Bill Yuan [Sun, 8 Apr 2018 07:50:55 +0000 (07:50 +0000)]
ipfw3nat: RB_tree to keep the translation records
Bill Yuan [Sun, 8 Apr 2018 07:37:54 +0000 (07:37 +0000)]
ipfw3nat: cleanup 2
Bill Yuan [Sun, 8 Apr 2018 07:04:17 +0000 (07:04 +0000)]
ipfw3nat: delete the libalias 3
Bill Yuan [Sun, 8 Apr 2018 07:00:22 +0000 (07:00 +0000)]
ipfw3nat: delete the libalias 2
Bill Yuan [Sun, 8 Apr 2018 06:54:50 +0000 (06:54 +0000)]
ipfw3nat: delete the libalias 1
Bill Yuan [Sun, 8 Apr 2018 06:41:43 +0000 (06:41 +0000)]
ipfw3nat: cleanup
Matthew Dillon [Sun, 8 Apr 2018 01:06:08 +0000 (18:06 -0700)]
kernel - Change AcpiOs*Lock() operation to reduce freezes
* Rework the AcpiOs*Lock() functions to use a lockmgr lock + critical
section instead of a spin lock. There are situations where the
ACPI contrib code is holding one of these locks and then issues
high level calls that might block.
* When called from the idle thread, AcpiOs*Lock() uses a separate
spin-lock as before. The idle thread of course is not allowed to
block (except for HLT related activities). This lock will be
independent of the lockmgr lock, so we are assuming that cpu->idle
operation is independent from ACPI execution on other cpu threads.
* Fixes at least one panic, and also possibly lockups here and
there.
Matthew Dillon [Sun, 8 Apr 2018 01:03:23 +0000 (18:03 -0700)]
kernel - Fix pageout / sbusy race
* For now, vm_page_sbusy_try() cannot safely increment m->busy_count,
and then decrement it if it turns out to be busy. This interferes
with any other thread that is holding the page hard-busy and expects
the soft-busy mask to remain 0.
* Eventually we should be able to rework the fetchadd optimization back
in. Use atomic_cmpset_*() for now though.
Sascha Wildner [Sat, 7 Apr 2018 22:03:47 +0000 (00:03 +0200)]
ufs: Rename a number of UFS constants throughout the tree.
ROOTINO -> UFS_ROOTINO
WINO -> UFS_WINO
NDADDR -> UFS_NDADDR
NIADDR -> UFS_NIADDR
MAXSYMLINKLEN -> UFS1_MAXSYMLINKLEN
This helps to better sepate them from identically named EXT2 constants
which I'll rename in a later commit.
It also helps with my makefs(8) port.
I went with FreeBSD's names, so it is UFS1_MAXSYMLINKLEN even though
we'll probably never have UFS2, but since they are in public headers,
some stuff in dports might benefit from it in the future.
For safety, bump <sys/param.h> too.
Sascha Wildner [Sat, 7 Apr 2018 19:37:05 +0000 (21:37 +0200)]
getty(8): Sync with FreeBSD.
* Change gettytab's 'he' capability to accept a POSIX extended regexp
to mess with the system's hostname better (r313107).
* Add a new 'iM' capability to specify a program to be run that
generates the banner message displayed before the login prompt.
This extends the functionality of 'im'.
* Do not loop indefinitely if the device node doesn't exist (think
USB serial adapters e.g.).
* Various bug fixes and minor issue adjustments.
Sascha Wildner [Sat, 7 Apr 2018 10:48:56 +0000 (12:48 +0200)]
Remove usage of syslog(3)'s LOG_ODELAY because it is the default.
Sascha Wildner [Fri, 6 Apr 2018 06:34:19 +0000 (08:34 +0200)]
sdp.3: Adjust protos to follow the code.
Sascha Wildner [Fri, 6 Apr 2018 06:25:08 +0000 (08:25 +0200)]
ypclnt.3: Adjust yp_first() and yp_next() protos to follow the code.
Sascha Wildner [Fri, 6 Apr 2018 06:18:01 +0000 (08:18 +0200)]
ioctl.9: Remove ioctl() from SYNOPSIS, it's described in ioctl.2.
Sascha Wildner [Fri, 6 Apr 2018 06:10:19 +0000 (08:10 +0200)]
zone.9: Update a bit better.
* zinitna() doesn't take a vm_object anymore (see
e16c650d787872d60f6ca),
so this commit reverts
35f324174f2040a627e18e8b6f008e7a37d818e9.
* Use .Fn instead of .Fa for functions.
* Remove old documentation about the zone functions not being SMP-safe.
Sascha Wildner [Thu, 5 Apr 2018 17:05:45 +0000 (19:05 +0200)]
pax(8): Remove old prototypes.
Sascha Wildner [Thu, 5 Apr 2018 15:47:45 +0000 (17:47 +0200)]
zone.9: Include <vm/vm_object.h> for struct vm_object.
Sascha Wildner [Thu, 5 Apr 2018 07:36:00 +0000 (09:36 +0200)]
pax(8): Switch to using libc's pwcache instead of a local implementation.
Based-on: NetBSD
Sascha Wildner [Wed, 4 Apr 2018 19:00:50 +0000 (21:00 +0200)]
libc/pwcache: Some further work on the pwcache import.
* Add prototypes
* Adjust Symbol.map and Versions.def
* Add MLINKS
* Bump .Dd in the manual page to the date it came into our tree
François Tigeot [Wed, 4 Apr 2018 17:37:42 +0000 (19:37 +0200)]
Update pwcache(3) support files from NetBSD
* Bringing the API up-to-date with NetBSD and FreeBSD
* This effectively adds the uid_from_user() and gid_from_group()
functions, the reverse of user_from_uid() and groups_from_gid()
* (void) casts for discarded return values have been removed
Sascha Wildner [Wed, 4 Apr 2018 09:04:10 +0000 (11:04 +0200)]
rmt(8): Remove unneeded include of <sgtty.h>.
Sascha Wildner [Wed, 4 Apr 2018 07:54:24 +0000 (09:54 +0200)]
<sys/ioctl_compat.h>: Add a reminder comment about IOCTLTRIM.
This ioctl really doesn't belong in this header. Unfortunately, it is
used by the benchmarks/fio port, so we can't fix it right away.
Sascha Wildner [Wed, 4 Apr 2018 07:45:04 +0000 (09:45 +0200)]
ioctl.9: Document that 'C' is used by some cam(4) ioctls as well.
Sascha Wildner [Tue, 3 Apr 2018 21:50:31 +0000 (23:50 +0200)]
libcompat: Remove the old and unused insque() and remque() functions.
We have a conforming implementation of those two functions in libc, so
nothing was taking the ones from libcompat. However, libcompat's
outdated manual pages were shadowing the good ones from libc.
Reported-by: zrj
Sascha Wildner [Sun, 1 Apr 2018 18:13:31 +0000 (20:13 +0200)]
kernel/vtnet: Add opt_ifpoll.h to the Makefile.
Sascha Wildner [Sun, 1 Apr 2018 12:24:15 +0000 (14:24 +0200)]
Sync zoneinfo database with tzdata2018d from ftp://ftp.iana.org/tz/releases
* Palestine started DST a week earlier, on March 24.
* Various changes to past time stamps and zone abbreviations.
For a detailed list of changes, see share/zoneinfo/NEWS.
Sascha Wildner [Sun, 1 Apr 2018 11:22:51 +0000 (13:22 +0200)]
re_format.7: Make regex.7 an MLINK.
Suggested-by: aly
Sascha Wildner [Sun, 1 Apr 2018 10:54:05 +0000 (12:54 +0200)]
man(1): Make -a default.
Sascha Wildner [Fri, 23 Mar 2018 09:19:01 +0000 (10:19 +0100)]
Fix -Wsizeof-pointer-memaccess better.
My previous commits:
56267d362d5769c8df07bf26d5e322610e0d24b4
562a180a58780b719830629307e6283c30c85229
caused out of bounds write by one byte. This commit fixes the
warning and restores the code what was there before.
Sascha Wildner [Fri, 30 Mar 2018 09:04:48 +0000 (11:04 +0200)]
kernel/vkernel: Add __printflike() to two internal functions, for -Wformat.
Aaron LI [Sun, 1 Apr 2018 06:28:45 +0000 (14:28 +0800)]
loader/menu: Fix 'vfs.root.real_root' to be 'vfs.root.realroot'
This should be the same variable as used in /boot/loader.conf to specify
the real root partition. See loader(8) and mkinitrd(8) for more info.
The intent here by unsetting it for the *rescue boot mode* is to prevent
the real root been mounted, and hence the system will stay in the md0
rescue root.
Verified-by: Matt Dillon
Aaron LI [Sun, 21 Jan 2018 07:16:36 +0000 (15:16 +0800)]
cpdup(1): Fix a typo in the usage
Matthew Dillon [Sun, 1 Apr 2018 05:02:22 +0000 (22:02 -0700)]
hammer2 - Increase bulkfree buffer size request and cap
* Increase the kernel buffer size cap for hammer2 cleanups
(aka bulkfree's) from 64MB to 1/16 physical memory. Systems
with more memory can accomodate larger in-memory bulkfree bitmap
buffers. Systems with less memory simply do multiple scan
passes as per usual.
* Increase the buffer size requested by the hammer2 utility
from 8MB to 1/16 physical memory. The actual amount of memory
allocated by the kernel will be 32MB per 1TB of filesystem, up to
the cap.
* This will generally reduce the number of scans required for
a 4TB filesystem from 16 to 1, meaning that bulkfree will take
1/16 the time as it did before for a filesystem of that size.
For example, a system with 128GB of ram would be able to
request up to 8GB of buffer which can accomodate a bulkfree
operation on a 256TB filesystem in a single pass.
Aaron LI [Sat, 31 Mar 2018 02:37:53 +0000 (10:37 +0800)]
mkinitrd(8): Fix a wrong variable name, and more improvements
* Fix the wrong variable name 'CONTENT_DIR' to be 'CONTENT_DIRS', which is
used in the configuration file as well as this tool.
* Add a function check_dirs() to check the existence of directories.
* Append the PID to the temporary build directory to make it unique.
* Remove the temporary build directory after use.
* Backup the old initrd.img.gz before copy the new one over.
Imre Vadász [Thu, 29 Mar 2018 21:34:20 +0000 (23:34 +0200)]
if_vtnet - Add polling support.
* We need to tear down the interrupts, to reliably stop them from firing,
when enabling polling. So when disabling polling, we need to hope that
we can get all interrupts set up again, as before.
* When polling, handle the rx and tx virtqueues on the same cpu.
Imre Vadász [Thu, 29 Mar 2018 19:23:10 +0000 (21:23 +0200)]
if_vtnet - Factor out the virtio_bind_intr() calls into a separate function.
* And most importantly, store the bindings in the vtnet_softc struct, so we
can restore the interrupt <-> virtqueue bindings when disabling polling
mode.
Aaron LI [Fri, 30 Mar 2018 08:30:49 +0000 (16:30 +0800)]
mkinitrd(8): improve the man page and styles
Suggested-by: swildner
Aaron LI [Fri, 30 Mar 2018 06:32:59 +0000 (14:32 +0800)]
mkinitrd(8): determine the initrd size from the contents size
* Calculate the contents size to determine the required initrd size.
The used initrd size is determined from the total contents size, rounded up
to align with 1 MiB, added additional 1 MiB.
The calculated initrd size is 12 MiB at the moment, so we have 3 MiB
more usable memory now (compared to the fixed 15 MiB before).
* Add options "-s" to specify the required initrd size as before. But a check
against the above determined required initrd size is performance to avoid the
out-of-space error, resulting in a broken initrd image.
* Add option "-S" to constrain the allowed maximum initrd size, because a too
big initrd image will also lead to boot failure, due to the boot loader
limitation.
* Some minor cleanups and improvements, better error handle and messages.
* Update the mkinitrd.conf and man page accordingly.
Reviewed-by: zrj
zrj [Fri, 30 Mar 2018 06:22:21 +0000 (09:22 +0300)]
kernel/usb4bsd: Add missing descriptions for C610/X99 controllers.
Codename Wellsburg.
While there, fix few typos and match entries in share/misc/pci_vendors.
Matthew Dillon [Fri, 30 Mar 2018 05:12:35 +0000 (22:12 -0700)]
kernel - Fix UFS bug on filesystem full conditions
* UFS was not properly translating the lbn to a block offset
when trying to back-out an indirect block allocation during
a write() which fails due to lack of space.
* Fixes a panic which could occur in the above situation.
Reported-by: Aaron LI
Aaron LI [Thu, 29 Mar 2018 14:12:33 +0000 (22:12 +0800)]
rc.d: Add ipfw3 rc script
To use ipfw3, add 'ipfw3_enable=YES' to /etc/rc.conf . By default, only
the 'ipfw3' and 'ipfw3_basic' modules will be loaded. If you need more
ipfw3 modules, set them with 'ipfw3_modules="<modules> ..."' in
/etc/rc.conf . See also the rc.conf(5) man page.
This ipfw3 rc script is largely based on the ipfw script.
The /etc/ipfw3.rules (a shell script) should be present to setup the
firewall rules, otherwise the default setup will block all incoming
connections, including your current connection!
NOTE:
I wrote this script for use in my VPS last year, but switched to
PF later, so more tests and improvements (e.g., flush tables on stop)
to this script are needed.
Reviewed-by: swildner
Aaron LI [Tue, 27 Mar 2018 15:52:35 +0000 (23:52 +0800)]
rc.subr: Add _rc_subr_loaded to avoid sourcing it again
* Once /etc/rc.subr been sourced, $_rc_subr_loaded="YES", then /etc/rc.subr
takes care of avoiding sourcing it again. (obtained from FreeBSD)
* Remove the unnecessary "$_rc_subr_loaded" prefix from several rc scripts,
which also make the "$_rc_subr_loaded . /etc/rc.subr" command fail.
Aaron LI [Thu, 15 Mar 2018 14:30:32 +0000 (22:30 +0800)]
rcorder(8): Multiple cleanups
* Add static for rcorder.c (obtained from FreeBSD)
* Remove commented lines (commit:
70c6c3571b5250b7cabfad3d20d4f7c09a5aadaf)
* Describe function arguments in hash.c
* Fix several typos (obtained from FreeBSD)
* Break two long lines
* Clean up whitespaces
Imre Vadász [Thu, 29 Mar 2018 20:10:07 +0000 (22:10 +0200)]
virtio_pci - Fix vtpci_reinit MSI-X vector reinitialization.
Imre Vadász [Thu, 29 Mar 2018 22:33:03 +0000 (00:33 +0200)]
if_vtnet - Fixup serialized assertion in vtnet_is_link_up().
* We check vtnet_is_link_up() from the config interrupt, which runs with
just the sc->vtnet_slz serializer.
Matthew Dillon [Wed, 28 Mar 2018 22:09:18 +0000 (15:09 -0700)]
bridge - Fix dereference race
* bri->bri_dead can be set asynchronously vs gc scans, catching a
scan in the middle.
* Add a ref count, bri_refs, to be able to properly determine when
we can kfree() the bri. Before we kfree()'d it in the last chained
netisr, but that had the above race.
Matthew Dillon [Wed, 28 Mar 2018 00:41:04 +0000 (17:41 -0700)]
bridge - Fix arp issue
* Fix an issue when an arp route is associated with a bridge. When an
arp request comes in on an interface that is not destined for us, do
not reply to it if the interface belongs to the bridge the arp route
is associated with.
Replying to such arp requests would result in a double-reply, one from
the bridge MAC and one from the target MAC.
* The request will be relayed over the bridge normally, as intended.
Imre Vadász [Wed, 28 Mar 2018 20:46:14 +0000 (22:46 +0200)]
iwm - Always clear watchdog timer, when bringing down firmware state.
Imre Vadász [Wed, 28 Mar 2018 20:26:18 +0000 (22:26 +0200)]
iwm - Clear Time Event active state, when receiving End Notification.
* This hopefully avoids some firmware panics, I was occasionally seeing,
when iwm disconnects upon losing signal to an access point at some point.
* This is synchronizing the if_iwm_time_event.c file a bit more from the
corresponding Linux iwlwifi/mvm/time-event.c.
Taken-From: Linux iwlwifi
Imre Vadász [Sun, 10 Sep 2017 20:27:50 +0000 (22:27 +0200)]
vtnet - Split up serializers, now has separate rx and tx serializers.
* This uses the MULTI SERIALIZERS MODE in the DragonFly net framework.
* This allows the rx and tx paths of if_vtnet to run in parallel.
François Tigeot [Wed, 28 Mar 2018 06:07:04 +0000 (08:07 +0200)]
sys/file.h: Expand struct file for drm/linux shim
The struct file.private_data field will be used to pass drm_file information,
removing the need for some of the existing devfs cdevpriv code.
Distinct commit suggested-by: zrj
Aaron LI [Wed, 28 Mar 2018 04:39:01 +0000 (12:39 +0800)]
Makefile.usr: allow easier change of repo URL and cleanups
* Introduce variables GITURL_SRC and GITURL_DPORTS to replace GITHOST and
GITHOST_DPORTS, respectively, allowing easier change the git repo URLs, e.g.,
changing to use the HTTPs clone from GitHub.
* Introduce variable BRANCH to specify the src branch to be checked out, either
"master" (the development branch) or "DragonFly_RELEASE_x_y" (for stable
releases). Also print the checkout branch in the help.
* Simplify "src-create" and "src-create-shallow" targets to a one-line git clone
command with the --depth and --branch options.
* Target "src-create-repo" is merged into "src-create", and target
"src-checkout" is removed as it is no longer needed.
* Protect "release-sys-extract" target with the existence check of
src-sys.tar.bz2 .
* Other minor updates and cleanups.
Justin C. Sherrill [Tue, 27 Mar 2018 22:41:29 +0000 (18:41 -0400)]
Changes for DragonFly 5.3.
Imre Vadász [Tue, 27 Mar 2018 18:40:14 +0000 (20:40 +0200)]
if_vtnet - Cleanup ifalt_subque handling a bit.
Sascha Wildner [Tue, 27 Mar 2018 18:39:48 +0000 (20:39 +0200)]
kernel/{ext2,u}fs: Remove unused 'in_exists' from struct indir.
Imre Vadász [Mon, 26 Mar 2018 21:44:30 +0000 (23:44 +0200)]
if_vtnet - Get rid of separate vtnet_{rx,tx}_intr_task functions.
* Just inline them into vtnet_rx_vq_intr and vtnet_tx_vq_intr respectively.
Imre Vadász [Mon, 26 Mar 2018 20:56:29 +0000 (22:56 +0200)]
if_vtnet - Fixup previous commit.
Imre Vadász [Sun, 25 Mar 2018 23:17:40 +0000 (01:17 +0200)]
if_vtnet - Use ifsq_watchdog_* functions as the watchdog.
* The existing watchdog code was commented out.
Tomohiro Kusumi [Mon, 26 Mar 2018 15:12:59 +0000 (00:12 +0900)]
sys/vfs/autofs: Cleanups (sync with NetBSD and OpenBSD)
FreeBSD has these in autofs.c, but only vfsops needs init/uninit.
Aaron LI [Sat, 10 Mar 2018 04:12:11 +0000 (12:12 +0800)]
locate.updatedb(8): improve default search paths with small updates
* Make locate.updatedb(8) default to search from all mount points with
the allowed filesystem types. Therefore, it is allowed to exclude the
root filesystem type, while still correctly index other mounted
filesystem with allowed types. Otherwise, the whole filesystem is
simply ignored if the root filesystem type is excluded.
* Although we can now ignore the root filesystem type from being indexed,
but this is generally a configuration mistake, so we warn about this.
* Use stat(1) instead of the obscure "find -c" to help check whether the
created database is empty.
* Print out the full find command by default (to help config/debug), also print
a message when the database is successfully built.
* Update locate.rc accordingly.
* Explain why use cat instead of cp for the final database creation.
Imre Vadász [Sun, 25 Mar 2018 19:59:43 +0000 (21:59 +0200)]
if_vtnet - Don't drop vtnet_slz in some of the virtqueue_notify() calls.
* There are several cases where the vtnet_slz serializer is dropped around
virtqueue_notify() calls, with the hope that this allows other cores to
run if_vtnet code in parallel.
* In vtnet_exec_ctrl_cmd(), it shouldn't make a significant performance
difference, and avoids some theoretical race possibilities.
* In the vtnet_init_rx_vq initialization code, it also is irrelevant for
performance, and only complicates things.
* In the vtnet_rxeof() and vtnet_start() cases it might actually be relevant
for performance, so we keep those for now. When changing the driver to
use the MULTI_SERIALIZERS MODE, the serializer dropping can also be
removed in those two cases.
Imre Vadász [Sun, 25 Mar 2018 19:05:04 +0000 (21:05 +0200)]
if_vtnet - Remove vtnet_tick_ch and vtnet_cfgchg_task from softc struct.
* Forgot to remove the softc variables in the changes
c5f8e1538bcc13f7a39b805bcc5b774780 and
603be9646446bdb4292b6cd27544788c2c.
* Also remove a leftover taskqueue_drain() call.
Imre Vadász [Sun, 25 Mar 2018 18:55:23 +0000 (20:55 +0200)]
if_vtnet - Pass the vtnet serializer in ether_ifattach().
* This is likely just an intermediate step, the plan is to eventually use
the MULTI SERIALIZER MODE.
Imre Vadász [Sun, 25 Mar 2018 18:54:41 +0000 (20:54 +0200)]
if_vtnet - Remove unused vtnet_tick() code.
Imre Vadász [Sun, 25 Mar 2018 14:35:18 +0000 (16:35 +0200)]
if_vtnet - No need to use the taskqueue to run vtnet_update_link_status().
Eitan Adler [Sun, 25 Mar 2018 17:18:02 +0000 (17:18 +0000)]
[pci_vendors] Update pci_vendors to 2018.03.21
Aaron LI [Sun, 25 Mar 2018 12:27:26 +0000 (20:27 +0800)]
[test] Hi Fred :-)
François Tigeot [Sun, 25 Mar 2018 08:34:02 +0000 (10:34 +0200)]
drm/linux: Add signal_pending_state
Imre Vadász [Thu, 22 Mar 2018 23:08:43 +0000 (00:08 +0100)]
virtio - Get rid of unused flags option in virtio_alloc_virtqueues().
Imre Vadász [Sat, 24 Mar 2018 13:44:19 +0000 (14:44 +0100)]
virtio - Explicitly teardown interrupts in driver detach methods.
* This seems to be an easy way to avoid potential races during driver
detachements.
Imre Vadász [Wed, 13 Sep 2017 19:59:25 +0000 (21:59 +0200)]
virtio - Move virtqueue_pending() checks into the child driver.
* This makes it easier for now, to deal with virtqueue locking, when
multiple virtqueues are mapped to a single interrupt vector.
Imre Vadász [Tue, 12 Sep 2017 22:32:22 +0000 (00:32 +0200)]
virtio - Get rid of config_change devmethod, treat like vq interrupts.
Imre Vadász [Mon, 11 Sep 2017 20:59:58 +0000 (22:59 +0200)]
virtio - Specify handler and argument in bind_intr, instead of vq alloc.
Imre Vadász [Thu, 7 Sep 2017 22:22:49 +0000 (00:22 +0200)]
virtio - Refactor virtio bus API, to untangle virtuques from IRQ allocation.
* These API changes allow each virtio device driver to explicitly control
the assignment of virtqueues to IRQ vectors.
* This allows explicitly assigning IRQ vectors to cpus. And virtio device
drivers also get to know, on which cpu each virtqueue interrupt is
running.
Sascha Wildner [Sat, 24 Mar 2018 14:15:34 +0000 (15:15 +0100)]
groff/tmac: Revert doc-common back to vendor state.
Missing .Fx etc. versions should be added to mdoc.local instead, so
add any which were not in it yet.
While here, fix a typo from
12041c96816e1e7cccac302fd5656529fe9c9e9b
and add NetBSD 9.0 for fstyp.8.
Tomohiro Kusumi [Sat, 24 Mar 2018 09:18:00 +0000 (18:18 +0900)]
sys/vfs/autofs: Revert
1bfb99b1b1("prevent assert on unmount")
1bfb99b1b1 originally brought from NetBSD was not correct.
Autofs can't unmount when more than one level of directories
are created via indirect map.
François Tigeot [Sat, 24 Mar 2018 07:21:31 +0000 (08:21 +0100)]
drm/linux: Add cmpxchg_relaxed()
Obtained-from: FreeBSD
Matthew Dillon [Fri, 23 Mar 2018 23:18:22 +0000 (16:18 -0700)]
kernel - Fix namecache leak / broken hysteresis
* We were double-using NCF_DESTROYED for both dummy ncp's for
list iteration and to indicate a removed ncp. This prevented
the hysteresis code from properly cleaning up such ncp's.
Symptoms were growing stalls every few seconds on one cpu as
the hysteresis code would try to futiley clean up excess
namecache entries.
* Use NCF_DUMMY instead of NCF_DESTROYED to check for dummy
list iterators in the hysteresis code, allowing it to process
NCF_DESTROYED ncps.
* These ncps could accumulate only in situations where the vnode
nlinks count is greater than 1, since a file deletion when nlinks
is 1 will delete the vnode and all related ncp's.
When nlinks is greater than 1, a file entry deletion does not
necessarily cause the vnode to be deleted, leaving the namecache
records intact, but flagged NCF_DESTROYED, and prevented the
hystersis code from operating correctly.
Reported-by: ftigeot
Sascha Wildner [Fri, 23 Mar 2018 20:15:34 +0000 (21:15 +0100)]
kernel/machdep: Remove unused and #if 0'd finishidentcpu() prototype.
François Tigeot [Fri, 23 Mar 2018 14:09:48 +0000 (15:09 +0100)]
drm/linux: Add a few functions (mostly stubs)
Sascha Wildner [Fri, 23 Mar 2018 12:47:29 +0000 (13:47 +0100)]
vfs_unmountall.9: Document the new halting argument.
François Tigeot [Fri, 23 Mar 2018 09:59:33 +0000 (10:59 +0100)]
kernel: Add missing atomic_fcmpset_*() variants
They will be useful for newer drm code from Linux 4.8.
Sascha Wildner [Fri, 23 Mar 2018 09:30:18 +0000 (10:30 +0100)]
kernel/devfs: Use 1UL << ... in bitmap bit operations.
Sascha Wildner [Fri, 23 Mar 2018 08:15:41 +0000 (09:15 +0100)]
Remove obsolete stabs.7 manual page.
Imre Vadász [Sat, 11 Mar 2017 20:13:19 +0000 (21:13 +0100)]
virtio_blk - Inline vtblk_complete into vtblk_vq_intr. Dedup sglist_reset.
Imre Vadász [Mon, 6 Mar 2017 19:33:56 +0000 (20:33 +0100)]
virtio_pci - Remove int return value from interrupt handler functions.
Matthew Dillon [Thu, 22 Mar 2018 16:38:09 +0000 (09:38 -0700)]
kernel - devfs cleanup
* staticize two functions, remove *_fff() from manual pages (now
static).
* Adjust documentation.
* Rename internal function from devfs_clone_bitmap_resize() to
devfs_clone_bitmap_extend(). Document the slight weirdness in
the extent of the passed-in 'newchunks' parameter, but do not change
the API.
Submitted-by: Aaron LI, with added code comments by Matt
Matthew Dillon [Thu, 22 Mar 2018 16:09:50 +0000 (09:09 -0700)]
kernel - Fix tapN creation >= 32 units, fix pty issues >= 32 ptys (2)
* Also fix incorrect shift type in devfs_clone_bitmap_set().
Submitted-by: Aaron LI
Matthew Dillon [Thu, 22 Mar 2018 01:11:10 +0000 (18:11 -0700)]
kernel - Fix tapN creation >= 32 units, fix pty issues >= 32 ptys
* Fix ifconfig tapN create for N >= 32.
* devfs_clone_bitmap_chk(), which only if_tap uses, was returning
the wrong default value for unit numbers beyond the current
dynamic size of the bitmap.
* Both devfs_clone_bitmap_chk() and devfs_clone_bitmap_put()
were improperly using (1 << unit) instead of (1L << unit)
when masking the 64-bit (long) bitmap elements, resulting in
chaos for units >= 32.
* This also should fix pty issues >= 32 ptys.
Reported-by: Aaron Li
Sascha Wildner [Wed, 21 Mar 2018 21:02:46 +0000 (22:02 +0100)]
idr.9: Improve this manual page a bit.
Based on a patch submitted by Aaron LI <aly@aaronly.me>.
Matthew Dillon [Wed, 21 Mar 2018 18:22:14 +0000 (11:22 -0700)]
kernel - Fix quickhalt panic-on-reboot
* Fix an issue with the quickhalt feature that shortcuts the
unmounting of procfs, devfs, and tmpfs.
* Fixes a panic-on-reboot due to quickhalt still cleaning out
the vnode ops even though it leaves the vnodes intact. Don't
clean out the vnode ops.
Reported-by: zrj
Sascha Wildner [Wed, 21 Mar 2018 15:29:23 +0000 (16:29 +0100)]
BSD.include.dist: Use spaces to indent (see etc/mtree/README).
Bill Yuan [Wed, 21 Mar 2018 03:08:56 +0000 (03:08 +0000)]
ipfw3: use the headers in include folder
Bill Yuan [Wed, 21 Mar 2018 03:07:54 +0000 (03:07 +0000)]
ipfw3: move the headers into include folder
Eitan Adler [Wed, 21 Mar 2018 09:07:51 +0000 (09:07 +0000)]
[bsd-family-tree]: announce NetBSD 7.1.2
Sascha Wildner [Tue, 20 Mar 2018 20:49:36 +0000 (21:49 +0100)]
mmap.2: Remove reference to no longer existing BUGS section.
Sascha Wildner [Tue, 20 Mar 2018 09:11:06 +0000 (10:11 +0100)]
boot/common/fs.h: Add a common definition of CTASSERT().
Sascha Wildner [Tue, 20 Mar 2018 08:12:02 +0000 (09:12 +0100)]
<net/if_poll.h>: Add license. It was based on sys/net/if_poll.c.
Reported-by: Aaron LI <aly@aaronly.me>
Discussed-with: sephe
Matthew Dillon [Tue, 20 Mar 2018 01:39:21 +0000 (18:39 -0700)]
kernel - Attempt to fix high vnlru cpu use
* In certain situations _cache_cleanpos() gets into a livelock
of some sort. It is limited by a counter but the livelock creates
havoc when it happens.
* It appears that _cache_cleanpos() can get into a situation where
the head element of all available nchashtbl[] slots is not freeable
(e.g. due to having multiple refs). When this occurs, trailing
elements in the chained hash table cannot be recycled.
nchashtbl[] is typically very large, so this situation does not
happen very often. There are usually plenty of entries to pick
away at. However, it looks like situations can develop where
enough entries get into this masked state that the hysteresis is
unable to complete, resulting in a cpubound loop.
* Refactor the chained list from LIST to TAILQ and modify the
_cache_cleanpos() function to cycle ncp's to the end of the
list before trying to zap them. If the zap fails, the next
iteration will encounter a different head.
Reported-by: ftigeot
Imre Vadász [Mon, 19 Mar 2018 22:39:17 +0000 (23:39 +0100)]
atkbdc - Add additional ACPI PNP IDs specified by Microsoft for Keyboards.
These IDs are from Microsofts device IDs list at
http://download.microsoft.com/download/1/6/1/
161ba512-40e2-4cc9-843a-
923143f3456c/devids.txt
which is linked from http://www.uefi.org/PNP_ACPI_Registry.
This should fix some machines which failed to attach the PS2 Keyboard
after
131acb0380e93664f106117ebf8ba322a3d1b7c4, unless the acpi(4) module
was disabled.