dragonfly.git
6 years agokernel/pc64: Adjust some references to already removed i386.
zrj [Tue, 19 Dec 2017 15:59:05 +0000 (17:59 +0200)]
kernel/pc64: Adjust some references to already removed i386.

While there, perform some whitespace fixes.
No functional change.

6 years agofsck_msdosfs(8): Use __printflike.
Sascha Wildner [Wed, 20 Dec 2017 04:59:44 +0000 (05:59 +0100)]
fsck_msdosfs(8): Use __printflike.

6 years agorouted(8): Clean up a bit.
Sascha Wildner [Wed, 20 Dec 2017 04:55:50 +0000 (05:55 +0100)]
routed(8): Clean up a bit.

* Use __printflike and __unused.

* Remove sccsid[] etc.

6 years agoAdd __attribute__((__noreturn__)) to various function prototypes in sbin/.
Sascha Wildner [Wed, 20 Dec 2017 03:33:04 +0000 (04:33 +0100)]
Add __attribute__((__noreturn__)) to various function prototypes in sbin/.

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

6 years agolibc/sysconf: Add _SC_LEVEL1_DCACHE_LINESIZE.
Sascha Wildner [Tue, 19 Dec 2017 18:51:17 +0000 (19:51 +0100)]
libc/sysconf: Add _SC_LEVEL1_DCACHE_LINESIZE.

nginx can make use of it (see http://hg.nginx.org/nginx/rev/057adb2a9d23).

Requested-by: sephe
6 years agokernel/pc64: Rename i386_protection_init().
zrj [Tue, 19 Dec 2017 15:26:44 +0000 (17:26 +0200)]
kernel/pc64: Rename i386_protection_init().

No functional change.

6 years agoapply(1): Add const attributes.
zrj [Tue, 19 Dec 2017 15:25:48 +0000 (17:25 +0200)]
apply(1): Add const attributes.

6 years agotcp: Stringent assertion for pru_connect.
Sepherosa Ziehau [Tue, 19 Dec 2017 05:00:51 +0000 (13:00 +0800)]
tcp: Stringent assertion for pru_connect.

6 years agogettytab: Add 3wire and std as terminal types/classes.
Eitan Adler [Tue, 19 Dec 2017 06:03:02 +0000 (22:03 -0800)]
gettytab: Add 3wire and std as terminal types/classes.

These are similar to
the existing terminal types/classes that have the baudrate suffix,
but differ in that no baudrate is set/defined.

Obtained-from: FreeBSD (r262955)

6 years agolibc/nmalloc: Handle size_t overflows in posix_memalign().
zrj [Sun, 17 Dec 2017 15:54:01 +0000 (17:54 +0200)]
libc/nmalloc: Handle size_t overflows in posix_memalign().

Some algorithms/programs try to heuristically deduce alignment rules.
Instead of confusing them with minimal allocations for size == (size_t)-7 and
friends at different alignment values, just return ENOMEM since POSIX does not
explictly specify how alignment overflows should behave. Programs should keep
track of pointers they try to allocate (including size == 0 case too).

6 years agolibc/nmalloc: Whistespace cleanup.
zrj [Sun, 17 Dec 2017 15:47:30 +0000 (17:47 +0200)]
libc/nmalloc: Whistespace cleanup.

No functional change.

6 years agokernel - Relax requirements for sysctl operations on longs
Matthew Dillon [Mon, 18 Dec 2017 05:45:34 +0000 (21:45 -0800)]
kernel - Relax requirements for sysctl operations on longs

* Relax the requirements for sysctl operations on longs.  Allow the
  case where userland is reading the sysctl variable as an integer
  instead of a long.

* Allows us to promote various vm.stats.vm_* fields that must now be
  longs in order to support > 8TB of ram as well as avoid internal
  calculation overflows.

* Fixes an issue where chrome issues this sysctl as an int:

  sysctlbyname("vm.stats.vm.v_page_count", &pages, &size, NULL, 0);

  And fails to check the return value / error code, resulting in
  a garbage value for 'pages' which chrome then uses to size the
  browser's memory restrains.  This in turn was causing chrome
  to give up multiple tabs thinking it had run out of memory when,
  in fact, there is plenty of memory available.

* Generally speaking, I've wanted to have this sort of backwards
  compatibility for a while now.  Userland shouldn't get errors
  accesing integral sysctl values simply because it uses an integer
  of a different size than the sysctl.

6 years agomd5(1): Add support for sha512.
zrj [Wed, 13 Dec 2017 07:39:00 +0000 (09:39 +0200)]
md5(1): Add support for sha512.

6 years agomd5(1): Minor cleanup.
zrj [Wed, 13 Dec 2017 07:04:08 +0000 (09:04 +0200)]
md5(1): Minor cleanup.

 * add static and const attributes;
 * use RUSAGE_SELF;
 * fix some whitespace errors.

No functional change.

6 years agoboot/efi: Remove our old efi include files.
Sascha Wildner [Sun, 17 Dec 2017 13:42:03 +0000 (14:42 +0100)]
boot/efi: Remove our old efi include files.

The remaining local header is eficonsctl.h which I will deal with later.

6 years agoboot/efi: Port most of our efi boot code to the EDK II headers.
Sascha Wildner [Sun, 17 Dec 2017 13:41:22 +0000 (14:41 +0100)]
boot/efi: Port most of our efi boot code to the EDK II headers.

6 years agoboot/efi: Bring in a bunch of additional TianoCore EDK II headers.
Sascha Wildner [Sun, 17 Dec 2017 13:40:41 +0000 (14:40 +0100)]
boot/efi: Bring in a bunch of additional TianoCore EDK II headers.

Taken from the MdePkg of the UDK2017 branch of the
git://github.com/tianocore/edk2 repository.
All I modified was killing CRs using 'tr -d "\015"'.

This is a work in progress.

6 years agoigb.4: Sort chip names.
Sepherosa Ziehau [Sun, 17 Dec 2017 03:36:49 +0000 (11:36 +0800)]
igb.4: Sort chip names.

6 years agoix.4: Sort chip names.
Sepherosa Ziehau [Sun, 17 Dec 2017 03:33:04 +0000 (11:33 +0800)]
ix.4: Sort chip names.

6 years agomount_udf(8): Staticize
Sascha Wildner [Sun, 17 Dec 2017 09:39:55 +0000 (10:39 +0100)]
mount_udf(8): Staticize

6 years agodumpfs(8): Staticize
Sascha Wildner [Sun, 17 Dec 2017 09:26:31 +0000 (10:26 +0100)]
dumpfs(8): Staticize

6 years agodmesg(8): Staticize
Sascha Wildner [Sun, 17 Dec 2017 09:11:02 +0000 (10:11 +0100)]
dmesg(8): Staticize

6 years agodevfsctl(8): syntax_error() never returns, so change it to 'void'.
Sascha Wildner [Sat, 16 Dec 2017 21:53:56 +0000 (22:53 +0100)]
devfsctl(8): syntax_error() never returns, so change it to 'void'.

While here, fix indent a bit and add a copyright to devfsctl.h.

6 years agoigb: Sync with Intel's igb 2.5.3
Sepherosa Ziehau [Sat, 2 Dec 2017 17:43:05 +0000 (01:43 +0800)]
igb: Sync with Intel's igb 2.5.3

- Enable RX packets dropping, if flow control is disabled.
- Clear bad data at chip initialization time.

Tested-by: Tim Darby <t timdarby net> (i211)
   Thomas Nikolajsen <thomas nikolajsen mail dk> (i210)

6 years agoRemove /boot/efiboot.img via 'make upgrade'.
Sascha Wildner [Sat, 16 Dec 2017 10:39:05 +0000 (11:39 +0100)]
Remove /boot/efiboot.img via 'make upgrade'.

This is needed in our ISOs only to allow booting in EFI mode from
CD-ROM. In the installed system it is not needed anymore, so it
can be removed.

6 years agobsd-family-tree: add NetBSD 7.0.2; Add recent DragonFly
Eitan Adler [Fri, 15 Dec 2017 03:38:21 +0000 (19:38 -0800)]
bsd-family-tree: add NetBSD 7.0.2; Add recent DragonFly

Also changes exact location of NetBSD 7.0.1
Gives some breathing room for adding new releases

Obtained-from: NetBSD (partially)

6 years agoSync our gcc5 manual pages with gcc-5.4.0
Sascha Wildner [Wed, 13 Dec 2017 14:08:59 +0000 (15:08 +0100)]
Sync our gcc5 manual pages with gcc-5.4.0

6 years agoRemove <sys/ioccom.h> inclusion from a number of files.
Sascha Wildner [Wed, 13 Dec 2017 12:27:41 +0000 (13:27 +0100)]
Remove <sys/ioccom.h> inclusion from a number of files.

<sys/ioccom.h> should only be included from header files that define
ioctls.

6 years agotmpfs - Fix arbitrary maximum file size limitation
Matthew Dillon [Tue, 12 Dec 2017 22:45:04 +0000 (14:45 -0800)]
tmpfs - Fix arbitrary maximum file size limitation

* tmpfs's maxfilesize was limited to the original tmpfs mount
  storage limit, prior to argument overrides.  In addition,
  the -f argument override would not increase the maxfilesize
  limit beyond the original storage limit anyway.

* Remove this limit calculation entirely.  Instead the limit
  is based on the storage limit which can be optioned at mount
  time.

* Fixes expectations when tmpfs is used to hold just a few (or
  even just one) file.

Reported-by: kerma
6 years agoboot/efi: Sync IndustryStandard/Bluetooth.h with UDK2017 branch of EDK II.
Sascha Wildner [Tue, 12 Dec 2017 14:03:19 +0000 (15:03 +0100)]
boot/efi: Sync IndustryStandard/Bluetooth.h with UDK2017 branch of EDK II.

6 years agoboot/efi: Bring in two more TianoCore EDK II headers.
Sascha Wildner [Tue, 12 Dec 2017 08:59:07 +0000 (09:59 +0100)]
boot/efi: Bring in two more TianoCore EDK II headers.

Taken from the MdePkg of the UDK2017 branch of the
git://github.com/tianocore/edk2 repository.
All I modified was killing CRs using 'tr -d "\015"'.

This is a work in progress.

6 years agoboot/efi: Remove an unused header.
Sascha Wildner [Tue, 12 Dec 2017 08:55:37 +0000 (09:55 +0100)]
boot/efi: Remove an unused header.

If we ever need it we'd take it from TianoCore EDK II, i.e.
MdePkg/Include/IndustryStandard/Mbr.h.

6 years agokernel - Revert part of the contig allocation work
Matthew Dillon [Sun, 10 Dec 2017 22:09:03 +0000 (14:09 -0800)]
kernel - Revert part of the contig allocation work

* Revert part of the contig allocation work.  Always use the alist
  allocation from the DMA reserve first, and revert to scanning the
  VM page list only if it fails.

* Fixes performance problems with X.

* What this means, generally, is that we will probably need to rely
  directly on the dma reserve for most of our DMA allocation and
  freeing just to ensure reasonable performance.  Scanning page
  lists is a big lose.

Reported-by: mneumann
6 years agolibc: Forgotten adjustment in Versions.def
Sascha Wildner [Sun, 10 Dec 2017 15:56:53 +0000 (16:56 +0100)]
libc: Forgotten adjustment in Versions.def

Reported-by: Aaron LI <aly@aaronly.me>
6 years agoMakefile_upgrade.inc: Add adding_user.8 to the end of the list.
Sascha Wildner [Sun, 10 Dec 2017 07:15:19 +0000 (08:15 +0100)]
Makefile_upgrade.inc: Add adding_user.8 to the end of the list.

6 years agolibc/termios: Add cfmakesane().
Sascha Wildner [Sun, 10 Dec 2017 07:08:27 +0000 (08:08 +0100)]
libc/termios: Add cfmakesane().

This makes net/bsdrcmds build, in preparation for removing rcp, rlogin,
rlogind, rsh and rshd from base.

Taken-from: FreeBSD

6 years agoUpdate bsd-family-tree
Eitan Adler [Sun, 10 Dec 2017 06:42:19 +0000 (22:42 -0800)]
Update bsd-family-tree

6 years agoRemove 'adding_user.8'
Eitan Adler [Sun, 10 Dec 2017 05:56:27 +0000 (21:56 -0800)]
Remove 'adding_user.8'

Not to be confused with adduser.
Not to be confused with useful information.

6 years agoflowers: correct spelling of lily, Camellia
Eitan Adler [Sun, 10 Dec 2017 05:09:54 +0000 (21:09 -0800)]
flowers: correct spelling of lily, Camellia

6 years agoscsi_modes - bring up to date
Eitan Adler [Sun, 10 Dec 2017 05:06:41 +0000 (21:06 -0800)]
scsi_modes - bring up to date

Obtained-from:   FreeBSD

6 years agoascii: Use standard name for ASCII control codes
Eitan Adler [Sun, 10 Dec 2017 05:02:44 +0000 (21:02 -0800)]
ascii: Use standard name for ASCII control codes

Obtained-from: FreeBSD r293281

6 years agorlogind(8): Staticize and remove unused declarations.
Sascha Wildner [Sat, 9 Dec 2017 19:16:32 +0000 (20:16 +0100)]
rlogind(8): Staticize and remove unused declarations.

6 years agoAdd test to dports-download-check target
Pierre-Alain TORET [Wed, 29 Nov 2017 09:37:03 +0000 (10:37 +0100)]
Add test to dports-download-check target

Signed-off-by: Pierre-Alain TORET <pierre-alain.toret@protonmail.com>
Dragonfly-bug: <https://bugs.dragonflybsd.org/issues/3053>

6 years agoAdd __attribute__((__noreturn__)) to various function prototypes in bin/.
Sascha Wildner [Sat, 9 Dec 2017 08:09:51 +0000 (09:09 +0100)]
Add __attribute__((__noreturn__)) to various function prototypes in bin/.

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

6 years agorcp(1): Staticize.
Sascha Wildner [Sat, 9 Dec 2017 07:57:02 +0000 (08:57 +0100)]
rcp(1): Staticize.

6 years agopwd(1): Staticize, remove void cast.
Sascha Wildner [Sat, 9 Dec 2017 07:47:50 +0000 (08:47 +0100)]
pwd(1): Staticize, remove void cast.

6 years agoAdd dports-download-check target
Pierre-Alain TORET [Wed, 29 Nov 2017 09:16:42 +0000 (10:16 +0100)]
Add dports-download-check target

Signed-off-by: Pierre-Alain TORET <pierre-alain.toret@protonmail.com>
Dragonfly-bug: <https://bugs.dragonflybsd.org/issues/3053>

6 years agodrm/radeon: Prevent a crash on load with newer drm code
François Tigeot [Fri, 8 Dec 2017 21:57:34 +0000 (22:57 +0100)]
drm/radeon: Prevent a crash on load with newer drm code

* The existing drm/radeon code base contained an invalid unlock/lock
  sequence in radeon_irq_kms_init() which caused the kernel to crash
  during driver initialization with the drm code base from Linux 4.8

* Fixed by replacing the problematic function with its version from
  Linux 4.8

6 years agokernel/evdev: Fix Makefile permission.
Sascha Wildner [Fri, 8 Dec 2017 14:45:08 +0000 (15:45 +0100)]
kernel/evdev: Fix Makefile permission.

6 years agokernel - Add evdev support to kbdmux, ukbd and ums
Peeter Must [Fri, 8 Dec 2017 12:17:22 +0000 (14:17 +0200)]
kernel - Add evdev support to kbdmux, ukbd and ums

* Allows kbdmux, ukbd and ums to send input events to evdev if kernel
  is built with evdev support.

* The sysctl kern.evdev.rcpt_mask determines which drivers are allowed
  to send events to evdev. Its default value is 3, meaning that kbdmux
  and sysmouse are enabled. Since the latter is not yet supported, one
  might want to change the value to 6 (kbdmux and ums) until sysmouse
  will be added.

Obtained-from: FreeBSD

6 years agokernel - Bring in evdev from FreeBSD
Peeter Must [Fri, 8 Dec 2017 10:43:27 +0000 (12:43 +0200)]
kernel - Bring in evdev from FreeBSD

* This is a port of evdev, the generic input event interface, from
  FreeBSD. Wikipedia: "Evdev generalizes raw input events from device
  drivers and makes them available through character devices in the
  /dev/input/ directory."

* It is currently considered experimental since in rare cases it may cause
  kernel crash when a device, e.g. usb mouse or keyboard, is detached
  while a userland program reading from the corresponding input/eventX
  device.

* In order to enable evdev, kernel needs to be rebuilt with 'device evdev'
  and 'options EVDEV_SUPPORT'. For debugging, add 'options EVDEV_DEBUG'.

* At present, only ums, kbd and kbdmux can send events to their respective
  input/eventX devices. More drivers will be added in due course. The sysctl
  kern.evdev.rcpt_mask determines which drivers send events to evdev.

6 years agokernel - Fix crash when cdev is destroyed
Peeter Must [Fri, 8 Dec 2017 09:53:22 +0000 (11:53 +0200)]
kernel - Fix crash when cdev is destroyed

* If a device is destroyed while a userland program is blocked
  in a read() or device ioctl, the underlying node will be deleted.
  Returning EIO will prevent kernel from crashing.

* This is a temporary hack until a better solution is found.

6 years agotmpfs - Correct some timestamp update issues
Matthew Dillon [Fri, 8 Dec 2017 02:20:49 +0000 (18:20 -0800)]
tmpfs - Correct some timestamp update issues

* write() was updating ACCESSED and MODIFIED when it should only
  be updating MODIFIED.

* unlink() (aka rm) was updating CHANGE, ACCESSED, and MODIFIED
  which would be visible if a file has multiple links.  It should
  only update the CHANGE time.

* rmdir() was doing the same thing as UNLINKED.  Fixed this too,
  but the bug would not be readily visible anyway since directories
  cannot be hardlinked.  However, fstat() on an open descriptor
  would still reveal it.

6 years agoddb.4: Document i command addition.
zrj [Thu, 7 Dec 2017 09:06:17 +0000 (11:06 +0200)]
ddb.4: Document i command addition.

While there, document call and reset commands too.

6 years agokernel - Improve contiguous DMA memory allocation
Matthew Dillon [Thu, 7 Dec 2017 03:09:20 +0000 (19:09 -0800)]
kernel - Improve contiguous DMA memory allocation

* vm_page_alloc_contig() will use vm_page_alloc() for degenerate
  requests instead of the free page reserve.  This case often
  occurs when called from kmem_alloc_attr().

* vm_contig_pg_alloc() will scan the VM page queues first for
  unrestricted memory requests, falling back to vm_page_alloc_contig()
  if this fails.

  vm_contig_pg_alloc() will call vm_page_alloc_contig() first if the
  memory request has an address range restriction, falling back to
  a VM page queue scan if it fails.

* This should hopefully reduce instances where device drivers probed
  prior to e.g. NATA eat up all available low-memory DMA space,
  leaving none for NATA (or other drivers) that might need low-memory
  DMA space.

6 years agokernel - Fix ranging test in memory probe, adjust memory test (2)
Matthew Dillon [Wed, 6 Dec 2017 18:51:28 +0000 (10:51 -0800)]
kernel - Fix ranging test in memory probe, adjust memory test (2)

* Correct bug in ranging test commit.

6 years agokernel - Fix vmstats.v_dma_pages
Matthew Dillon [Wed, 6 Dec 2017 08:23:16 +0000 (00:23 -0800)]
kernel - Fix vmstats.v_dma_pages

* Fix an atomic adjustment on v_dma_pages that we believe was
  messing it up, due to not sign-extended a negation of an unsigned
  field.

Reported-by: swildner
6 years agokernel - Fix hw.usermem, hw.physmem, hw.availpages
Matthew Dillon [Wed, 6 Dec 2017 07:00:22 +0000 (23:00 -0800)]
kernel - Fix hw.usermem, hw.physmem, hw.availpages

* Fix the hw.usermem, hw.physmem, and hw.availpages sysctls.
  Some of these were improperly 32-bit sysctls when they need
  to be 64-bits, causing silly values to be returned on
  machines with more than 2G of ram.

Reported-by: zrj
6 years agokernel - Adjust vfs.nbuf, kern.maxvnodes, and debug.max_softdeps counts
Matthew Dillon [Wed, 6 Dec 2017 04:11:51 +0000 (20:11 -0800)]
kernel - Adjust vfs.nbuf, kern.maxvnodes, and debug.max_softdeps counts

* Cap some system limits on machines with tons of memory.

* vfs.nbuf is capped at roughly 350,000 buffer cache buffers.  This
  also caps buffer cache data to approximately 11G.

  The reason for this is to avoid wiring too much physical memory in
  the default configuration since programs might want to use the
  memory fo r other purposes.  Buffer cache buffers provide very fast
  access when reading data from cached vnodes, but the new KVABIO ABI
  will soon allow us to acquire and dispose of such buffers more
  efficiently, increasing the efficiency of accessing data cached via
  the block device and making long-term caching via vnode-backed
  buffers somewhat less important.

* kern.maxvnodes is capped at approximately 3 million vnodes.  This
  can be set higher at run-time via a sysctl, if desired.

  NOTE: Our go-to filesystems are designed to or can be set to cache
file data via the block device instead of the vnode, which is
preferable on systems with lots of memory.  vnode caching is
still important, but a bit less so now than in prior years.

* debug.max_softdeps is capped at 1000000, because even a million
  is an obscenely high value for softdep structures.

6 years agokernel - Fix ranging test in memory probe, adjust memory test
Matthew Dillon [Wed, 6 Dec 2017 02:17:53 +0000 (18:17 -0800)]
kernel - Fix ranging test in memory probe, adjust memory test

* Fix a ranging test in the memory probe code.  The 'incr' value can
  potentially skip over blacked-out areas of memory and thus accidently
  include them in the range.

* We no longer exhaustively test memory.  Instead, just test the first
  and last page for each range.

6 years agokernel - add 'i' command to ddb
Matthew Dillon [Tue, 5 Dec 2017 23:29:00 +0000 (15:29 -0800)]
kernel - add 'i' command to ddb

* Add the 'i' command, which just issues a cpu_invltlb() on the
  current cpu.  Useful when testing MMU/PMAP issues.

6 years agoefisetup - Correct bugs
Matthew Dillon [Tue, 5 Dec 2017 22:40:22 +0000 (14:40 -0800)]
efisetup - Correct bugs

* nvme0 was hardwired in one place, use ${drive}.  This fixes
  issues with efisetup not working on e.g. drives other than nvme0.

* Use -f in the newfs_hammer to avoid complaints about small
  partition sizes.

6 years agodrm/linux: Improve linux/pci.h
François Tigeot [Tue, 5 Dec 2017 20:52:31 +0000 (21:52 +0100)]
drm/linux: Improve linux/pci.h

* Add pci_get_drvdata and pci_set_drvdata()

* Add pci_register_driver()

6 years agofdc(4): Check return error from isa_dma_init().
zrj [Tue, 5 Dec 2017 10:49:16 +0000 (12:49 +0200)]
fdc(4): Check return error from isa_dma_init().

Must have been a bad sector in
0bc821c68ba308f17d4f0a7bfd300c432d575a09

Reported-by: dcb <https://bugs.dragonflybsd.org/issues/3108>
6 years agokernel/ipfw3: Minimal K&R cleanup.
Sascha Wildner [Tue, 5 Dec 2017 09:46:57 +0000 (10:46 +0100)]
kernel/ipfw3: Minimal K&R cleanup.

6 years agolibcipher: Minimal K&R cleanup.
Sascha Wildner [Tue, 5 Dec 2017 09:45:11 +0000 (10:45 +0100)]
libcipher: Minimal K&R cleanup.

6 years agolibcrypt: K&R cleanup.
Sascha Wildner [Tue, 5 Dec 2017 09:43:43 +0000 (10:43 +0100)]
libcrypt: K&R cleanup.

6 years agolibcrypt: Remove some __STDC__ checks.
Sascha Wildner [Tue, 5 Dec 2017 09:43:16 +0000 (10:43 +0100)]
libcrypt: Remove some __STDC__ checks.

6 years agortld - Fix ifunc relocations
Matthew Dillon [Tue, 5 Dec 2017 08:14:20 +0000 (00:14 -0800)]
rtld - Fix ifunc relocations

* ifunc relocations had a bug that caused an immediate seg-fault.
  Apparently this type of relocation is not used very much, we've
  never encountered it before.  But the qemu port uses it.

* Fix the bug.  Basically missing lockstate and the top level in
  rtld was expected to hold the bind lock.

6 years agoRemove some orphaned externs.
Sascha Wildner [Tue, 5 Dec 2017 08:05:52 +0000 (09:05 +0100)]
Remove some orphaned externs.

6 years agokernel - Correct vm_map_entry_reserve() call depth issue w/vkernel
Matthew Dillon [Mon, 4 Dec 2017 23:43:13 +0000 (15:43 -0800)]
kernel - Correct vm_map_entry_reserve() call depth issue w/vkernel

* vm_map was making vaddr tests against VM_MAX_USER_ADDRESS to
  determine when to call vm_map_entry_reserve().  This test works
  fine on real hardware but will not work properly for a vkernel.

* Instead, just use td_nest_count to prevent excessive recursion,
  or test against &kernel_map when necessary.

* Fixes SIGILL in vkernel due to stack overflow.

Reported-by: tuxillo
6 years agolibc - Adjust comment
Matthew Dillon [Mon, 4 Dec 2017 20:26:47 +0000 (12:26 -0800)]
libc - Adjust comment

* Adjust comment for dmalloc.c

6 years agodebug - Adjust kmapinfo
Matthew Dillon [Mon, 4 Dec 2017 20:25:27 +0000 (12:25 -0800)]
debug - Adjust kmapinfo

* Use formatnum() for kmapinfo's segment size output when
  printing EMPTY segments.  This makes the output more readable.

6 years agokernel - Expand physical memory support to 64TB
Matthew Dillon [Mon, 4 Dec 2017 07:50:07 +0000 (23:50 -0800)]
kernel - Expand physical memory support to 64TB

* Make NKPML4E truly programmable and change the default from 1
  PDP page to 16 PDP pages.  This increases KVM from 512G to
  8TB, which should be enough to accomodate a maximal 64TB
  configuration.

  Note that e.g. 64TB of physical ram certainly requires more
  than one kernel PDP page, since the vm_page_array alone
  would require around 2TB, never mind everything else!

  PDP entries in the PML4E (512 total @ 512GB per entry):
256 User space
112 (unused, avail for NKPML4E)
128 DMAP (64TB max physical memory)
 16 KVM NKPML4E default (8TB) (recommend 64 max)

* Increase the DMAP from 64 PDP pages to 128 PDP pages, allowing
  support for up to 64TB of physical memory.

* Changes the meaning of KPML4I from being 'the index of the only
  PDP page in the PML4e' to 'the index of the first PDP page in
  the PML4e'.  There are NKPML4E PDP pages starting at index KPML4I.

* NKPDPE can now exceed 512.  This is calculated to be the maximmum
  number of PD pages needed for KVM, which is now (NKPML4E*NPDPEPG-1).

  We now pre-allocate and populate only enough PD pages to accomodate
  the page tables we are pre-installing.  Those, in turn, are calculated
  to be sufficient for bootstrapping mainly vm_page_array and a large
  initial set of pv_entry structures.

* Remove nkpt, it was not being used any more.

6 years agokernel - Remove DISABLE_PSE option and related code
Matthew Dillon [Mon, 4 Dec 2017 03:06:03 +0000 (19:06 -0800)]
kernel - Remove DISABLE_PSE option and related code

* Remove the DISABLE_PSE option and related code.  Note that the pmap
  code was already forcing DISABLE_PSE since we are 64-bit only and
  the pmap code never supported PSE anyway.

6 years agovmstat(8): Change few printf formats for -s.
zrj [Mon, 4 Dec 2017 10:02:16 +0000 (12:02 +0200)]
vmstat(8): Change few printf formats for -s.

Catch up with b7ea2f3f550021bfaabd8ed8d50b54c157fb55bf

6 years agoFix mdoc issues in various manual pages.
Sascha Wildner [Mon, 4 Dec 2017 08:25:32 +0000 (09:25 +0100)]
Fix mdoc issues in various manual pages.

6 years agokernel - Expand page count fields to 64 bits
Matthew Dillon [Mon, 4 Dec 2017 01:14:15 +0000 (17:14 -0800)]
kernel - Expand page count fields to 64 bits

* 32 bit page count fields limit us to 8TB of ram.  Expand to allow
  up to the DMAP limit (32TB).  Do an initial pass on various page
  count fields and change them from int's to long's or vm_pindex_t's.

* Fix a 32-bit overflow in the pv_entry initialization code.

  pv_entry_max = shpgperproc * maxproc + vm_page_array_size;
  2000 * 1046516 + pages_of_phys_memory;

  maxproc is 1046516 @ 512GB.  This calculation overflows its 32
  bit signed variable somewhere between 256G and 512G of ram.  This
  can lead to a zinitna() allocation in pvzone that is much too
  large.

Reported-by: zrj
6 years agotest/vkernel - Add hammer2 support
Antonio Huete Jimenez [Mon, 4 Dec 2017 00:55:05 +0000 (16:55 -0800)]
test/vkernel - Add hammer2 support

- Fix fstab and loader.conf settings.

6 years agodrm/linux: Implement sort()
François Tigeot [Sun, 3 Dec 2017 10:31:41 +0000 (11:31 +0100)]
drm/linux: Implement sort()

6 years agodrm/linux: Fix linux/dma-mapping.h compilation...
François Tigeot [Sun, 3 Dec 2017 09:32:58 +0000 (10:32 +0100)]
drm/linux: Fix linux/dma-mapping.h compilation...

... with the drm/i915 codebase from Linux 4.8

* Move functions from .h to .c files in order to hide their implementations

* This avoids leaking DragonFly headers to Linux code or Linux headers
  to DragonFly code with associated bad side-effects like type redefinitions

6 years agodrm/linux: Add i2c_get_adapdata and i2c_set_adapdata()
François Tigeot [Sun, 3 Dec 2017 07:15:04 +0000 (08:15 +0100)]
drm/linux: Add i2c_get_adapdata and i2c_set_adapdata()

6 years agodrm/linux: Add system_unbound_wq
François Tigeot [Sun, 3 Dec 2017 06:37:04 +0000 (07:37 +0100)]
drm/linux: Add system_unbound_wq

6 years agosyscons: Unbreak compile w/o splash
Sepherosa Ziehau [Sat, 2 Dec 2017 17:13:43 +0000 (01:13 +0800)]
syscons: Unbreak compile w/o splash

6 years agohammer2 - Fix indefinite dirty chains due to rename
Matthew Dillon [Sat, 2 Dec 2017 19:06:30 +0000 (11:06 -0800)]
hammer2 - Fix indefinite dirty chains due to rename

* Do not call cache_setunresolved() on tnch before calling
  cache_rename().  Doing so prevents cache_rename() from properly
  setting VREF_FINALIZE on the vnode.  Since the deleted inode is
  no longer in the chain hierarchy the related chains can remain
  dirty indefinitely (until the vnode is reused or the filesystem
  is unmounted).

* This fixes inmem dirty chain accounting and 'chnmem' stalls.

6 years agohammer2 - optimize hammer2_pfs_memory_wakeup()
Matthew Dillon [Sat, 2 Dec 2017 19:05:58 +0000 (11:05 -0800)]
hammer2 - optimize hammer2_pfs_memory_wakeup()

* Optimize hammer2_pfs_memory_wakeup() to use fetchadd instead of
  cmpset, reducing contention.

6 years agohammer2 - Fix rename race
Matthew Dillon [Sat, 2 Dec 2017 09:02:59 +0000 (01:02 -0800)]
hammer2 - Fix rename race

* Fix a rename race in hammer2 which can result in unwanted namecache
  overloading.  When this occurs, a rename operation can wind up
  deleting the same inode more than once.

* This fixes a "hammer2_chain_getparent: no parent" panic in H2.

Reported-by: kerma
6 years agotop(1): Avoid segfaults on systems with more cores.
zrj [Sat, 2 Dec 2017 17:44:42 +0000 (19:44 +0200)]
top(1): Avoid segfaults on systems with more cores.

Fixes segfaults on 110 cpu core setup for top -M usage.

6 years agosystat(1): Print VNODE PAGER.
zrj [Sat, 2 Dec 2017 17:41:04 +0000 (19:41 +0200)]
systat(1): Print VNODE PAGER.

Looks a bit better.

6 years agohammer2(8): Document cleanup command.
zrj [Sat, 2 Dec 2017 17:40:08 +0000 (19:40 +0200)]
hammer2(8): Document cleanup command.

While there, do some misc whitespace cleanup.

6 years agokernel/vfs: Avoid diagnostic line breaks.
zrj [Sat, 2 Dec 2017 17:38:46 +0000 (19:38 +0200)]
kernel/vfs: Avoid diagnostic line breaks.

6 years ago<sys/systm.h>: Remove bcopyb() prototype.
Sascha Wildner [Sat, 2 Dec 2017 11:49:45 +0000 (12:49 +0100)]
<sys/systm.h>: Remove bcopyb() prototype.

It was a i386 specific function which was removed in f4c88b0b86f1810.

Reported-by: zrj
6 years agopthread_join.3: Document pthread_timedjoin_np() (taken from FreeBSD).
Sascha Wildner [Sat, 2 Dec 2017 08:55:05 +0000 (09:55 +0100)]
pthread_join.3: Document pthread_timedjoin_np() (taken from FreeBSD).

6 years agopthread_mutex_*.3: Add {get,set}prioceiling to SEE ALSO.
Sascha Wildner [Sat, 2 Dec 2017 07:57:03 +0000 (08:57 +0100)]
pthread_mutex_*.3: Add {get,set}prioceiling to SEE ALSO.

6 years agopf.os: Add OpenBSD:6.1
Eitan Adler [Sat, 2 Dec 2017 06:24:18 +0000 (22:24 -0800)]
pf.os: Add OpenBSD:6.1

Obtained From: OpenBSD

6 years agokernel - Increase syscons history buffer size
Matthew Dillon [Sat, 2 Dec 2017 02:56:06 +0000 (18:56 -0800)]
kernel - Increase syscons history buffer size

* Increase syscons history from 4 screens to 10 screens
  (4*rows to 10*rows).

6 years agokernel - Fix CAM unit allocation race
Matthew Dillon [Sat, 2 Dec 2017 02:52:27 +0000 (18:52 -0800)]
kernel - Fix CAM unit allocation race

* Fix an issue where multiple targets can sometimes allocate the
  same unit when probed asynchronously.  The search and insertion
  functions were being independently locked.

6 years agoRevert "Remove advertising clause from pr(1)"
Eitan Adler [Fri, 1 Dec 2017 21:29:50 +0000 (13:29 -0800)]
Revert "Remove advertising clause from pr(1)"

This reverts commit 4586ab4e1609825eb2d4503c92578933c3479239.

See https://www.mail-archive.com/svn-src-all@freebsd.org/msg153666.html

6 years agosyscons - Partly implement FBIO_BLANK ioctl, for display powersaving.
Imre Vadász [Sat, 25 Nov 2017 17:25:54 +0000 (18:25 +0100)]
syscons - Partly implement FBIO_BLANK ioctl, for display powersaving.

* Only implement the parts relevant for display powersaving when a KMS
  graphics driver is loaded. Adjust the existing V_DISPLAY_* constant
  definitions to match the corresponding definitions in FreeBSD.

6 years agoMinor fixes to pr(1)
Eitan Adler [Thu, 30 Nov 2017 08:02:23 +0000 (00:02 -0800)]
Minor fixes to pr(1)

Obtained-from: FreeBSD