mav [Fri, 5 Dec 2014 07:25:02 +0000 (07:25 +0000)]
MFC r274805:
Make cfiscsi_offline() synchronous, waiting for connections termination
before return. This should make ctld restart more clean and predictable.
In this mode one head is in Active state, supporting all commands, while
another is in Standby state, supporting only minimal LUN discovery subset.
It is still incomplete since Standby state requires reservation support,
which is impossible to do right without having interlink between heads.
But it allows to run some basic experiments.
delphij [Fri, 5 Dec 2014 00:32:33 +0000 (00:32 +0000)]
MFC r274172 (avg)
fix l2arc compression buffers leak
We have observed that arc_release() can be called concurrently with a
l2arc in-flight write.
Also, we have observed that arc_hdr_destroy() can be called from
arc_write_done() for a zio with ZIO_FLAG_IO_REWRITE flag in similar
circumstances.
Previously the l2arc headers would be freed while leaking their
associated compression buffers. Now the buffers are placed on
l2arc_free_on_write list for delayed freeing. This is similar to what
was already done to arc buffers that were supposed to be freed
concurrently with in-flight writes of those buffers.
In addition to fixing the discovered leaks this change also adds some
protective code to assert that a compression buffer associated with a
l2arc header is never leaked.
A new kstat l2_cdata_free_on_write is added. It keeps a count of
delayed compression buffer frees which previously would have been leaks.
Tested by: Vitalij Satanivskij <satan@ukr.net> et al
Requested by: many
Sponsored by: HybridCluster / ClusterHQ
delphij [Thu, 4 Dec 2014 23:20:44 +0000 (23:20 +0000)]
MFC r274276: MFV r274271:
Improve zdb -b performance:
- Reduce gethrtime() call to 1/100th of blkptr's;
- Skip manipulating the size-ordered tree;
- Issue more (10, previously 3) async reads;
- Use lighter weight testing in traverse_visitbp();
Initialize tqent_flags in the userland taskq implementation. Without
this the assertion of tq->tq_freelist != NULL may fail in taskq_destroy.
The problem is that tqent_flags is never initialized in the userland
implementation while the kernel one does initialize it. Without proper
initialization, the flag may have its lowest bit set, making it treated
as TQENT_FLAG_PREALLOC and never removing taskq_ent_t from tq_freelist.
dim [Thu, 4 Dec 2014 01:21:56 +0000 (01:21 +0000)]
MFC r275366:
Pull in r209785 from upstream libc++ trunk (by Marshall Clow):
Fix a problem exposed by r208825, which caused bind (and other bits of
libc++) to stop working. And tests
This fix is needed to support clang 3.5.0 and higher, which are more
strict about forming pointer-to-function types with cv-qualifiers or
ref-qualifiers. See also the upstream PR <http://llvm.org/PR19742> and
<http://llvm.org/viewvc/llvm-project?rev=208825&view=rev>
dim [Thu, 4 Dec 2014 01:10:50 +0000 (01:10 +0000)]
MFC r274922:
Fix the following -Werror warning from clang 3.5.0, while building the
ath kernel module:
sys/dev/ath/ath_hal/ar5212/ar5212_reset.c:2642:7: error: taking the absolute value of unsigned type 'unsigned int' has no effect [-Werror,-Wabsolute-value]
if (abs(lp[0] * EEP_SCALE - target) < EEP_DELTA) {
^
sys/dev/ath/ah_osdep.h:74:18: note: expanded from macro 'abs'
#define abs(_a) __builtin_abs(_a)
^
sys/dev/ath/ath_hal/ar5212/ar5212_reset.c:2642:7: note: remove the call to '__builtin_abs' since unsigned values cannot be negative
sys/dev/ath/ah_osdep.h:74:18: note: expanded from macro 'abs'
#define abs(_a) __builtin_abs(_a)
^
1 error generated.
This warning occurs because both lp[0] and target are unsigned, so the
subtraction expression is also unsigned, and calling abs() is a no-op.
However, the intention was to look at the absolute difference between
the two unsigned quantities. Introduce a small static function to
clarify what we're doing, and call that instead.
Reviewed by: adrian
Differential Revision: https://reviews.freebsd.org/D1212
hselasky [Wed, 3 Dec 2014 09:08:04 +0000 (09:08 +0000)]
MFC r275328 and r275330:
Add missing libraries when linking and fix building of some infiniband
utilities by updating some header file locations and compiler include
directives.
brooks [Tue, 2 Dec 2014 22:04:27 +0000 (22:04 +0000)]
MFC r274821:
Merge from CheriBSD:
commit d0c7d235c09fc65dbdb278e7016a96f79c6a49cc
Make the Altera JTAG UART device driver slightly more forgiving of
the foibles of a sub-par hrdware interface by increasing the timeout
for spotting JTAG polling from one to two seconds.
commit 8edfe803f033cc8e33229f99894c2b7496a44d5f
Add a comment about a device-driver race condition that could cause the BERI
pipeline to wedge awaiting JTAG in the event that both the low-level console
and the tty layer decide to write to the JTAG FIFO just before JTAG is
disconnected. Resolving this race is a bit tricky as it looks like there
isn't a way to 'give the character back' to the tty layer when we discover
the race. The easy fix is to drop the character, which we don't yet do, but
perhaps should as that is a better outcome than wedging the pipeline.
tijl [Tue, 2 Dec 2014 14:21:06 +0000 (14:21 +0000)]
MFC r269634:
drm: fix usage of vm_phys_fictitious_to_vm_page
vm_phys_fictitious_to_vm_page should not be called directly, even when
operating on a range that has been registered using
vm_phys_fictitious_reg_range. PHYS_TO_VM_PAGE should be used instead
because on arches that use VM_PHYSSEG_DENSE the page might come
directly from vm_page_array.
Reported by: nwhitehorn
Tested by: nwhitehorn, David Mackay <davidm.jx8p@gmail.com>
Sponsored by: Citrix Systems R&D
tijl [Tue, 2 Dec 2014 14:09:54 +0000 (14:09 +0000)]
MFC r273862,273902:
Port the TTM AGP backend to the FreeBSD agp driver and enable AGP support
in the radeonkms driver.
Note: In PCI mode virtual addresses on the graphics card that map to system
RAM are translated to physical addresses by the graphics card itself. In
AGP mode address translation is done by the AGP chipset so fictitious
addresses appear on the system bus. For the CPU cache management to work
correctly when the CPU accesses this memory it needs to use the same
fictitious addresses (and let the chipset translate them) instead of using
the physical addresses directly.
tijl [Tue, 2 Dec 2014 13:46:13 +0000 (13:46 +0000)]
MFC r273856,273863,273963-273965
- Add two new functions to the AGP driver KPI to bind/unbind arbitrary sets
of pages into the GTT.
- Avoid possible overflow in agp_generic_alloc_memory.
- In agp(4) avoid the need to flush all cpu caches with wbinvd between
updating the GTT and flushing the AGP TLB by storing the GTT in
write-combining memory.
- In agp_amd_bind_page don't flush the AGP TLB. It's done by the calling
function.
- agp_generic_unbind_memory: flush AGP TLB before unwiring pages
agp_bind_pages: assert that pages have been wired down
Currently, the TCP slow timer can starve TCP input processing while it
walks the list of connections in TIME_WAIT closing expired connections
due to contention on the global TCP pcbinfo lock.
To remediate, introduce a new global lock to protect the list of
connections in TIME_WAIT. Only acquire the TCP pcbinfo lock when
closing an expired connection. This limits the window of time when
TCP input processing is stopped to the amount of time needed to close
a single connection.
hselasky [Mon, 1 Dec 2014 07:34:25 +0000 (07:34 +0000)]
MFC r274918:
Don't use the synchronization endpoint unless referenced by the
isochronous endpoint descriptor used for the data transfers, hence the
synchronization feature might not be supposed to be supported [yet].
This makes seamless playback synced with the USB HOST clock work with
the DN32-USB module for Midas audio systems and possibly other similar
products from Klark Teknik.
bryanv [Sat, 29 Nov 2014 23:05:01 +0000 (23:05 +0000)]
MFC r274325:
Enable LRO by default when available on vtnet interfaces
The prior change to not enable LRO by default has confused several
people. The configurations where LRO is problematic is not the
typical use case for VirtIO, and due to other issues, this often
requires checksum offloading to be disabled anyways.
trasz [Sat, 29 Nov 2014 15:41:55 +0000 (15:41 +0000)]
MFC r273849:
Add support for "timeo", "actimeo", "noac", and "proto" options
to mount_nfs(8). They are implemented on Linux, OS X, and Solaris,
and thus can be expected to appear in automounter maps.
trasz [Sat, 29 Nov 2014 15:32:15 +0000 (15:32 +0000)]
MFC r273813:
Add discovery-filter. This makes it possible to restrict which targets
are returned during discovery based on initiator portal, name, and CHAP
credentials.
hselasky [Fri, 28 Nov 2014 19:21:46 +0000 (19:21 +0000)]
MFC r274017, r274088 and r275205:
Provide an on-stack temporary buffer for small IOCTL requests.
Avoiding a memory allocation per IOCTL request can give a significant
speedup for applications which heavily rely on IOCTLs.
trasz [Thu, 27 Nov 2014 10:45:55 +0000 (10:45 +0000)]
MFC r273768:
Remove the distinction between strings and numbers from ctld(8) yacc parser.
This fixes problems with passing strings that look like numbers to clauses
that expect strings; previously it caused syntax errors and had to be worked
by user, using quotes. The workaround introduced in r267833 is no longer
neccessary.
hselasky [Wed, 26 Nov 2014 09:37:35 +0000 (09:37 +0000)]
MFC r273774:
Make some infiniband example utilities easily buildable:
- Add new Makefiles.
- Add more include directories when building.
- Fixed a printf() formatting string.
dteske [Tue, 25 Nov 2014 13:47:53 +0000 (13:47 +0000)]
MFC r274116:
Add new libraries/utilities for data throughput visualization.
dpv(3): dialog progress view library
dpv(1): stream data from stdin or multiple paths with dialog progress view
figpar(3): configuration file parsing library
MFC r274120, r274121, r274123, r274124, r274144, r274146, r274159, r274192,
r274203, r274209, r274226, r274270, and r274851: Fixes following r274116
Reviews: D714
Relnotes: New libdpv/libfigpar and dpv(1) utility
Reviewed by: jelischer, shurd
Discussed at: MeetBSD California 2014 Vendor/Dev Summit
Discussed on: -current
Thanks to: ngie, ian, jelischer, shurd, bapt
dim [Tue, 25 Nov 2014 13:29:13 +0000 (13:29 +0000)]
MFC r274900:
Fix the following -Werror warnings from clang 3.5.0, while building
bsnmpd's snmp_hostres module:
usr.sbin/bsnmpd/modules/snmp_hostres/hostres_snmp.c:204:20: error: absolute value function 'abs' given an argument of type 'const long' but has parameter of type 'int' which may cause truncation of value [-Werror,-Wabsolute-value]
str[9] = (u_char)(abs(tm->tm_gmtoff) / 3600);
^
usr.sbin/bsnmpd/modules/snmp_hostres/hostres_snmp.c:204:20: note: use function 'labs' instead
str[9] = (u_char)(abs(tm->tm_gmtoff) / 3600);
^~~
labs
usr.sbin/bsnmpd/modules/snmp_hostres/hostres_snmp.c:205:22: error: absolute value function 'abs' given an argument of type 'const long' but has parameter of type 'int' which may cause truncation of value [-Werror,-Wabsolute-value]
str[10] = (u_char)((abs(tm->tm_gmtoff) % 3600) / 60);
^
usr.sbin/bsnmpd/modules/snmp_hostres/hostres_snmp.c:205:22: note: use function 'labs' instead
str[10] = (u_char)((abs(tm->tm_gmtoff) % 3600) / 60);
^~~
labs
Since tm::tm_gmtoff is a long, use labs(3) instead.
dim [Tue, 25 Nov 2014 13:12:45 +0000 (13:12 +0000)]
MFC r274898:
Fix the following -Werror warnings from clang 3.5.0, while building
usr.sbin/rtadvd:
usr.sbin/rtadvd/rtadvd.c:1291:7: error: taking the absolute value of unsigned type 'unsigned int' has no effect [-Werror,-Wabsolute-value]
abs(preferred_time - pfx->pfx_pltimeexpire) > rai->rai_clockskew) {
^
usr.sbin/rtadvd/rtadvd.c:1291:7: note: remove the call to 'abs' since unsigned values cannot be negative
abs(preferred_time - pfx->pfx_pltimeexpire) > rai->rai_clockskew) {
^~~
usr.sbin/rtadvd/rtadvd.c:1324:7: error: taking the absolute value of unsigned type 'unsigned int' has no effect [-Werror,-Wabsolute-value]
abs(valid_time - pfx->pfx_vltimeexpire) > rai->rai_clockskew) {
^
usr.sbin/rtadvd/rtadvd.c:1324:7: note: remove the call to 'abs' since unsigned values cannot be negative
abs(valid_time - pfx->pfx_vltimeexpire) > rai->rai_clockskew) {
^~~
2 errors generated.
These warnings occur because both preferred_time and pfx_pltimeexpire
are uint32_t's, so the subtraction expression is also unsigned, and
calling abs() is a no-op.
However, the intention was to look at the absolute difference between
the two unsigned quantities. Introduce a small static function to
clarify what we're doing, and call that instead.
dim [Tue, 25 Nov 2014 12:58:21 +0000 (12:58 +0000)]
MFC r274856:
Avoid undefined behaviour in gas's rotate_left() macro for n == 0.
Otherwise, clang can effectively remove the first iteration of the for
loops where this macro is invoked, and as a result, "cmp r0, #99" fails
to assemble.
dim [Tue, 25 Nov 2014 12:45:31 +0000 (12:45 +0000)]
MFC r274847:
Fix the following -Werror warnings from clang 3.5.0, while building
usr.bin/locate:
usr.bin/locate/locate/util.c:249:29: error: taking the absolute value of unsigned type 'unsigned int' has no effect [-Werror,-Wabsolute-value]
MAXPATHLEN, abs(i) < abs(htonl(i)) ? i : htonl(i));
^
usr.bin/locate/locate/util.c:249:29: note: remove the call to 'abs' since unsigned values cannot be negative
MAXPATHLEN, abs(i) < abs(htonl(i)) ? i : htonl(i));
^~~
usr.bin/locate/locate/util.c:274:32: error: taking the absolute value of unsigned type 'unsigned int' has no effect [-Werror,-Wabsolute-value]
MAXPATHLEN, abs(word) < abs(htonl(word)) ? word :
^
usr.bin/locate/locate/util.c:274:32: note: remove the call to 'abs' since unsigned values cannot be negative
MAXPATHLEN, abs(word) < abs(htonl(word)) ? word :
^~~
The problem is that ntohl() always returns an unsigned quantity. In
this case, it's expected to be cast back to a signed integer, but to
stop complaints about abs() we just store it into an integer, and don't
call ntohl() again.
dim [Tue, 25 Nov 2014 12:19:05 +0000 (12:19 +0000)]
MFC r274846:
Fix the following -Werror warning from clang 3.5.0, while building
usr.bin/cpio on amd64 (or any arch with 64-bit time_t):
contrib/libarchive/cpio/cpio.c:1143:6: error: absolute value function 'abs' given an argument of type 'long' but has parameter of type 'int' which may cause truncation of value [-Werror,-Wabsolute-value]
if (abs(mtime - now) > (365/2)*86400)
^
contrib/libarchive/cpio/cpio.c:1143:6: note: use function 'labs' instead
if (abs(mtime - now) > (365/2)*86400)
^~~
labs
1 error generated.
This is because time_t is a long on amd64. To avoid the warning, just
copy the equivalent test from a few lines before, which is used in the
Windows case, and which is type safe.
grehan [Mon, 24 Nov 2014 01:25:19 +0000 (01:25 +0000)]
MFC r274407
Fix incorrect reading of 32-bit modinfo by 64-bit loaders.
The various structures in the mod_metadata set of a FreeBSD kernel and
modules contain pointers. The FreeBSD loader correctly deals with a
mismatch in loader and kernel pointer size (e.g. 32-bit i386/ppc
loader, loading 64-bit amd64/ppc64 kernels), but wasn't dealing with
the inverse case where a 64-bit loader was loading a 32-bit kernel.
Simply remove the tmp dir before creating the symlink. We are doing
chroots all over the place, so there is bound to be a stale file lying
around in there (in my case samba lock files from creating accounts). If
we don't do that, the symlink later will fail.
dumbbell [Sat, 22 Nov 2014 17:37:51 +0000 (17:37 +0000)]
drm: Take vt(4) default mode from loader tunables
By default, vt(4) gets the "preferred mode" from DRM, when using a DRM
video driver as its backend. The preferred mode is usually the native
screen resolution.
Now, if this mode isn't appropriate, a user can use loader tunables to
select a mode. The tunables are read in the following order:
1. kern.vt.fb.modes.$connector_name
2. kern.vt.fb.default_mode
For example, to set a 1024x768 mode, no matter the connector:
kern.vt.fb.default_mode="1024x768"
To set a 800x600 mode only on the laptop builtin screen:
kern.vt.fb.modes.LVDS-1="800x600"
Beside r274031, this MFC includes:
r274049:
drm: When reading connector mode tunables, list connectors
... and their associated tunables. This gives a way to know the list of
available connectors, no matter the driver.
The problem is that xrandr(1) can list connectors but it uses a
different naming.
r274050:
vt(4): Document kern.vt.fb.default_mode and kern.vt.fb.modes.*
Those tunables are used to set a specific mode in vt(4) instead of using
the default mode.
dumbbell [Sat, 22 Nov 2014 17:10:57 +0000 (17:10 +0000)]
vt(4): Fix keyboard allocation when kbdmux(4) isn't used
The problem was that only the kbdmux keyboard index was saved in
vd->vd_keyboard. This index is -1 when kbdmux isn't used. In this
case, the keyboard was correctly allocated, but the returned index was
discarded.
dumbbell [Sat, 22 Nov 2014 16:55:55 +0000 (16:55 +0000)]
vt(4): Adjust the cursor position after changing the window size
A new terminal_set_cursor() is added: it wraps the existing
teken_set_cursor() function.
In vtbuf_grow(), the cursor position is adjusted at the end of the
function. In vt_change_font(), we call terminal_set_cursor() just after
terminal_set_winsize_blank(), while the terminal is mute.
This fixes a bug where, after loading a kernel video driver which
increases the terminal window size, the cursor remains at its old
position, in other words, in the middle of the display content.