Matthew Dillon [Tue, 6 Dec 2016 22:36:32 +0000 (14:36 -0800)]
hammer - Disallow modifying ioctls when filesystem is read-only
* Disallow modifying ioctls if the filesystem has been mounted read-only
or gone into read-only mode due to an I/O error.
* This is only a partial fix. There are still error-pathing problems
in numerous procedures, particularly the node locking code, that might
result in a token life-lock.
Reported-by: Peter Avalos
Matthew Dillon [Tue, 6 Dec 2016 22:34:24 +0000 (14:34 -0800)]
libc - Take care of minor buffer overrun in link_ntoa()
* Take care of a minor buffer overrun in link_ntoa(). It is unlikely
that any program produces the conditions required to trigger the
problem.
Taken-from: FreeBSD-SA-16:37.libc
Reported-by: swildner, zrj, others
Tomohiro Kusumi [Tue, 6 Dec 2016 18:19:14 +0000 (03:19 +0900)]
sbin/hammer: Fix direntry message in hammer recover
name could have already been free'd, so move it to the beginning.
Also enable it only on -v, just like inode/data rectype cases.
Tomohiro Kusumi [Tue, 6 Dec 2016 17:47:04 +0000 (02:47 +0900)]
sbin/hammer: Fix inode/data messages in hammer recover
Based on other printf messages where "file" indicates regfile,
the first one should be "inode" rather than "file" because it
could be both directory and regfile.
The second one could be "file" because it's for file data, but
just sync with the first format.
Tomohiro Kusumi [Tue, 6 Dec 2016 13:14:06 +0000 (22:14 +0900)]
sbin/hammer: Minor cleanup for hammer recover
The reason for moving "info.pfs_id = dict->pfs_id;" is because
PT_FIGURE only requires strlen of "PFS%05d" (max 65535), and
dict->pfs_id never changes during path lookup by design.
Matthew Dillon [Tue, 6 Dec 2016 18:13:11 +0000 (10:13 -0800)]
dntpd - Fix memory leak
* Every log line leaked a bit of memory. Fixed.
zrj [Tue, 6 Dec 2016 08:14:32 +0000 (10:14 +0200)]
vkernel: Add a dummy cpu_smp_stopped() function (unbreaks build).
Follow-up to
63cff0361caa40216fcb16f79855de833431274b
Matthew Dillon [Tue, 6 Dec 2016 00:49:04 +0000 (16:49 -0800)]
kernel - Increase worst-case maximum exec rate
* The pid reuse algorithm limits the maximum fork rate. This limit
was set too low. Increase the limit from 10000/sec to 100000/sec.
Currently our opteron maxes out at 43000/sec.
Note that with 999999 pids and a 10-second mandatory reuse time
floor there isn't much of a point increasing the limit beyond
100000/sec.
100,000/sec. Currently our opteron maxes out at around
43,000/sec (vfork/exec/wait3/exit of a small static binary).
* The domain reuse array was increased to 1MB to accomodate this
change. In addition, update the array in a cache-friendly manner.
* Modify test/sysperf/exec1 to take a nprocesses argument for the
timing run.
Matthew Dillon [Mon, 5 Dec 2016 23:26:46 +0000 (15:26 -0800)]
kernel - Remove unused process_exit and process_exec eventhandlers
* Remove these two eventhandlers. They are not used in DragonFly.
* Fixes an unnecessary global lock bottleneck in exec and exit.
Suggested-by: Mateusz Guzik (mjg_)
Matthew Dillon [Mon, 5 Dec 2016 23:07:43 +0000 (15:07 -0800)]
kernel - Spiff up locks a bit
* Do a little optimization of _spin_lock_contested(). The critical path
is able to avoid two atomic ops in the initialization portion of the
contested path.
* Optimize _spin_lock_shared_contested() to use atomic_fetchadd_long()
to add a shared-lock count instead of atomic_cmpset_long(). Shared
spinlocks are used heavily and this will prevent a lot of unnecessary
spinning when many cpus are using the same lock at the same time.
* Hold fdp->fd_spin across fdp->fd_cdir and fdp->fd_ncdir modifications.
This completes other work which caches fdp->fd_ncdir and avoids having
to obtain the spin-lock when the cache matches.
Discussed-with: Mateusz Guzik (mjg_)
Matthew Dillon [Mon, 5 Dec 2016 23:01:10 +0000 (15:01 -0800)]
kernel - Make kern_proc cache-friendly
* Make the proc_tokens[], allprocs[], allpgrps[], and allsessn[]
arrays cache-friendly by aggregating them into a cache-aligned
struct procglob.
* Doesn't do much for the token array, but should help
allprocs/allpgrps/allsessn scans whos structures were previously
8-byte aligned.
Tomohiro Kusumi [Sun, 4 Dec 2016 09:57:15 +0000 (18:57 +0900)]
sbin/hammer: Add hammer strip command
This command is inspired by hammer recover command, and does
opposite of what recover command does.
This command zero clears zone-8(B-Tree) big-blocks, zone-9(meta)
big-blocks, and then the whole volume header, except that volume
signature field is overwritten with "STRIPPED" instead of zeros.
After running, a filesystem is no longer mountable or recoverable
with hammer recover command. This command is also fast as it only
zero clears good enough ondisk data to make it unmountable and
unrecoverable.
Keep in mind that this command does _not_ zero clear user data.
Users would normally use a software designed to completely shred
a filesystem. This command is not designed to shred a filesystem.
The name "strip" gives better idea of what it really does than
using "shred"/etc.
-- example
# newfs_hammer -L TEST /dev/da1 /dev/da2 /dev/da3 > /dev/null
# mount_hammer /dev/da1:/dev/da2:/dev/da3 /HAMMER
# cd /HAMMER
# dd if=/dev/urandom of=./out bs=1M count=120000
120000+0 records in
120000+0 records out
125829120000 bytes transferred in 1766.417077 secs (
71234094 bytes/sec)
# cd
# umount /HAMMER
# hammer -f /dev/da1:/dev/da2:/dev/da3 strip
You have requested that HAMMER filesystem (TEST) be stripped
Do you really want to do this? [y/n] y
Stripping HAMMER filesystem (TEST) in 5 4 3 2 1.. starting destruction pass
8000000021000000
9000000021800000
800000019c000000
800000030c000000
800000047e000000
80000005f7000000
8000000767000000
80000008d8000000
8000000a51800000
8000000bc5000000
8000000d37800000
8000000ead000000
800000101e800000
8000001193000000
8000001304000000
8000001478800000
80000015ee000000
8000001760800000
80000018d1800000
8000001a47000000
8000001bb6000000
801000013c000000
/dev/da1
/dev/da2
/dev/da3
# mount_hammer /dev/da1:/dev/da2:/dev/da3 /HAMMER
mount: Invalid argument
mount_hammer: /dev/da1: Invalid volume signature
4445505049525453
Tomohiro Kusumi [Sun, 4 Dec 2016 09:26:22 +0000 (18:26 +0900)]
sbin/hammer: Make hammer_parsedevs() take open(2) flag
This is for the next commit.
No functional change.
Tomohiro Kusumi [Sun, 4 Dec 2016 14:44:00 +0000 (23:44 +0900)]
sbin/hammer: Add "[y/n]" before getyn()
Tomohiro Kusumi [Sun, 4 Dec 2016 09:21:17 +0000 (18:21 +0900)]
sbin/hammer: Fix recursively called hammer_parsedevs()
c2b74c42 had to change recursively called hammer_parsedevs() as well.
Matthew Dillon [Mon, 5 Dec 2016 17:25:56 +0000 (09:25 -0800)]
kernel - Remove debugging kprintf
* Remove the 'exit race handled' debugging kprintf.
Matthew Dillon [Mon, 5 Dec 2016 17:21:19 +0000 (09:21 -0800)]
kernel - Try to idle cpus when in panic()
* Try to use MONITOR/MWAIT to idle cpus while they are stopped in a panic(),
instead of hard-looping. This significantly reduces power consumption while
in a paniced state and is particularly helpful on laptops.
Reported-by: tuxillo
Matthew Dillon [Mon, 5 Dec 2016 17:15:44 +0000 (09:15 -0800)]
kernel - more kmalloc and nlookup performance optimizations
* Give the pcpu counters in struct malloc_type their own cache line per
cpu. This removes a large kmalloc/kfree bottleneck on multi-socket
systems
* Avoid having to ref, lock, and GETATTR intermediate directory components
in nlookup() by adding the NCF_WXOK flag. This flag is set in the ncp
when the directory permissions are at least 555. This saves significant
overhead in all situations, including single-threaded.
Discussed-with: Mateusz Guzik (mjg_)
zrj [Mon, 5 Dec 2016 14:08:47 +0000 (16:08 +0200)]
ncurses: Fix parallel buildworld race.
* MKexpanded.sh uses cc -E on source having curses.priv.h -> term.h chain.
Be safe and delay until all GENHDRS are created first.
No funtional change for final libprivate_ncurses{,w}.a,
just for reproducibility of buildworld intermediates.
Matthew Dillon [Mon, 5 Dec 2016 00:09:23 +0000 (16:09 -0800)]
kernel - Skip vmm_shutdown() when in a panic
* vmm_shutdown() calls cpusync and assumes that the system is basically
in working order. This is definitely not the case from a panic.
Conditionalize the code to be skipped if shutting-down due to a panic.
* Fixes a number of situations where the panic code locks up and refuses
to dump when called with debug.debugger_on_panic=0.
Imre Vadász [Sat, 3 Dec 2016 20:13:03 +0000 (21:13 +0100)]
syscons - Set cons_unavail when blanking framebuffer screen from syscons.
* This makes us treat panics while the screen is blanked in KMS mode
similar to panics while X is running. In both cases we have no
possibility left to display the syscons console.
* This also sets the cons_unavail flag, when the KMS driver is unloaded
and the framebuffer console becomes unusable, until a new framebuffer
is registered for syscons (which will clear cons_unavail again).
Imre Vadász [Thu, 17 Nov 2016 22:34:13 +0000 (23:34 +0100)]
drm: Use fb_blank hooks from drm_fb_helper.c as a syscons screensaver.
* When the KMS framebuffer is registered in syscons, a screensaver is
registered in syscons, which uses the fb_blank callback (if provided
by the drm driver) for turning the display off.
* The syscons screensaver can be used by e.g. running "vidcontrol -t 30"
for a 30s "screensaver" timeout), or via the Shift-Pause shortcut (for
instantly enabling the "screensaver").
Imre Vadász [Sat, 3 Dec 2016 20:15:13 +0000 (21:15 +0100)]
kernel: Fix stop_cpus()/restart_cpus() usages when panicing.
* If we are panicing (i.e. panicstr != NULL), Debugger() should make sure
that cpus are stopped when it returns. So call stop_cpus() explicitly
if Debugger() does an early return (i.e. in the cons_unavail case), and
don't call restart_cpus() at the end if we are panicing.
* This should make sure that Debugger()'s behaviour matches the
expectations of panic() in sys/kern/kern_shutdown.c.
Matthew Dillon [Sun, 4 Dec 2016 17:45:53 +0000 (09:45 -0800)]
kernel - set default locale name for smbfs
* Set default locale name for smbfs to "ISO8859-1".
Matthew Dillon [Sun, 4 Dec 2016 17:44:45 +0000 (09:44 -0800)]
test - Adjust blib and sc1
* convert loops to loops/sec, and add a parameter to syscall1.c to
allow the number of threads to be specified.
Matthew Dillon [Sun, 4 Dec 2016 17:21:45 +0000 (09:21 -0800)]
kernel - Overhaul namecache operations to reduce SMP contention
* Overhaul the namecache code to remove a significant amount of cacheline
ping-ponging from the namecache paths. This primarily effects
multi-socket systems but also improves multi-core single-socket systems.
Cacheline ping-ponging in the critical path can constrict a multi-core
system to roughly ~1-2M operations per second running through that path.
For example, even if looking up different paths or stating different
files, even something as simple as a non-atomic ++global_counter
seriously derates performance when it is being executed on all cores at
once.
In the simple non-conflicting single-component stat() case, this improves
performance from ~2.5M/second to ~25M/second on a 4-socket 48-core opteron
and has a similar improvement on a 2-socket 32-thread xeon, as well as
significantly improves namecache perf on single-socket multi-core systems.
* Remove the vfs.cache.numcalls and vfs.cache.numchecks debugging counters.
These global counters caused significant cache ping-ponging and were only
being used for debugging.
* Implement a poor-man's referenced-structure pcpu cache for struct mount
and struct namecache. This allows atomic ops on the ref-count for these
structures to be avoided in certain critical path cases. For now limit
to ncdir and nrdir (nrdir particularly, which is usually the same across
nearly all processes in the system). Eventually we will want to expand
this cache to handle more cases.
Because we are holding refs persistently, add a bit of infrastructure to
clear the cache as necessary (e.g. when doing an unmount, for example).
* Shift the 'cachedvnodes' global to a per-cpu accumulator, then roll-up
the counter back to the global approximately once per second. The code
critical paths adjust only the per-cpu accumulator, removing another
global cache ping-pong from nearly all vnode and nlookup paths.
* The nlookup structure now 'Borrows' the ucred reference from td->td_ucred
instead of crhold()ing it, removing another global ref/unref from all
nlookup paths.
* We have a large hash table of spinlocks for nchash, add a little pad
from 24 to 32 bytes. Its ok that two spin locks share the same cache
line (its a huge table), adding the pad cleans up cacheline-crossing
cases.
* Add a bit of pad to put mount->mnt_refs on its own cache-line verses
prior fields which are accessed shared. But don't bother isolating it
completely.
Matthew Dillon [Sun, 4 Dec 2016 17:10:25 +0000 (09:10 -0800)]
kernel - Reduce spinning on shared spinlocks
* Improve spinlock performance by removing unnecessary extra reads,
using atomic_fetchadd_int() to avoid a cmpxchg loop, and allowing
the SHARED flag to remain soft-set on the 1->0 transition.
* The primary improvement here is that multiple cpu's obtaining the
same shared spinlock can now do so via a single atomic_fetchadd_int(),
whereas before we had multiple atomics and cmpxchg loops. This does not
remove the cacheline ping-pong but it significantly reduces unnecessary
looping when multiple cpu cores are heavily loading the same shared spin
lock.
* Trade-off is against the case where a spinlock's use-case switches from
shared to exclusive or back again, which requires an extra atomic op to
deal with. This is not a common case.
* Remove spin->countb debug code, it interferes with hw cacheline operations
and is no longer desireable.
Discussed-with: Mateusz Guzik (mjg_)
Matthew Dillon [Sun, 4 Dec 2016 17:06:25 +0000 (09:06 -0800)]
hammer - Remove global VOP counters
* Remove global VOP counters. These were only used for debugging.
Removing these globals significantly improves concurrent VOP operations
on multi-core systems, particularly multi-socket systems, by removing
a cache ping-pong bottleneck.
Discussed-with: Mateusz Guzik (mjg_)
Sascha Wildner [Sun, 4 Dec 2016 11:34:08 +0000 (12:34 +0100)]
Rename cpumask.9 -> CPUMASK.9
Sascha Wildner [Sat, 3 Dec 2016 13:29:52 +0000 (14:29 +0100)]
libefivar: Fix wrong MLINKS.
Sascha Wildner [Sat, 3 Dec 2016 13:12:58 +0000 (14:12 +0100)]
Add a minimal nvmectl(8) manual page.
Sascha Wildner [Sat, 3 Dec 2016 10:35:59 +0000 (11:35 +0100)]
libusbhid: Remove some unneeded compat code.
Matthew Dillon [Fri, 2 Dec 2016 21:57:02 +0000 (13:57 -0800)]
kernel - Fix smbfs readdir and umount
* Fix a bug in smbfs's readdir that was causing an empty list and also
a later deadlock.
* Fix a bug in smbfs's umount that was panicing the machine.
Matthew Dillon [Fri, 2 Dec 2016 21:33:19 +0000 (13:33 -0800)]
kernel - Remove NETSMBCRYPTO option (make it the default)
* Remove the NETSMBCRYPTO option entirely, and make its function the
default.
Sascha Wildner [Fri, 2 Dec 2016 21:47:56 +0000 (22:47 +0100)]
kernel: Remove some ep(4) driver remains.
It was removed in August (
3a609f3dbc).
Matthew Dillon [Fri, 2 Dec 2016 00:32:27 +0000 (16:32 -0800)]
smbfs - Fix mount_smbfs authentication error (but 'ls' still broken)
* Fixes an authentication error with mount_smbfs. Most windows file servers
require a later crypto rev and man-in-the-middle protection.
* Note however that while mounting works, and files can be copied by name,
'ls' currently returns empty and the mount appears to get stuck, so more
work is needed.
Imre Vadász [Tue, 29 Nov 2016 22:04:46 +0000 (23:04 +0100)]
syscons - Add need_unlock arg to sc_clean_up and sc_wait_scrn_saver_stop.
* This should fix issues with VT switching when the screen is blanked.
Matthew Dillon [Thu, 1 Dec 2016 20:52:08 +0000 (12:52 -0800)]
kernel - Cleanup macros
* Fix ssb_insert_knote() and friends to use do { ... } while(0) intead
of { }, which fixes a double-statement issue with the macro.
Matthew Dillon [Thu, 1 Dec 2016 20:50:01 +0000 (12:50 -0800)]
kernel - Fix bug in socket_wait() (used by samba)
* socket_wait() was not properly initializing the temporary kqueue
structure, resulting in corruption that prevented the event from being
properly deleted.
* Fixes mount_smbfs panic.
Reported-by: dflyum
Imre Vadász [Thu, 1 Dec 2016 20:31:02 +0000 (21:31 +0100)]
poll.2: Add some information about the ppoll() syscall.
* Explain timeout and newsigmask arguments. (inspired by FreeBSD's poll(2))
* Mention that ppoll() can be used for precise timeouts like the ppoll()
in Linux.
Sascha Wildner [Thu, 1 Dec 2016 07:36:38 +0000 (08:36 +0100)]
<sys/{ipc,stat}.h>: Remove duplicate inclusion of <sys/cdefs.h>.
Sascha Wildner [Wed, 30 Nov 2016 07:26:31 +0000 (08:26 +0100)]
Sort SEE ALSO in some manual pages.
François Tigeot [Tue, 29 Nov 2016 21:39:09 +0000 (22:39 +0100)]
drm/i915: build firmware handling code
François Tigeot [Tue, 29 Nov 2016 21:23:16 +0000 (22:23 +0100)]
firmware(9): Change type of data pointer
* from const void * to const uint8_t *
* Avoiding tedious and error-prone void * conversions in Linux drivers
Sascha Wildner [Tue, 29 Nov 2016 21:16:01 +0000 (22:16 +0100)]
examples/rconfig: Fix CD/PXE check in two scripts.
Sascha Wildner [Tue, 29 Nov 2016 18:54:01 +0000 (19:54 +0100)]
Fix manual pages for s/struct device */device_t/.
Sascha Wildner [Tue, 29 Nov 2016 18:23:19 +0000 (19:23 +0100)]
mbuf.9: Adjust for removal of m_devget()'s 'copy' argument.
Imre Vadász [Mon, 28 Nov 2016 23:31:08 +0000 (00:31 +0100)]
syscons - Add range check for color value in KDSBORDER ioctl.
* Only values 0..15 are handled correctly for scp->border at the moment.
Imre Vadász [Mon, 28 Nov 2016 22:22:27 +0000 (23:22 +0100)]
syscons - Add option for the default terminal border color to kernel config.
* The default terminal border can be set to e.g. darkgrey by adding
option SC_BORDER_COLOR="FG_DARKGREY"
to the kernel configuration.
Sascha Wildner [Tue, 29 Nov 2016 13:07:47 +0000 (14:07 +0100)]
efivar.3: Fix prototypes in the SYNOPSIS.
Sascha Wildner [Tue, 29 Nov 2016 12:52:06 +0000 (13:52 +0100)]
Fix some typos in manual pages.
Sascha Wildner [Mon, 28 Nov 2016 22:36:52 +0000 (23:36 +0100)]
boot/efi: Add paging to the 'memmap' and 'configuration' commands.
Taken-from: FreeBSD
Sascha Wildner [Mon, 28 Nov 2016 22:35:25 +0000 (23:35 +0100)]
boot/efi: Add three new commands, efi-show, efi-set and efi-unset.
For accessing EFI variables. These replace the nvram command.
Taken-from: FreeBSD
Sascha Wildner [Mon, 28 Nov 2016 22:34:44 +0000 (23:34 +0100)]
libstand: Add uuid_{from,to}_string() and uuid_create_nil() (from libc).
Taken-from: FreeBSD
Matthew Dillon [Mon, 28 Nov 2016 20:18:54 +0000 (12:18 -0800)]
hammer2 - Revamp flush and xopq mechanism, stabilization
* Originally the xopq mechanism was meant to allow multiple worker
threads to share the same queue in a NxM configuration, but for
all intents and purposes we changed to a one-queue-per-thread
mechanic. Formalize this by associating the queue directly with
the thread structure.
* Separate out strategy-related XOPs from vnop-related XOPs, using half
the worker threads for strategy XOPs and the other half for vnop XOPs.
This fixes at least one deadlock which could occur if a strategy XOP
was queued after a vnop on the same worker thread. Since the
buffers related to the strategy XOP were locked, it could prevent
the vnop ahead of it from being able to execute.
* Cleanup the xopq sleep/wakeup mechanism and also fix a 30-60 second
stall when a xop-xop dependency is present on the same queue. The
clearing of the older xop was not waking up the worker thread.
* Revamp the transaction code, putpages via UIO_NOCOPY vop_write's,
and vop_strategy, to allow all buffer-cache-related I/Os to execute
concurrent with a flush. This fixes several deadlocks and long stalls
that could occur during flushes. Nominal modifying vnops will still
interlock with flushes, for now.
* Remove the PREFLUSH flag, it is no longer relevant.
* NOTE - Flush code probably doesn't completely handle modifying races
yet and needs more work. However, such races should be restricted to
buffer-cache-related I/Os only and thus it should be possible to
interlock them on an inode-by-inode basis.
Matthew Dillon [Mon, 28 Nov 2016 18:47:17 +0000 (10:47 -0800)]
fsstress - Fix compiler warning
* Remove annoying compiler warning.
Matthew Dillon [Mon, 28 Nov 2016 18:45:54 +0000 (10:45 -0800)]
kernel - Remove buffer flush debug message
* Remove the "dirty bufs left after final pass" debug message. This case
can occur normally during heavy I/O, particularly with the concurrency
DragonFly now allows.
Matthew Dillon [Mon, 28 Nov 2016 18:41:53 +0000 (10:41 -0800)]
kernel - Fix deadlock in vm_page_repurpose()
* vm_page_repurpose() was hard+soft busying the underlying VM page,
which can deadlock against putpages or other I/O.
* Only hard-busy the page, then add an SBUSY test to the failure case
(we don't want to repurpose a page undergoing I/O after all!).
Sascha Wildner [Mon, 28 Nov 2016 17:46:11 +0000 (18:46 +0100)]
kernel/ath: Limit check index variables before using them.
Sascha Wildner [Mon, 28 Nov 2016 14:49:05 +0000 (15:49 +0100)]
boot/efi: Use strtoul().
Reduces diffs with FreeBSD.
Sascha Wildner [Mon, 28 Nov 2016 14:47:18 +0000 (15:47 +0100)]
libstand: Add strtoul().
Taken-from: FreeBSD
Sascha Wildner [Mon, 28 Nov 2016 14:40:02 +0000 (15:40 +0100)]
<machine/limits.h>: Add some i386 limits which libstand's strtol() needs.
Sascha Wildner [Mon, 28 Nov 2016 07:45:51 +0000 (08:45 +0100)]
Sync zoneinfo database with tzdata2016j from ftp://ftp.iana.org/tz/releases
* Saratov, Russia switches from +03 to +04 on 2016-12-04 at 02:00.
This hives off a new zone Europe/Saratov from Europe/Volgograd.
(Thanks to Yuri Konotopov and Stepan Golosunov.)
* The new zone Asia/Atyrau for Atyraū Region, Kazakhstan, is like
Asia/Aqtau except it switched from +04/+05 to +05/+06 in spring
1999, not fall 1994. (Thanks to Stepan Golosunov.)
* Asia/Gaza and Asia/Hebron now use "EEST", not "EET", to denote
summer time before 1948. The old use of "EET" was a typo.
* The 'Theory' file now documents UT.
* iso3166.tab now accents "Curaçao", and commentary now mentions
the names "Cabo Verde" and "Czechia". (Thanks to Jiří Boháč.)
Imre Vadász [Mon, 21 Nov 2016 22:12:05 +0000 (23:12 +0100)]
syscons - Clean up log messages for KMS console.
* This was printing two lines of kernel messages for every virtual terminal
that was switched to a new KMS framebuffer configuration.
Instead only print the display resolution when the framebuffer is
registered with syscons, and print the font scaling information for
each VT only when debugging is enabled in syscons.
Imre Vadász [Sun, 20 Nov 2016 21:01:40 +0000 (22:01 +0100)]
drm: Improve integration with syscons. Move taskqueue handling to syscons.
* Adds asynctd_lk to syscons' softc to synchronize unregister_framebuffer
code with the asynchronous screen refresh threads.
* Use a generation counter in syscons to check if sc_update_render() needs
to update the VT resolution/mode. Remove struct fb_info *fbi pointer
from scr_stat, and always check the struct fb_info *fbi pointer in the
softc struct instead.
* Use videio_in_progress flag to make initial fb_set_par call a bit safer.
* Moves driver callbacks from struct fb_info itself into struct fb_ops,
and adds fb_blank and fb_debug_enter callbacks which will be wired up
in future commits.
* Add unregister_framebuffer() function to syscons, which is needed for
cleanly detaching a drm graphics driver which is used as framebuffer
console. The screen will turn off, blank or display a frozen screen,
after the graphics driver has detached.
* Use drm_fb_helper_unregister_framebuffer() and drm_fb_helper_release_fbi()
functions in radeon and i915, which are needed for cleanly detaching the
drivers.
* With radeonkms detaching and re-attaching is working very reliably
on an AMD G-T44R APU (i.e. HD6250 graphics).
* Detaching i915 on a Haswell notebook barely worked with several errors,
and trying to re-attach i915 caused the machine to hang.
Imre Vadász [Sun, 27 Nov 2016 22:01:30 +0000 (23:01 +0100)]
syscons - Make sure we draw the terminal border during syscons startup.
* Try drawing the border first in scinit. If we use the UEFI framebuffer,
and efi_fb_info.vaddr is still NULL, we call sc_set_border again in
scmeminit() when efi_fb_info.vaddr will be definitely non-NULL.
* This makes sure that we clear any garbage left from the UEFI bootloader
stage on the screen, after
ae88e4ee083dbf5a51e10cae2cdf213e2a7c5225.
Sascha Wildner [Mon, 28 Nov 2016 00:06:51 +0000 (01:06 +0100)]
sigpause.3: Add missing comma.
Sascha Wildner [Mon, 28 Nov 2016 00:04:54 +0000 (01:04 +0100)]
efivar.{3,8}: Fix some mdoc issues.
efivar.3: Remove extra semicolons, fix section ordering, add SEE ALSO.
efivar.8: Fix wrong .El, add SEE ALSO, add parentheses.
Sascha Wildner [Sun, 27 Nov 2016 23:51:45 +0000 (00:51 +0100)]
libefivar: Remove unneeded WARNS setting (lib/Makefile.inc has it).
Sascha Wildner [Sun, 27 Nov 2016 23:51:18 +0000 (00:51 +0100)]
mdoc.local: Add definition for '.Lb libefivar'.
Sascha Wildner [Sun, 27 Nov 2016 23:36:28 +0000 (00:36 +0100)]
kernel/efirt: Make efirt compilable into the kernel and add to LINT64.
Sascha Wildner [Sun, 27 Nov 2016 23:35:40 +0000 (00:35 +0100)]
kernel/efirt: Hook efirt.ko into the build.
Sascha Wildner [Sun, 27 Nov 2016 23:35:00 +0000 (00:35 +0100)]
Move efirt.ko's source to sys/dev/misc/efirt.
Michael Neumann [Sun, 27 Nov 2016 21:51:40 +0000 (22:51 +0100)]
rc.subr: Add ${name}_env and ${name}_prepend.
${name}_env is an argument list which will be passed to env(1).
${name}_prepend is simply prepended to the command line for $command.
This fixes issues I had with starting up www/h2o as described [here][1].
Taken-From: FreeBSD (commit
77616d41a2ce3744b35e68b8a3c12e562283a322)
[1]: https://github.com/DragonFlyBSD/DeltaPorts/issues/663
Tomohiro Kusumi [Sun, 27 Nov 2016 05:35:27 +0000 (14:35 +0900)]
sys/dev/disk/dm: Make init/destroy handlers optional
Some targets don't need to do anything in init/destroy handler,
as there is no private data to init/cleanup, so make them optional.
table_en->target_config can be set to NULL regardless of destroy
handler or its result.
Tomohiro Kusumi [Sun, 27 Nov 2016 05:21:11 +0000 (14:21 +0900)]
sys/dev/disk/dm: Remove dm_table_init_target() call with NULL
These can be removed after
41a68322.
Imre Vadász [Sat, 26 Nov 2016 22:49:01 +0000 (23:49 +0100)]
syscons - Fix syscons with UEFI fb when drm fb is registered during boot.
* Fix sc->fbi related checks in register_framebuffer() and scinit(), to
patch up the case where the drm framebuffer is registered before
syscons' late initialization (i.e. the actual syscons attach function)
runs.
* Change some of the SC_EFI_FB checks in scinit() to check instead for
sc->fbi != NULL.
Imre Vadász [Sun, 27 Nov 2016 00:43:19 +0000 (01:43 +0100)]
kernel/pc64: Don't fill uefi framebuffer with 0x77 during early boot.
* This memset was only meant for debugging uefi boot, and just produces an
annoying flick or a short delay during UEFI boot now.
Tomohiro Kusumi [Sat, 26 Nov 2016 10:21:01 +0000 (19:21 +0900)]
sys/dev/disk/dm: Remove obvious/redundant/disabled kprintfs
Tomohiro Kusumi [Sat, 26 Nov 2016 05:58:26 +0000 (14:58 +0900)]
sys/dev/disk/dm: Cleanups
This part is unclear for no reason.
Tomohiro Kusumi [Sat, 26 Nov 2016 05:17:43 +0000 (14:17 +0900)]
sys/dev/disk/dm: Remove dm_dev::dev_type
This is unused, and also not necessary.
dm core shouldn't need to be aware of target driver type.
Target dependent actions are handled by target dependent handlers.
dm targets have oop-like structure, so we don't want/need
dm core to be able to do things like below.
Also see
d471f1f9 and
49784e7d.
switch (dev->dev_type) {
case DM_LINEAR_DEV:
do_something_specific_to_linear(); break;
case DM_STRIPE_DEV:
do_something_specific_to_striped(); break;
case ...:
...; break;
}
Tomohiro Kusumi [Sat, 26 Nov 2016 03:46:41 +0000 (12:46 +0900)]
sys/dev/disk/dm: Remove dm/targets/ directory and move its entries to dm/
Device mapper targets' path
sys/dev/disk/dm/targets/<name>/dm_target_<name>.c
are too deep for no reason.
This commit removes targets/ directory which isn't necessary at all,
assuming userspace has no dependencies on this (and it shouldn't either).
Neither NetBSD (sys/dev/dm/dm_target_*.c) nor Linux kernel (drivers/md/dm*.c)
has a directory tree like this, which is better.
Sascha Wildner [Sat, 26 Nov 2016 15:55:19 +0000 (16:55 +0100)]
libc: Fix up compat-43 a bit.
* Functions in libc/compat-43 aren't syscalls so move their manual
pages to section 3.
* Move the stuff that is still POSIX to libc/gen: creat, gethostid,
getwd, killpg. compat-43 isn't the right place for them.
* While here, a non-existant .PATH in compat-43/Makefile.inc.
Bill Yuan [Sat, 26 Nov 2016 11:51:21 +0000 (11:51 +0000)]
ipfw3: 'or' supports more filters
Imre Vadász [Sat, 26 Nov 2016 00:25:56 +0000 (01:25 +0100)]
syscons - Remove unneeded NULL check from register_framebuffer().
* While there use tabs for indenting register_framebuffer() to reduce
diff of a following change.
Imre Vadász [Fri, 25 Nov 2016 00:26:32 +0000 (01:26 +0100)]
syscons - Handle kms framebuffer activation better when not in text mode.
* We shouldn't just silently remove the GRAPHICS_MODE or UNKNOWN_MODE flags,
instead try to handle those cases a bit more graceful.
* This avoid's Xorg's VT being silently switched back from UNKNOWN_MODE
to text mode, when the drm driver module is automatically loaded while
Xorg is starting.
Imre Vadász [Sat, 19 Nov 2016 20:33:54 +0000 (21:33 +0100)]
drm/radeon: restore the fbdev mode in lastclose
restore the fbdev state if a drm app like X is killed.
Taken-From: Linux git
8c70e1cda04b966b50ddfefafbd0ea376ed8edd5
Imre Vadász [Fri, 25 Nov 2016 22:13:42 +0000 (23:13 +0100)]
iwm - Reduce gratuitous differences with Linux iwlwifi in struct naming.
* Rename some structs and struct members for firmware handling.
zrj [Fri, 25 Nov 2016 10:41:12 +0000 (11:41 +0100)]
drm/linux: add circ_buf.h needed for guc client bits
Found on github implemented by mattmacy@ FreeBSD repo
François Tigeot [Fri, 25 Nov 2016 09:07:48 +0000 (10:07 +0100)]
drm/linux: Add request_firmware_nowait()
François Tigeot [Fri, 25 Nov 2016 08:06:16 +0000 (09:06 +0100)]
drm/linux: Improve request_firmware() compatibility
request_firmware() takes a pointer to a Linux struct device as
third parameter.
zrj [Sun, 20 Nov 2016 14:06:25 +0000 (16:06 +0200)]
binutils227: Symlink liblto_plugin.so from default base gcc compiler too.
Now that alternative binutils are updated, further work towards making
binutils 2.27 as default can continue.
Also documment bfd-plugins/ directory in hier(7).
zrj [Fri, 16 Sep 2016 04:47:19 +0000 (07:47 +0300)]
rtld: Fix handling of DT_TEXTREL for object with multiple read-only segments.
Taken-from: FreeBSD (r296319)
zrj [Sun, 20 Nov 2016 13:47:14 +0000 (15:47 +0200)]
Retire the binutils-2.24.
Remove contrib sources and build Makefiles.
zrj [Sat, 19 Nov 2016 18:52:45 +0000 (20:52 +0200)]
Replace binutils 2.24 with latest binutils 2.27.
zrj [Sat, 19 Nov 2016 18:50:38 +0000 (20:50 +0200)]
binutils227: Add build dirs.
It is unclear weather -Wl,z,relro should be on by default in ld.gold.
zrj [Sat, 19 Nov 2016 17:46:44 +0000 (19:46 +0200)]
binutil227: Add DF README's and local modifications.
zrj [Thu, 24 Nov 2016 15:28:14 +0000 (17:28 +0200)]
Merge remote-tracking branch 'origin/vendor/BINUTILS227'
zrj [Thu, 24 Nov 2016 07:20:38 +0000 (09:20 +0200)]
Initial import of binutils 2.27 on vendor branch
Sascha Wildner [Wed, 23 Nov 2016 20:04:50 +0000 (21:04 +0100)]
Sync ACPICA with Intel's version
20161117.
* Fixed regression introduced in
20160729 in the table manager. Also,
add a new public interface, AcpiPutTable().
* Fix a problem where runtime implicit conversion as incorrectly
disabled for some ASL operators.
* Various fixes and improvements for iasl(8).
* In acpixtract(8), add handling for both LF and CR/LF line
termination to accept acpidump output from any system.
* In acpibin(8), add new -a and -o options.
For a more detailed list, please see sys/contrib/dev/acpica/changes.txt.