Using the device pager with /dev/kmem is not stable since KVA mappings
are transient, but the device pager caches the PA associated with a
given offset forever. Interestingly, mips' implementation of
memmap() already refused requests for /dev/kmem.
Note that kvm_read/kvm_write do not use mmap, but use read and write on
/dev/kmem, so this should not affect libkvm users.
Reviewed by: kib
MFC after: 2 months
etc. can find out where the SMBIOS entry point is located. In pure
UEFI mode the BIOS is not mapped into the standard address space so the
SMBIOS table might not appear between 0xf0000 and 0xfffff. The
UEFI environment can report this the location of the anchor. If it is
reported then expose it as hint.smbios.0.mem. This can then be used
by other tools. However, we should make smbios(4) useful and have it
take this value and provide accesor function so ipmi(4) etc. don't
have to parse and figure things about the SMBIOS table. I have some
simple patches to smbios(4) to expose this address as sysctl and
for ipmi(4) to get the base address. However, the real fix is to
have ipmi(4) ask smbios(4) for what it wants and have smbios(4)
parse it out and return it. This would make smbios(4) useful and reduce
duplicated code. If this address doesn't point to the anchor then
finding SMBIOS info. will fail as if this didn't exist. So there should
be no harm.
With this change and the following hack, dmidecode works on a bunch of
UEFI machines that I tested:
if kenv hint.smbios.0.mem > /dev/null
then
mkdir -p /sys/firmware/efi
mount -t tmpfs -o size=8k tmpfs /sys/firmware/efi
echo "SMBIOS=`kenv hint.smbios.0.mem`" > /sys/firmware/efi/systab
fi
Linux exposes this information via the /sys/firmware/efi/systab file which
dmidecode looks at. We should update dmidecode to do this the FreeBSD
way when we determine what that is!
Reviewed by: jhb
functions to call at the appropriate time to register new forth
words. In the past we've done this with ifdef soup, but now if the
file is included in the build, we'll get the new forth words.
Use this new functionality to move the pci bios stuff out of loader.c
by moving it to biospci.c.
Move the pnp functionality to common/pnp.c.
Move the inb/outb forth words to the i386 sysdep.c file where their
implementation is defined.
Adjust the efi linker scripts and build machinery to cope.
his should be an invisible change to forth scripts and user
experience.
Differential Revision: https://reviews.freebsd.org/D8145
than to store the location of a forth word that is subsequently never
used. It was last used before the 2.03 ficl upgrade in r51786. It was
only used from r43614 (so Feb-Sept 1999) on head and in the 3.x branch
(merged r43715 3.1 -> EOL). Remove it since nobody cared enough to
report the bug in the last 18 years rather than fix it. It's need
seems to have passed in the 2.03 ficl update.
Differential Revision: https://reviews.freebsd.org/D8150
replaced by /boot/loader.rc for 3.1 (r42682). In May 2000, this was
documented as deprecated (r61942) (between FreeBSD 4.0 and
4.1). Remove it since it's not been the preferred method in 17 years
and has been deprecated for 16.
Differential Revision: https://reviews.freebsd.org/D8142
received on a TCP session that has entered the ESTABLISHED state. This
results in a lot of calls to reset the keepalive timer.
This patch changes the behavior so we set the keepalive timer for the
keepalive idle time (TP_KEEPIDLE). When the keepalive timer fires, it will
first check to see if the session has been idle for TP_KEEPIDLE ticks. If
not, it will reschedule the keepalive timer for the time the session will
have been idle for TP_KEEPIDLE ticks.
For a session with regular communication, the keepalive timer should fire
approximately once every TP_KEEPIDLE ticks. For sessions with irregular
communication, the keepalive timer might fire more often. But, the
disruption from a periodic keepalive timer should be less than the regular
cost of resetting the keepalive timer on every packet.
(FWIW, this change saved approximately 1.73% of the busy CPU cycles on a
particular test system with a heavy TCP output load. Of course, the
actual impact is very specific to the particular hardware and workload.)
Reviewed by: gallatin, rrs
MFC after: 2 weeks
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D8243
6988 spa_sync() spends half its time in dmu_objset_do_userquota_updates
Using a benchmark which creates 2 million files in one TXG, I observe
that the thread running spa_sync() is on CPU almost the entire time we
are syncing, and therefore can be a performance bottleneck. About 50% of
the time in spa_sync() is in dmu_objset_do_userquota_updates().
The problem is that dmu_objset_do_userquota_updates() calls
zap_increment_int(DMU_USERUSED_OBJECT) once for every file that was
modified (or created). In this benchmark, all the files are owned by the
same user/group, so all 2 million calls to zap_increment_int() are
modifying the same entry in the zap. The same issue exists for the
DMU_GROUPUSED_OBJECT.
We should keep an in-memory map from user to space delta while we are
syncing, and when we finish, iterate over the in-memory map and modify
the ZAP once per entry. This reduces the number of calls to
zap_increment_int() from "number of objects modified" to "number of
owners/groups of modified files".
This reduced the time spent in spa_sync() in the file create benchmark
by ~33%, from 11 seconds to 7 seconds.
Closes#107
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Steve Gonczi <steve.gonczi@delphix.com>
Reviewed by: Ned Bass <bass6@llnl.gov>
Reviewed by: Jinshan Xiong <jinshan.xiong@intel.com>
Author: Matthew Ahrens <mahrens@delphix.com>
openzfs/openzfs@5fc46359c5
5120 zfs should allow large block/gzip/raidz boot pool (loader project)
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Yuri Pankov <yuri.pankov@nexenta.com>
Reviewed by: Andrew Stormont <andyjstormont@gmail.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Toomas Soome <tsoome@me.com>
openzfs/openzfs@c8811bd3e2
FreeBSD still does not support booting from gzip-compressed datasets,
so keep one chunk of this commit out.
Using a benchmark which creates 2 million files in one TXG, I observe
that the thread running spa_sync() is on CPU almost the entire time we
are syncing, and therefore can be a performance bottleneck. About 50% of
the time in spa_sync() is in dmu_objset_do_userquota_updates().
The problem is that dmu_objset_do_userquota_updates() calls
zap_increment_int(DMU_USERUSED_OBJECT) once for every file that was
modified (or created). In this benchmark, all the files are owned by the
same user/group, so all 2 million calls to zap_increment_int() are
modifying the same entry in the zap. The same issue exists for the
DMU_GROUPUSED_OBJECT.
We should keep an in-memory map from user to space delta while we are
syncing, and when we finish, iterate over the in-memory map and modify
the ZAP once per entry. This reduces the number of calls to
zap_increment_int() from "number of objects modified" to "number of
owners/groups of modified files".
This reduced the time spent in spa_sync() in the file create benchmark
by ~33%, from 11 seconds to 7 seconds.
Closes#107
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Steve Gonczi <steve.gonczi@delphix.com>
Reviewed by: Ned Bass <bass6@llnl.gov>
Reviewed by: Jinshan Xiong <jinshan.xiong@intel.com>
Author: Matthew Ahrens <mahrens@delphix.com>
openzfs/openzfs@5fc46359c5
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Yuri Pankov <yuri.pankov@nexenta.com>
Reviewed by: Andrew Stormont <andyjstormont@gmail.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Toomas Soome <tsoome@me.com>
openzfs/openzfs@c8811bd3e2
Submitted by: Hongjiang Zhang <honzhan microsoft com>
Reported by: Lili Deng <v-lide microsoft com>
MFC after: 3 days
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D8238
RPI3 kernel config builds kernel compatible with latest upstream device
tree and firmware: https://github.com/raspberrypi/firmware/tree/master/boot
As of today it's 597c662a613df1144a6bc43e5f4505d83bd748ca
Default console is PL01x, so pi3-disable-bt dt overlay should be configured
in config.txt and stock U-Boot should be patched to use proper serial port.
Yet unsupported: SMP, VCHIQ, RNG driver. RNG requires some work due to
upstream device tree incompatibility.
Multiple people contributed to this work over time: db@, loos@, manu@
Cases other than MK_* (e.g. ${MACHINE_CPUARCH} == "i386") have been left
as is.
Reviewed by: imp
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D8246
- Revert BUS_SPACE_PHYSADDR back to rman_get_start. BUS_SPACE_PHYSADDR was
introduced in 2013 as temporary wrapper until proper solution appears.
It's ARM only and since we need this file for ARM64 build and no proper
API has been introduced - just revert the change and make sure it's
going to appear when people grep for BUS_SPACE_PHYSADDR in sources.
- Fix printf format for size_t variables
dev_net.c code.
The NETIF_OPEN_CLOSE_ONCE flag was added in r201932 to prevent that behaviour
on some architectures (sparc64 and powerpc64) the default was left to always
open and close the device for each open and close of a file by the loader
because it was necessary for u-boot on arm.
Since it has been added, the flag was turned on for every arches including the
u-boot loader for arm.
This also fixes netbooting on RPi3 (tested by gonzo@)
For the loader.efi it greatly speeds up netbooting
Reviewed by: emaste, gonzo, tsoome
Approved by: gonzo
MFC after: 1 month
Sponsored by: Gandi.net
Differential Revision: https://reviews.freebsd.org/D8230
Ignore the ECN bits on 'tos' and 'set-tos' and allow to use
DCSP names instead of having to embed their TOS equivalents
as plain numbers.
Obtained from: OpenBSD
Sponsored by: OPNsense
Differential Revision: https://reviews.freebsd.org/D8165
compiler-rt's complex division support routines contain calls to
compiler builtins such as `__builtin_scalbnl`. Unfortunately Clang
turns these back into a call to `scalbnl`.
For now link libm's C version of the required support routines.
Reviewed by: ed
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D8190
They are not yet connected to the build, but I am adding them to allow
for easier testing, ports exp-runs, etc.
Reviewed by: ed
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D8188
Back in 2015 when I reimplemented these functions to use an AVL tree, I
was annoyed by the weakness of the typing of these functions. Both tree
nodes and keys are represented by 'void *', meaning that things like the
documentation for these functions are an absolute train wreck.
To make things worse, users of these functions need to cast the return
value of tfind()/tsearch() from 'void *' to 'type_of_key **' in order to
access the key. Technically speaking such casts violate aliasing rules.
I've observed actual breakages as a result of this by enabling features
like LTO.
I've filed a bug report at the Austin Group. Looking at the way the bug
got resolved, they made a pretty good step in the right direction. A new
type 'posix_tnode' has been added to correspond to tree nodes. It is
still defined as 'void' for source-level compatibility, but in the very
far future it could be replaced by a proper structure type containing a
key pointer.
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D8205
buildkernel failed with GCC 5.3 with
error: comparison of constant '852736' with boolean expression is always true
Sponsored by: The FreeBSD Foundation
Linux's copy of the Cavium SDK does not have these non-ASCII characters
and this reduces noise in diffs when comparing the two.
Sponsored by: The FreeBSD Foundation
Suppose that we have an exclusively busy page, and a thread which can
accept shared-busy page. In this case, typical code waiting for the
page xbusy state to pass is
again:
VM_OBJECT_WLOCK(object);
...
if (vm_page_xbusied(m)) {
vm_page_lock(m);
VM_OBJECT_WUNLOCK(object); <---1
vm_page_busy_sleep(p, "vmopax");
goto again;
}
Suppose that the xbusy state owner locked the object, unbusied the
page and unlocked the object after we are at the line [1], but before we
executed the load of the busy_lock word in vm_page_busy_sleep(). If it
happens that there is still no waiters recorded for the busy state,
the xbusy owner did not acquired the page lock, so it proceeded.
More, suppose that some other thread happen to share-busy the page
after xbusy state was relinquished but before the m->busy_lock is read
in vm_page_busy_sleep(). Again, that thread only needs vm_object lock
to proceed. Then, vm_page_busy_sleep() reads busy_lock value equal to
the VPB_SHARERS_WORD(1).
In this case, all tests in vm_page_busy_sleep(9) pass and we are going
to sleep, despite the page being share-busied.
Update check for m->busy_lock == VPB_UNBUSIED in vm_page_busy_sleep(9)
to also accept shared-busy state if we only wait for the xbusy state to
pass.
Merge sequential if()s with the same 'then' clause in
vm_page_busy_sleep().
Note that the current code does not share-busy pages from parallel
threads, the only way to have more that one sbusy owner is right now
is to recurse.
Reported and tested by: pho (previous version)
Reviewed by: alc, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D8196
