42 +0100)]
Add __attribute__((__noreturn__)) to various function prototypes in libexec/.
This general cleanup is useful in various ways. It helps the compiler
optimize better and it also helps later GCCs detect better when to give
-Wimplicit-fallthrough warnings and when not.
Found-with: -Wsuggest-attribute=noreturn
Sascha Wildner [Sun, 31 Dec 2017 15:16:09 +0000 (16:16 +0100)]
rpc.sprayd(8): Staticize.
Sascha Wildner [Sun, 31 Dec 2017 15:12:01 +0000 (16:12 +0100)]
tftpd(8): Staticize.
Sascha Wildner [Sun, 31 Dec 2017 14:59:45 +0000 (15:59 +0100)]
telnetd(8): Remove duplicate prototypes.
zrj [Sun, 31 Dec 2017 14:13:54 +0000 (16:13 +0200)]
tools: Do not hardcode path to perl.
Solves issue with DPorts and Ravenports.
zrj [Sun, 31 Dec 2017 14:12:54 +0000 (16:12 +0200)]
world: Avoid extended ASCII.
Use UTF-8 where needed, use plain ASCII where possible.
No functional change.
zrj [Sun, 31 Dec 2017 14:08:21 +0000 (16:08 +0200)]
sys: Avoid extended ASCII.
Use UTF-8 where needed, use plain ASCII where possible.
Minor cleanup in dev/raid/arcmsr/arcmsr.h (for readability).
No functional change.
Sascha Wildner [Sun, 31 Dec 2017 13:53:38 +0000 (14:53 +0100)]
Remove various 'r' commands and daemons.
rcp(1), rlogin(1), rlogind(1), rsh(1) and rshd(1). In the unlikely case
that someone still needs them, they can use the net/bsdrcmds port to
bring them back.
FreeBSD removed them in October 2017 and I took their work as a guideline.
zrj [Sun, 31 Dec 2017 07:56:17 +0000 (09:56 +0200)]
lib/libc_rtld: Fix 'make install'.
Do not install nls categories and uuid(3) manpages. These are handled by libc.
Eitan Adler [Sun, 31 Dec 2017 07:03:29 +0000 (23:03 -0800)]
skel: Quote default alias for pager
Sascha Wildner [Sun, 31 Dec 2017 03:47:16 +0000 (04:47 +0100)]
mt(1): Staticize.
Sascha Wildner [Sun, 31 Dec 2017 03:40:39 +0000 (04:40 +0100)]
mkstr(1): Staticize.
Sascha Wildner [Sun, 31 Dec 2017 03:30:52 +0000 (04:30 +0100)]
m4(1): Add some __printflike.
Sascha Wildner [Sun, 31 Dec 2017 03:11:16 +0000 (04:11 +0100)]
locate(1): Staticize.
Sascha Wildner [Sun, 31 Dec 2017 03:00:29 +0000 (04:00 +0100)]
localedef(1): Add some __printflike.
Sascha Wildner [Sun, 31 Dec 2017 02:34:15 +0000 (03:34 +0100)]
locale(1): Staticize and remove an unused variable.
Sascha Wildner [Sun, 31 Dec 2017 02:30:05 +0000 (03:30 +0100)]
leave(1): Staticize.
Sascha Wildner [Sun, 31 Dec 2017 02:25:19 +0000 (03:25 +0100)]
gencat(1): Staticize and remove a wrong prototype.
zrj [Thu, 28 Dec 2017 12:30:07 +0000 (14:30 +0200)]
libc: Remove 'table' argument from call to collate_range_cmp() in __sccl().
This prevents segfaults when called with say incomplete locale.
Introduce a separate wide variant of __wcollate_range_cmp() and use proper
variant were needed. The __sccl() does not use wchar_t types.
Also use '\0' instead of plain 0.
Similar change was done in FreeBSD.
zrj [Thu, 28 Dec 2017 11:34:53 +0000 (13:34 +0200)]
libc/locale: Fix leading space issue in wcstof().
On failure correctly set endptr to report failure.
It was reported by libc++ testsuite.
Taken-from: FreeBSD
zrj [Sat, 30 Dec 2017 18:32:25 +0000 (20:32 +0200)]
newlocale(3): Add support for native complex locale.
Now newlocale(LC_ALL_MASK, setlocale(LC_ALL, NULL), NULL) should work
as intended even with combinations like LC_CTYPE=lv_LT LC_TIME=lt_LT.
While there, add compat support in setlocale(3) and fix a wrong git am
in previous newlocale(3) commit.
Eitan Adler [Fri, 29 Dec 2017 23:14:52 +0000 (15:14 -0800)]
bsd-family-tree: Update to latest
- add NetBSD 6.06, 7.1.1
- add HardenedBSD 10, 11
- correct 'NDB' to 'NBD'
Obtained-from: FreeBSD
Sascha Wildner [Fri, 29 Dec 2017 19:24:10 +0000 (20:24 +0100)]
shutdown(8): Fix building with -DDEBUG.
Sascha Wildner [Fri, 29 Dec 2017 19:20:51 +0000 (20:20 +0100)]
csplit(1): Staticize.
zrj [Fri, 29 Dec 2017 13:51:32 +0000 (15:51 +0200)]
newlocale(3): Add compat support for c++ runtime libs.
Implement simple 'private' env support that mimics the internal
__get_locale_env(). Not to be used in a context of C.
While there, fix off by one in array index (_LC_LAST vs XLC_LAST).
zrj [Thu, 28 Dec 2017 11:05:18 +0000 (13:05 +0200)]
libc/locale: Prevent bad builds on partial XLC_ enum changes.
Add static asserts and document them.
Taken-from: FreeBSD
zrj [Thu, 28 Dec 2017 11:01:36 +0000 (13:01 +0200)]
mbtowc(3): Set EILSEQ on incomplete character case.
There are several cases where EILSEQ is already handled, yet not handled
in other cases. Make it consistent.
In-discussion-with: swildner
Taken-from: OpenBSD
zrj [Fri, 29 Dec 2017 13:37:16 +0000 (15:37 +0200)]
libc/locale: Misc cleanup.
Mostly whitespace, use NULL and fix typo.
Sascha Wildner [Fri, 29 Dec 2017 12:08:30 +0000 (13:08 +0100)]
crunchgen(1)/crunchide(1): Staticize.
Sascha Wildner [Fri, 29 Dec 2017 11:51:09 +0000 (12:51 +0100)]
compress(1): Staticize.
Sascha Wildner [Fri, 29 Dec 2017 11:44:09 +0000 (12:44 +0100)]
chat(8): Staticize.
Sascha Wildner [Fri, 29 Dec 2017 11:32:49 +0000 (12:32 +0100)]
cap_mkdb(1): Staticize and remove an unused variable.
Sascha Wildner [Fri, 29 Dec 2017 11:28:03 +0000 (12:28 +0100)]
c99(1): Staticize.
Sascha Wildner [Fri, 29 Dec 2017 11:25:16 +0000 (12:25 +0100)]
btpin(1): Staticize.
Sascha Wildner [Fri, 29 Dec 2017 11:21:26 +0000 (12:21 +0100)]
basename(1): Staticize.
Sascha Wildner [Thu, 28 Dec 2017 21:57:41 +0000 (22:57 +0100)]
Add __attribute__((__noreturn__)) to various function prototypes in games/.
This general cleanup is useful in various ways. It helps the compiler
optimize better and it also helps later GCCs detect better when to give
-Wimplicit-fallthrough warnings and when not.
Found-with: -Wsuggest-attribute=noreturn
Sascha Wildner [Thu, 28 Dec 2017 21:50:37 +0000 (22:50 +0100)]
wump(6): Staticize.
Sascha Wildner [Thu, 28 Dec 2017 21:42:34 +0000 (22:42 +0100)]
worms(6): Staticize.
Sascha Wildner [Thu, 28 Dec 2017 21:36:51 +0000 (22:36 +0100)]
worm(6): Staticize and remove some "return" (crash() never returns).
Sascha Wildner [Thu, 28 Dec 2017 20:53:47 +0000 (21:53 +0100)]
sail(6): pl_main() never returns, so don't pretend it does.
Sascha Wildner [Thu, 28 Dec 2017 20:34:42 +0000 (21:34 +0100)]
pig(6): Staticize.
Sascha Wildner [Thu, 28 Dec 2017 20:24:57 +0000 (21:24 +0100)]
number(6): Staticize and remove a wrong prototype.
Sascha Wildner [Thu, 28 Dec 2017 20:09:26 +0000 (21:09 +0100)]
hangman(6): Move __dead2 to the prototype.
Sascha Wildner [Thu, 28 Dec 2017 19:40:08 +0000 (20:40 +0100)]
arithmetic(6): Staticize.
Sascha Wildner [Thu, 28 Dec 2017 09:54:35 +0000 (10:54 +0100)]
iso639: Add Montenegrin.
See http://www.loc.gov/standards/iso639-2/php/code_changes.php
Eitan Adler [Thu, 28 Dec 2017 08:19:34 +0000 (00:19 -0800)]
Update the pciconf(8) database.
2017-12-20 snapshot from https://pci-ids.ucw.cz/v2.2/pci.ids
zrj [Wed, 27 Dec 2017 16:10:40 +0000 (18:10 +0200)]
libc/nmalloc: Handle size_t overflows in malloc().
Similar to
e9586122ca8be1e743aa9cd2f9622dfe2f434ece.
zrj [Wed, 27 Dec 2017 06:31:35 +0000 (08:31 +0200)]
kerne: Move vfs.pbuf_{kva,mem,raw}_count sysctls to vm node.
The vm.pbud_kva_count, vm.pbud_mem_count and vm.pbud_raw_count
matches other sysctls in sys/vm/.
No userland usage was detected in base and dports.
Eitan Adler [Wed, 27 Dec 2017 13:07:25 +0000 (05:07 -0800)]
Fix several speelling errors
Reviewed by: swildner, zrj
zrj [Wed, 27 Dec 2017 06:28:33 +0000 (08:28 +0200)]
kernel: Remove 3 no longer used sysctls in vfs_bio.c
Sascha Wildner [Tue, 26 Dec 2017 13:04:58 +0000 (14:04 +0100)]
passwd.1/security.7: Some kerberos removal cleanup.
Sepherosa Ziehau [Tue, 26 Dec 2017 04:28:39 +0000 (12:28 +0800)]
socket: Use fast close if SO_LINGER and linger==0.
It is widely used to reset TCP connection by userland application.
Sascha Wildner [Tue, 26 Dec 2017 12:15:19 +0000 (13:15 +0100)]
Makefile.inc1: Adjust an obsolete comment and fix trailing whitespace.
Submitted-by: zrj
zrj [Tue, 26 Dec 2017 09:23:49 +0000 (11:23 +0200)]
ls(1): Add -_ option to include nanotime in ls -l.
Very useful when debugging Makefile races in rules with
vfs.timestamp_precision=1 set.
Yes, ls -l_ seems strage, but we are running out of letters and
FreeBSD already has the "-," flag in their ls(1).
zrj [Tue, 26 Dec 2017 09:22:04 +0000 (11:22 +0200)]
ls(1): Add -D option to set date-time format in ls -l.
Taken-from: FreeBSD
zrj [Tue, 26 Dec 2017 09:18:49 +0000 (11:18 +0200)]
ls(1): Sort flags by order in usage() and manpage.
Add missing 'S' in SYNOPSIS.
While there, move out trailing space separator handling to caller.
zrj [Mon, 25 Dec 2017 13:31:29 +0000 (15:31 +0200)]
ls(1): Minor cleanup.
No functional change.
zrj [Mon, 25 Dec 2017 12:47:52 +0000 (14:47 +0200)]
libc/stdtime: Use style(9) comments in strftime.c
Sascha Wildner [Mon, 25 Dec 2017 10:35:58 +0000 (11:35 +0100)]
Cleanup usr.sbin's Makefile a bit.
Sascha Wildner [Mon, 25 Dec 2017 01:48:05 +0000 (02:48 +0100)]
Remove obsolete comment from libexec's Makefile.
zrj [Sun, 24 Dec 2017 18:54:51 +0000 (20:54 +0200)]
fstyp(8): Don't hardcode UFS1 magic.
Suggested-by: swildner
Eitan Adler [Sun, 24 Dec 2017 18:34:06 +0000 (10:34 -0800)]
splitpatch: massage to function under python3
zrj [Sun, 24 Dec 2017 17:06:11 +0000 (19:06 +0200)]
.gitattributes: Use C language rules in git diff driver.
Solves a very annoying issue when goto labels are mistaken for function name.
Enable only on stuff we care: *.c and *.h. Remember, we had our fun with eof.
Approved-by: swildner
zrj [Sun, 24 Dec 2017 10:01:04 +0000 (12:01 +0200)]
fstyp(8): Remove direct dep on sys/boot/ headers.
Use the kernel headers with local additions to understand ufs2.
zrj [Sat, 23 Dec 2017 18:14:44 +0000 (20:14 +0200)]
boot: Remove unused loader.rc
We use dloader.rc and certainly do not have a loader.4th.
The loader.rc is not put into Makefile_upgrade.inc because we have not
detected it on any of setups and ficl removal was done back in 2010 in
116b7e3c174090f2c3e2f98e4d71f482b50700db