The variables that are extern in the netmap header file should be
defined in ixl_txrx.c (the file that is included in both ixl(4)/ixlv(4),
not in the main driver source files.
Reported by: ed@, dim@, ngie@
getdtablesize is "trivial global state" and is similar to
getrlimit(RLIMIT_NOFILE), so should be permitted in capability mode.
Reviewed by: oshogbo
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D7719
Calling cap_rights_contains() several times with the same inputs is not
going to produce a different output. The variable being iterated, i, is
never used inside the for loop.
The loop is actually done in cap_rights_contains()
Submitted by: Ryan Moeller <ryan@freqlabs.com>
Reviewed by: oshogbo, ed
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D7369
The swap_pager_swapoff() function uses trylock for the object lock
before pagein, which means that either i/o to md(4) over swap, or
intensive page faults over swap pager objects might prevent swapoff()
from making any progress. Then the retry < 100 check fails and machine
panics.
If trylock fails, acquire the object lock in the blockable way and
restart the hash bucket walk. Keep retries logic for now.
Reported and tested by: pho
Reviewed by: alc, markj
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D7688
Add a new 'netproto' variable which can be set for now to
NET_TFTP or NET_NFS (default to NET_NONE)
From the dhcp options if one sets the root-path option to:
"ip:path", the loader will consider it is booting over NFS
(meaning same behaviour as the default current behaviour)
if the dhcp option "tftp server address" is set (option 150)
the loader will consider it is booting over tftpfs, it will then
consider the root-path options with 2 possible case
1. "path" then the IP of the tftp server will be the one passed by
the option 150, and the files will be retrieved under "path" on the tftp
server
2. "ip:path" then the IP of the tftp server will be the one passed in
the option "overwritting the IP from the option 150.
We could not "abuse" the rootpath option in the form or tftp://ip:path because
this is already used for other purpose by iPXE preventing any chainload from
iPXE to the FreeBSD loader.
Given at each open(), the loader loops over all available filesystems and keep
the "best" error, we needed to prevent tftpfs to fallback on nfs and vice versa.
the tftpfs and nfs implementation in libstand now return EINVAL early if
'netproto' for that purpose.
Reviewed by: tsoome
Relnotes: yes
Sponsored by: Gandi.net
Differential Revision: https://reviews.freebsd.org/D7628
There are two writable hidden properties "iscsioptions" and "stmf_sbd_lu",
that have no default string value. Attempt to unset them or replicate
caused kernel panic. This simple bandaid seems fixes the problem nicely.
MFC after: 2 weeks
Keyboard input needs Giant locking, and that is not possible to do
correctly here. Use mtx_trylock() and proceed unlocked as before if
we can't acquire Giant (non-recursively), except in kdb mode don't
even try to acquire Giant. Everything here is a hack, but it often
works. Even if mtx_trylock() succeeds, this might be a LOR.
Keyboard input also needs screen locking, to handle screen updates
and switches. Add this, using the same simplistic screen locking
as for sc_cnputc().
Giant must be acquired before the screen lock, and the screen lock
must be dropped when calling the keyboard driver (else it would get a
harmless LOR if it tries to acquire Giant). It was intended that sc
cn open/close hide the locking calls, and they do for i/o functions
functions except for this complication.
Non-console keyboard input is still only Giant-locked, with screen
locking in some called functions. This is correct for the keyboard
parts only.
When Giant cannot be acquired properly, atkbd and kbdmux tend to race
and work (they assume that the caller acquired Giant properly and don't
try to acquire it again or check that it has been acquired, and the
races rarely matter), while ukbd tends to deadlock or panic (since it
does the opposite, and has other usb threads to deadlock with).
The keyboard (Giant) locking here does very little, but the screen
locking completes screen locking for console mode except for not
detecting or handling deadlock.
- By default, adjust time upon SYNC request. It can be disabled
through hw.hvtimesync.ignore_sync_req. SYNC request will be
sent by hypervisor the host is resumed, rebooted, etc.
- By default, adjust time upon SAMPLE request, if there is 100ms
difference between VM time and hypervisor time. This can be
disabled through hw.hvtimesync.sample_drift.
And nuke the unnecessary task, since channel callback is running
in a Hyper-V taskqueue nowadays.
Submitted by: YanZhe Chen <t-yachen microsoft com>
Discussed with: Dexuan Cui <decui microsoft com>, Hongjiang Zhang <honzhan microsoft com>, sephe
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7707
something (perhaps in loader.rc.local) that can read in .conf files
after all the other .conf files have been read and override settings
in them. This is quite handy if the .conf file name is determined
while the loader is running, but might be generically useful for other
things. If this hook exists, call it, otherwise don't do anything.
Doing it in these functions ensures that this file is reliably
read. It also works around a defect in forth where s" isn't allowed
outside a function (well, in a compile context) leading to gross
workarounds if one were to hack loader.rc like:
: maybe-some-func s" some-func" sfind if execute else drop then ;
maybe-some-func
which somehow seems worse. Though I'm sure there's some clever forthy
way of doing that with a macro.
Sponsored by: Netflix, Inc
because the bootloader, ie redboot, won't let them do so, and so used the
memory before the kernel for early memory allocation, such as pagetables,
stacks, etc...
Make a bit of an effort to try to get that memory mapped.
Previously this reported an error from Clang 3.9.0: implict conversion
from 'int' to 'char' changes value from 128 to -128.
Discussed with: dim, trasz
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D7699
there is a short period where functions that walk the kernel page table
without locking them may see an invalid entry. One solution would be to add
locking to these functions, however some may be called from locations where
we are unable to sleep.
Until a better solution can be found stop promoting pages in the kernel
pmap so these functions work as expected.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
this, and sc will soon depend on it again.
The on/off request is passed without modification to lower layers,
so the bug was smaller in this layer than in in lower layers (the
sequence on;on;off left polling off when it should be on, but the
sequence on;on;off;on;off... doesn't allow the interrupt handler
to eat the input after an "off" that should't turn off polled mode,
provided lower layers don't have the bug, since this layer is virtual.
The bug was small in lower layers too. Normally everything is Giant
locked for keyboards, and this locks out the interrupt handler in
on;on;off;on;off... sequences. However, PR 211884 says that fixing
this bug in ukbd in r303765 apparently causes the eating-by-interrupt
behaviour that the fix is to prevent.
Discussed with: emax
Restore an splx() lost in r228644. We aren't nearly ready to remove
spl's. They give hints about missing locking. This lost one was
misplaced. Dropping it early for convenience gave race windows for
accesses to the fkey buffer. Giant locking accidentally fixed this
for non-console cases.
Put the spl's around the whole function. Since there are many returns
that would need splx() just before them for a direct fix, split the
function into a wrapper that does the spl's and a "locked" function
that does the work.
Return earlier when no keyboard is attached to match the ordering in a
planned version. This breaks the dubious feature of returning keys
from the fkey buffer after the keyboard has gone away. Losing the keys
wouldn't matter, but we keep them too long now.
The removal of vm_fault_additional_pages() meant that a hard fault on
a swap-backed page would result in only that page being read in. This
change implements readahead and readbehind for the swap pager in
swap_pager_getpages(). swap_pager_haspage() is modified to return the
largest contiguous non-resident range of pages containing the requested
range.
Reviewed by: alc, kib
Tested by: pho
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D7677
Actually all OIDs defined in net/rndis.h are standard NDIS OIDs.
While I'm here, use the verbose macro name as in NDIS spec.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7679
Summary:
1) Attach problem - mpc85xx_probe() relies on fact that 0xfff0 mask matches all
QorIQ CPUs what is not true since e6500. This shall be reworked to match against
all supported CPUs.
2) There is no any reason for operating system to re-program or anyhow else
touch the LAWs programmed by firmware (u-boot). Right now mpc85xx_attach()
removes all LaW entries except for DRAM. This causes MCE to be generated when
later any of driver maps DTB-provided hardware addresses which do not exist
anymore because corresponding LaWs were removed.
Submitted by: Ivan Krivonos <int0dster_AT_gmail.com>
Differential Revision: https://reviews.freebsd.org/D7663
Adding the compatible property check isn't enough. Device trees for eTSEC2
devices are missing a 'reg' property on the eTSEC node itself, relegating it to
the queue group child nodes.
Still left to do: add Multigroup mode support (see QorIQ reference manuals s for
SoCs with eTSEC2).
MFC after: 2 weeks
Relnotes: Yes
Value passed as islocked argument to knlist_add should
be consistent with actual lock state so add lock/unlock
calls around knlist_add
PR: 212120
Submitted by: Manuel Stuhn
device argument to the stage-1 bootloader. In such cases, boot1 would
only try to read the entire device rather than checking for partitions.
Instead of panic'ing, fall back to reading the partitions as normal in
such situations. This was preventing boot of installed systems on some
versions of PowerKVM.
PR: kern/211599
MFC after: 2 days
Add a new 'clear driver' command for devctl along with the accompanying
ioctl and devctl_clear_driver() library routine to reset a device to
use a wildcard devclass instead of a fixed devclass. This can be used
to undo a previous 'set driver' command. After the device's name has
been reset to permit wildcard names, it is reprobed so that it can
attach to newly-available (to it) device drivers.
MFC after: 1 month
Sponsored by: Chelsio Communications
The device quiet flag is not automatically reset on detach, so it is
inherited by other device drivers (e.g. when switching a device driver
over to ppt for PCI pass through). Cope with this behavior by explicitly
marking the device verbose during detach so that the next driver can make
its own decision.
Sponsored by: Chelsio Communications
In r304602, I mistakenly removed the ioat_process_events check that we weren't
processing events before the hardware had completed the descriptor
("last_seen"). Reinstate that logic.
Keep the defensive loop condition and additionally make sure we've actually
completed a descriptor before blindly chasing the ring around.
In reset, queue and finish the startup command before allowing any event
processing or submission to occur. Avoid potential missed callouts by
requeueing the poll later.
can emulate efi_cons_poll(0 with a flag and caching the last key read with
ReadKeyStroke. This fixes the loader.efi countdown timer on Pine64 (and
other U-Boot + EFI using platforms).
Reviewed by: imp, manu
Differential Revision: https://reviews.freebsd.org/D7670
just use the same mutex locking as sc cn putc so they have the same
defects.
The locking calls to acquire the lock are actually in sc cn open and close.
Ungrab has to unlock, although this opens a race window.
Change the direct mutex lock calls in sc cn putc to the new locking
functions via the open and close functions. Putc also has to unlock, but
doesn't keep the screen open like grab. Screen open and close reduce to
locking, except screen open for grab also attempts to switch the screen.
Keyboard locking is more difficult and still null, even when keyboard
input calls screen functions, except some of the functions have locks
too deep to work right.
This organization gives a single place to fix some of the locking.
for zeroing pages in idle where nontemporal writes are clearly best.
This is almost a no-op since zeroing in idle works does nothing good
and is off by default. Fix END() statement forgotten in previous
commit.
Align the loop in sse2_pagezero(). Since it writes to main memory,
the loop doesn't have to be very carefully written to keep up.
Unrolling it was considered useless or harmful and was not done on
i386, but that was too careless.
Timing for i386: the loop was not unrolled at all, and moved only 4
bytes/iteration. So on a 2GHz CPU, it needed to run at 2 cycles/
iteration to keep up with a memory speed of just 4GB/sec. But when
it crossed a 16-byte boundary, on old CPUs it ran at 3 cycles/
iteration so it gave a maximum speed of 2.67GB/sec and couldn't even
keep up with PC3200 memory. Fix the alignment so that it keep up with
4GB/sec memory, and unroll once to get nearer to 8GB/sec. Further
unrolling might be useless or harmful since it would prevent the loop
fitting in 16-bytes. My test system with an old CPU and old DDR1 only
needed 5+ GB/sec. My test system with a new CPU and DDR3 doesn't need
any changes to keep up ~16GB/sec.
Timing for amd64: with 8-byte accesses and newer faster CPUs it is
easy to reach 16GB/sec but not so easy to go much faster. The
alignment doesn't matter much if the CPU is not very old. The loop
was already unrolled 4 times, but needs 32 bytes and uses a fancy
method that doesn't work for 2-way unrolling in 16 bytes. Just
align it to 32-bytes.
When fixing this module to build on PC98, I actually broke the build on
ARM64. On PC98 we need to pull in the sources from the MACHINE_CPUARCH
(i386), but on ARM64 we need to use the MACHINE, as MACHINE_CPUARCH is
set to aarch64 instead of just arm64.
same name as for i386). It is not reconnected yet.
Which method is better is too machine-dependent and system-dependent
to replace the old method unconditionally.
This fixes a tautological pointer comparison warning, but would also a
real bug for a platform where bus_dmamap_unload of a static allocation
is not a no-op.
Summary:
Some device trees put "fsl,ns16650" first in the compatible list. This causes
the probe code to choke, even though the device is compatible with ns16650, and
has it listed later in the tree.
Reviewed by: nwhitehorn
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D7676
beyond the end of disk. r298900 added code to prevent this. Some BIOSes
cause significant delays if asked to read past end-of-disk.
We never trusted the BIOS to accurately report the sectorsize of disks
before and this set of changes. Unfortuately they interact badly with
the infamous >2TB wraparound bugs. We have a number of relatively-recent
machines in the FreeBSD.org cluster where the BIOS reports 3TB disks as 1TB.
With pre-r298900 they work just fine. After r298900 they stop working if
the boot environment attempts to access anything outside the first 1TB on
the disk. 'ZFS: I/O error, all block copies unavailable' etc. It affects
both UFS and ZFS if they try to boot from large volumes.
This change replaces the blind trust of the BIOS end-of-disk reporting
with a read-ahead clip to prevent reads crossing the of end-of-disk
boundary. Since 2^32 (2TB) size reporting truncation is not uncommon,
the clipping is done on 2TB aliases of the reported end-of-disk.
ie: a 3TB disk reported as 1TB has readahead clipped at 1TB, 3TB, 5TB, ...
as one of them is likely to be the real end-of-disk.
This should make the loader on these broken machines behave the same as
traditional pre-r298900 loader behavior, without disabling read-ahead.
PR: 212139
Discussed with: tsoome, allanjude
macro is defined in lots of different places in ipfilter, so replace all
of the nonportable definitions with portable ones.
Pointy hat to: dim
X-MFC-With: r304959, r304953
MFC after: 3 days
This will allow us to perform bhndb(4) bridge configuration based on
the identified hardware, prior to performing full enumeration of the
child bhnd bus.
Approved by: adrian (mentor, implicit)
and negative shift counts.
Fix error messages: print "Division" instead of "Divide"; print
multiplier-like, addition-like and logical operator tokens instead of
garbage (usually the command name).
ddb has a primitive lexer with excessive information hiding that makes
it hard to find even the point in the line where a syntax error is
detected. Old ddb just printed "Syntax error" and this was unimproved
in most places by printing a garbage token.
Early use of vm86 depends on the PIC being reset to mask interrupts,
but r286667 moved PIC initialization to after where vm86 may be first
used.
Move the PIC initialization up to immdiately before vm86 initialization.
All invocations of diff that I tried display this move poorly so that it
looks like PIC and vm86 initialization was moved later.
r286667 was to move console initialization later. The diffs are again
unreadable -- they show a large move that doesn't seem to involve the
console. The PIC initialization stayed just below the console
initialization where it could still be debugged but no longer works.
Later console initialization breaks mainly debugging vm86 initialization
and memory sizing using ddb and printf(). There are several printf()s
in the memory sizing that now go nowhere since message buffer
initialization has always been too late. Memory sizing is done by loader
for most users, but the lost messages for this case are even more
interesting than for an auto-probe since they tell you what the loader
found.
vm86 uses the tss, but r273995 moved tss initialization to after where
it may be first used, just because tss_esp0 now depends on later
initializations and/or amd64 does it later.
vm86 is first used for memory sizing in cases where the loader can't
figure out the size or is not used. Its initialization is placed
immediately before memory sizing to support this, and the tss was
initialized a little earlier.
Move everything in the tss initialization except for tss_esp0 back to
almost where it was, immediately before vm86 initialization (the
combined move is from before dblflt_tss initialization to after). Add
only early initialization of tss_esp0, later reloading of the tss, and
comments. The initial tss_esp0 no longer has space for the pcb since
initially the size of the pcb is not known and no pcb is needed.
(Later changes broke debugging at this point, so the nonexistent pcb
cannot be used by debuggers, and at the time of 273995 when ddb was
almost able to debug this problem it didn't need the pcb.) The
iniitial tss_esp0 still has a magic 16 bytes reserved for vm86
although I think this is unused too.
cnv API is a set of functions for managing name/value pairs by cookie.
The cookie can be obtained by nvlist_next(), nvlist_get_parent() or
nvlist_get_pararr() function. This patch also includes unit tests.
Submitted by: Adam Starak <starak.adam@gmail.com>
handling.
- Extended PWRCTL/PMU APIs to support querying clock frequency during very
early boot, prior to bus attach.
- Implement generic PMU-based calculation of UART rclk values.
- Replaced use of static frequency tables (bcm_socinfo) with
runtime-determined values.
Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D7552
- Added bhnd_pmu driver implementations for PMU and PWRCTL chipsets,
derived from Broadcom's ISC-licensed HND code.
- Added bhnd bus-level support for routing per-core clock and resource
power requests to the PMU device.
- Lift ChipCommon support out into the bhnd module, dropping
bhnd_chipc.
Reviewed by: mizhka
Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D7492
This adds support for performing platform_reset() on all supported
devices, using early boot enumeration of chipc capabilities and
available cores.
- Added Broadcom-specific MIPS CP0 register definitions used by
BCM4785-specific reset handling.
- Added a bcm_platform structure for tracking chipc/pmu/cfe platform
data.
- Extended the BCMA EROM API to support early boot lookup of core info
(including port/region mappings).
- Extended platform_reset() to support PMU, PMU+AOB, and non-PMU
devices.
Reviewed by: mizhka
Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D7539
Rather than enabling the I/O MMU when the vmm module is loaded,
defer initialization until the first attempt to pass a PCI device
through to a guest. If the I/O MMU fails to initialize or is not
present, than fail the attempt to pass a PCI device through to a
guest.
The hw.vmm.force_iommu tunable has been removed since the I/O MMU is
no longer enabled during boot. However, the I/O MMU support can be
disabled by setting the hw.vmm.iommu.enable tunable to 0 to prevent
use of the I/O MMU on any systems where it is buggy.
Reviewed by: grehan
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D7448
can be in after receiving a FIN.
FWIW, NetBSD has this change for quite some time.
This has been tested at Netflix and Limelight in production traffic.
Reported by: Sam Kumar <samkumar99 at gmail.com> on transport@
Reviewed by: rrs
MFC after: 4 weeks
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D7475
- Return appropriate error code instead of ENOMEM when sosend() fails in
send_mpa_req.
- Fix for problematic race during destroy_qp.
- Abortive close in the failure of send_mpa_reject() instead of normal close.
- Remove the unnecessary doorbell flowcontrol logic.
Submitted by: Krishnamraju Eraparaju at Chelsio
MFC after: 1 month
Sponsored by: Chelsio communications
I-FORWARD-TSN chunk before any DATA or I-DATA chunk.
Thanks to Julian Cordes for finding this problem and prividing
packetdrill scripts to reporduce the issue.
MFC after: 3 days
And use new RNDIS set to configure NDIS offloading parameters.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7641
And switch MAC address query to use new RNDIS query function.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7639
Summary:
First time BSS is cleared in booke_init(), Second time it's cleared in
powerpc_init(). Any variable initialized between two those guys gets wiped out
what is wrong. In particular it wipes tlb1_entries initialized by tlb1_init(),
which was fine when tlb1_init() was called a second time, but this was removed
in r304656.
Submitted by: Ivan Krivonos <int0dster_gmail.com>
Differential Revision: https://reviews.freebsd.org/D7638
'show active trace', or 'acttrace' for short, prints backtraces from running
threads only.
Reviewed by: mjg
Differential Revision: https://reviews.freebsd.org/D7646
hardware send and receive PDU limits. Report these limits to ICL and
take them into account when setting the socket's send and receive buffer
sizes. The driver used a single hardcoded limit everywhere prior to
this change.
Sponsored by: Chelsio Communications
obliterate possible error from sleep with errors from
umtxq_check_susp(), when looping to clear URWLOCK_{READ,WRITE}_WAITERS.
Noted and reviewed by: vangyzen
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
If there was some error, e.g. the sleep was interrupted, as in the
referenced PR, do_rw_rdlock() did not cleared URWLOCK_READ_WAITERS.
Since unlock only wakes up write waiters when there is no read
waiters, for URWLOCK_PREFER_READER kind of locks, the result was
missed wakeups for writers.
In particular, the most visible victims are ld-elf.so locks in
processes which loaded libthr, because rtld locks are urwlocks in
prefer-reader mode. Normal rwlocks fall into prefer-reader mode only
if thread already owns rw lock in read mode, which is not typical and
correspondingly less visible. In the PR, unowned rtld bind lock was
waited for in the process where only one thread was left alive.
Note that do_rw_wrlock() correctly clears URWLOCK_WRITE_WAITERS in
case of errors.
Reported and tested by: longwitz@incore.de
PR: 211947
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
the l2 entry is a block type and not an l3 page.
While here fix the string to correct the level name and add a missing ')'.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
syscons spinlock for the output routine alone. It is better to extend
the coverage of the first syscons spinlock added in r162285. 2 locks
might work with complicated juggling, but no juggling was done. What
the 2 locks actually did was to cover some of the missing locking in
each other and deadlock less often against each other than a single
lock with larger coverage would against itself. Races are preferable
to deadlocks here, but 2 locks are still worse since they are harder
to understand and fix.
Prefer deadlocks to races and merge the second lock into the first one.
Extend the scope of the spinlocking to all of sc_cnputc() instead of
just the sc_puts() part. This further prefers deadlocks to races.
Extend the kdb_active hack from sc_puts() internals for the second lock
to all spinlocking. This reduces deadlocks much more than the other
changes increases them. The s/p,10* test in ddb gets much further now.
Hide this detail in the SC_VIDEO_LOCK() macro. Add namespace pollution
in 1 nested #include and reduce namespace pollution in other nested
#includes to pay for this.
Move the first lock higher in the witness order. The second lock was
unnaturally low and the first lock was unnaturally high. The second
lock had to be above "sleepq chain" and/or "callout" to avoid spurious
LORs for visual bells in sc_puts(). Other console driver locks are
already even higher (but not adjacent like they should be) except when
they are missing from the table. Audio bells also benefit from the
syscons lock being high so that audio mutexes have chance of being
lower. Otherwise, console drviver locks should be as low as possible.
Non-spurious LORs now occur if the bell code calls printf() or is
interrupted (perhaps by an NMI) and the interrupt handler calls
printf(). Previous commits turned off many bells in console i/o but
missed ones done by the teken layer.
tso_segsz pkthdr field during RX processing, and use the information in TCP for
more correct accounting and as a congestion control input. This is only a start,
and an audit of other uses for the data is left as future work.
Reviewed by: gallatin, rrs
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D7564
set later in the function. This fixes a potential NULL pointer dereference
found on arm64.
Obtained from: ABT Systems Ltd
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
to what the 32-bit arm code does, with the exception that it always assumes
the tag is non-coherent.
Tested by: jmcneill
Obtained from: ABT Systems Ltd
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
- Increasing queue depth gives ~100% performance improvement for
randwrite fio test in Azure.
- New channel selection, which takes LUN id and the current cpuid
into consideration, gives additional ~20% performance improvement
for ranwrite fio test in Azure.
Submitted by: Hongzhang Jiang <honzhan microsoft com>
Modified by: sephe
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7622
Decouple the send and receive limits on the amount of data in a single
iSCSI PDU. MaxRecvDataSegmentLength is declarative, not negotiated, and
is direction-specific so there is no reason for both ends to limit
themselves to the same min(initiator, target) value in both directions.
Allow iSCSI drivers to report their send, receive, first burst, and max
burst limits explicitly instead of using hardcoded values or trying to
derive all of them from the receive limit (which was the only limit
reported by the drivers prior to this change).
Display the send and receive limits separately in the userspace iSCSI
utilities.
Reviewed by: jpaetzel@ (earlier version), trasz@
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D7279
indicate (potentially partial) success of the open. Use these to
decide what to close in sccnclose(). Only grab/ungrab use open/close
so far.
Add a per-sc variable to count successful keyboard opens and use
this instead of the grab count to decide if the keyboad state has
been switched.
Start fixing the locking by using atomic ops for the most important
counter -- the grab level one. Other racy counting will eventually
be fixed by normal mutex or kdb locking in most cases.
Use a 2-entry per-sc stack of states for grabbing. 2 is just enough
to debug grabbing, e.g., for gets(). gets() grabs once and might not
be able to do a full (or any) state switch. ddb grabs again and has
a better chance of doing a full state switch and needs a place to
stack the previous state. For more than 3 levels, grabbing just
changes the count. Console drivers should try to switch on every i/o
in case lower levels of nesting failed to switch but the current level
succeeds, but then the switch (back) must be completed on every i/o
and this flaps the state unless the switch is null. The main point
of grabbing is to make it null quite often. Syscons grabbing also
does a carefully chosen screen focus that is not done on every i/o.
Add a large comment about grabbing.
Restore some small lost comments.
- in sccnopen(), open the keyboard before the screen. The keyboard
currently requires Giant (although it must be spinlocked to work
correctly as a console), so the previous order would be a LOR if
it has any semblance of locking.
- add a (currently dummy) state arg to scgetc().
As the support for large blocks was enabled in loader zfs code, the
heap in userboot was left not changed, resulting with failure of detecting
and accessing zfs pools for bhyve virtual machines.
This fix does set the heap to use same amount of memory as the zfsloader
is using. To make it possible to test and verify loader functions, bhyve
is providing very useful option, but it also means, we like to keep feature
parity with [zfs]loader as close as possible.
PR: 212038
Reported by: dfh0522@gmail.com
Reviewed by: allanjude, grehan
Approved by: allanjude (mentor)
Differential Revision: https://reviews.freebsd.org/D7635
The issue was introduced with adding support for salted checksums, and
was revealed by bhyve userboot.so.
During pool discovery the loader is reading pool label from disks, and
at that time the spa structure is not yet set up, so the NULL pointer
is passed for spa. This condition must be checked to avoid the corruption
of the memory and NULL pointer dereference.
PR: 212114
Reported by: tsoome@freebsd.com
Reviewed by: allanjude
Approved by: allanjude (mentor)
Differential Revision: https://reviews.freebsd.org/D7634
needed before enabling superpages on arm64. This code is based on the amd64
pmap with changes as needed to handle the differences between the two
architectures.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
iterate over superpages. We don't yet create these, but soon will.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Where the cloudabi64 kernel can be used to execute 64-bit CloudABI
binaries, this one should be used for 32-bit binaries. Right now it
works on i386 and amd64.
A nice thing about requiring a vDSO is that it makes it incredibly easy
to provide full support for running 32-bit processes on 64-bit systems.
Instead of letting the kernel be responsible for composing/decomposing
64-bit arguments across multiple registers/stack slots, all of this can
now be done in the vDSO. This means that there is no need to provide
duplicate copies of certain system calls, like the sys_lseek() and
freebsd32_lseek() we have for COMPAT_FREEBSD32.
This change imports a new vDSO from the CloudABI repository that has
automatically generated code in it that copies system call arguments
into a buffer, padding them to eight bytes and zero-extending any
pointers/size_t arguments. After returning from the kernel, it does the
inverse: extracting return values, in the process truncating
pointers/size_t values to 32 bits.
Obtained from: https://github.com/NuxiNL/cloudabi
The native CloudABI data types header file used to be pulled in by the
vDSOs when they were still written in C. Since they are now all
rewritten in assembly, this can go away.
In all of these source files, the userspace pointer size corresponds
with the kernelspace pointer size, meaning that casting directly works.
As I'm planning on making 32-bit execution on 64-bit systems work as
well, use TO_PTR() here as well, so that the changes between source
files remain minimal.
Use sbintime_t timeouts with precision control to get very accurate
timing. It costs little to always ask for about 1% accuracy, and the
not so new event timer implementation usual delivers that, and when
it can't it gets much closer than our previous coarse timeouts and
buggy simple countdown.
The 2 fastest atkbd repeat rates have periods 34 and 38 msec, and ukbd
pretended to support rates in between these. This requires
sub-microsecond precision and accuracy even to handle the 4 msec
difference very well, but ukbd asked the timeout subsystem for timeouts
of 25 msec and the buggy simple countdown of this gave a a wide range
of precisions and accuracies depending on HZ and other timer
configuration (sometimes better than 25 msec but usually more like 50
msec). We now ask for and usually get precision and accuracy of about
1% for each repeat and much better on average.
The 1% accuracy is overkill. Rounding of 30 cps to 34 msec instead of
33 already gives an error of +2% instead of -1%, and ut AT keyboards on
PS/2 interfaces have similar errors.
A timeout is now scheduled for every keypress and release. This allows
some simplifications that are not done. It allows removing the timeout
scheduling for exiting polled mode where it was unsafe in ddb mode. This
is done. Exiting polled mode had some problems with extra repeats. Now
exiting polled mode lets an extra timeout fire and the state is fudged
so that the timeout handler does very little.
The sc->time_ms variable is unsigned to avoid overflow. Differences of
it need to be signed. Signed comparisons were emulated by testing an
emulated sign bits. This only works easily for '<' comparisonss, but
we now need a '<=' comparison. Change the difference variable to
signed and use a signed comparison. Using unsigned types here didn't
prevent overflow bugs but just reduced them. Overflow occurs with
n repeats at the silly repeat period of [U]INT_MAX / n. The old countdown
had an off by 1 error, and the simplifications would simply count down
1 to 0 and not need to accumulate possibly-large repeat repeats.
And stringent input IC version negotiate message checks.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7614
RESET is not used by the hn(4) at all, and RESET_CMPLT does not even
have a rid to match with the pending requests. So, let's put it
onto an independent switch branch and log a warning about it.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7602
Summary:
Kernel maps only one page of FDT. When FDT is more than one page in size, data
TLB miss occurs on memmove() when FDT is moved to kernel storage
(sys/powerpc/booke/booke_machdep.c, booke_init())
This introduces a pmap_early_io_unmap() to complement pmap_early_io_map(), which
can be used for any early I/O mapping, but currently is only used when mapping
the fdt.
Submitted by: Ivan Krivonos <int0dster_gmail.com>
Differential Revision: https://reviews.freebsd.org/D7605
While I'm here, sort the RNDIS status in ascending order.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7594
The thermal sensor lives in the touch screen controller. Touch screen part
isn't done for now.
Temperature is read every ~2 seconds and exposed via sysctl.
kbdcontrol -r fast is documented to give a non-emulated atkbd's fastest
rate of 250.34, but is misimplemented to request this as 0.0. ukbd
supports many nonstandard rates, although it is currently too inaccurate
by a factor of several hundred for non-huge nonstandard rates to be
useful. It mapped 0.0 to 200.0. A repeat delay of 0 means a rate of
infinity which is quite fast, but physical constraints limit this to
a few MHz and the inaccuracies made it almost usable.
Convert 0.0 to the documented 250.34.
Also convert negative args and small args to the 250.34 minimal ones,
like atkbd does. This is for KDSETREPEAT -- the 2 versions of the
deprecated KDSETRAD have bounds checking. Keep not doing any bounds
checking or conversions for upper limits since nonstandard large
delays are useful for testing.
The inaccuracies are dependent on HZ and the timeout implementation.
With the old timeout implementation and HZ = 1000, 200.0 probably
worked better to emulate 250.34 than 250.34 itself. HZ = 100 gives
roundoff errors that accidentally reduce the inaaccuracies, and
event timers reduce the inaccuracies even more, so 200.0 was giving
more like itself (perhaps 215.15 on average but sometimes close to
10 msec repeat which is noticebly too fast). This commit makes 0.0
noticeably too slow, like 250.34 always was.
An optimization is in place to skip reading the .depend.* files with
'make install'. This was too strong and broke 'make all install' and
'make foo.o foo install'. Now only skip reading the dependency files
if all make targets ran are install targets.
The problem comes about because headers are only added in as a guessed
dependency if .depend.* files do not yet exist. If they do exist, even
if being skipped from being read, then the header dependencies are not
applied. This applies to all #included files, and not just headers.
Reported by: kib
MFC after: 1 day
Sponsored by: EMC / Isilon Storage Division
handling. This resulted in the window target being left uninitialized
when an underflow occured.
Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D7617
finding the vm_page_t in pmap_extract_and_hold. Previously it would return
the vm_page_t of the first page in a block. This would cause issues when,
for example, fsck reads from a device into the middle of a superpage. In
this case the read call would write to the start of the block, and not to
the buffer passed in.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Allow netbooting on efi without having to setup any NFS server by rebuilding the
loader with LOADER_TFTP_SUPPORT like for the i386 pxeloader
Sponsored by: Gandi.net
This code should be able to support later AMD chipsets as well, but that
hasn't been tested.
SB800 supports accessing several different SMBus buses using the same
set of constrol registeirs plus special PMIO registers that control which
bus is selected. This could be exposed to consumers as several smb devices
each talking to its bus. This feature is not implemented yet.
MFC after: 2 weeks
Summary:
There is no need to call tlb1_init() twice. Now it is called first time from
booke_init() and second time from powerpc_init() (where it is under BOOKE
switch). Although this does not cause immediate problems in the mainline kernel,
this can lead to undesirable side effects like two TLB entries with the same VA
in the TLB1. Presence of two TLB entries with the same VA can hang CPU.
Test Plan:
Add initial mapping for UART to the tlb1_init(), build and boot the kernel,
ensure that mapping presents only once (most convinient way - through Lauterbah
or similar hardware debugger)
Submitted by: Ivan Krivonos <int0dster_gmail.com>
Differential Revision: https://reviews.freebsd.org/D7607
Summary: Current booke/pmap code ignores mas7 and mas8 on e6500 CPU.
Submitted by: Ivan Krivonos <int0dster_gmail.com>
Differential Revision: https://reviews.freebsd.org/D7606
So that Hyper-V can leverage them instead of rolling its own definition.
Discussed with: hps
Reviewed by: hps
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7592
Move msix_disable_migration under #ifdef SMP since it doesn't make sense
for !SMP kernels.
PR: 212014
Reported by: Glyn Grinstead <glyn@grinstead.org>
MFC after: 3 days
its own job because this breaks the simplified QEMU XHCI TRB parser,
which expects the complete USB control transfer as a series of back to
back TRBs. The old behaviour is kept under #ifdef in case this change
breaks enumeration of any USB devices.
PR: 212021
MFC after: 1 week
Copy over amd64's cloudabi64_sysvec.c into i386 and tailor it to work.
Again, we use a system call convention similar to FreeBSD, except that
there is no support for indirect system calls (%eax == 0).
Where i386 differs from amd64 is that we have to store thread/process
entry arguments on the stack instead of using registers. We also have to
put an extra pointer on the stack for TLS (for GSBASE). Place that
pointer in the empty slot that is normally used to hold return
addresses. That seems to keep the code simple.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D7590
The previous fix was tested mainly on 3 AT keyboards with USB adaptors where
it works. 1 USB keyboard doesn't translate Alt-PrintScreen, so the software
has to do it.
Reorganize a little to share some code and to not translate the unusual usb
scan code0x8a unless an Alt modified is set. Remove redundant check of Alt
modifiers. Translation now more clearly filters out Alt-PrintScreen before
the check.
The table of errors fixed in the previous commit had many bugs. Correct
table:
K_RAW Ctl-PrintScreen: E0-2A-E0-37 -> E0-37
K_RAW Alt-PrintScreen (with 4 comb. of Ctl/Shift): 79 -> 54
K_RAW Pause/Break (with 4 comb. of Alt/Shift): E0-46 -> E1-1D-45
K_CODE PrintScreen (with 4 comb. of Ctl/Shift): 54 -> 5c
K_CODE Alt-PrintScreen (with 4 comb. of Ctl/Shift): 7e -> 54
K_CODE Pause/Break (with 8 comb. of Ctl/Alt/Shift): 6c -> 68
That is 25 of 32 shift combinations for 2 keys fixed. All 16 combinations
were broken for K_CODE and thus also for K_XLATE.
r304436 attempted to optimize the handling of incoming UDP packet by only
making an expensive call to in_broadcast() if the mbuf was marked as an
broadcast packet. Unfortunately, this cannot work in the case of point-to-
point L2 protocols like PPP, which have no notion of "broadcast".
Discussions on how to properly fix r304436 are ongoing, but in the meantime
disable the optimization to ensure that no existing network setups are broken.
Reported by: bms
pagetable is supported more will be added soon to support removing
superpages.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
is_completion_pending governs whether or not a callout will be scheduled
when new work is queued on the IOAT device. If true, a callout is
already scheduled, so we do not need a new one. If false, we schedule
one and set it true. Because resetting the hardware completed all
outstanding work but failed to clear is_completion_pending, no new
callout could be scheduled after a reset with pending work.
This resulted in a driver hang for polled-only work.
physical address. This is required when either mapping is writeable.
While here remove an unneeded call to pmap_pde, we already have the pde
from earlier in the function.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
configuring of EP0 and non-EP0 into xhci_cmd_evaluate_ctx() and
xhci_cmd_configure_ep() respectivly. This resolves some errors when
using XHCI under QEMU and gets is more in line with the XHCI
specification.
PR: 212021
MFC after: 1 week
And don't recreate chimney sending buffer for each primary channel
open, it is now created in device_attach DEVMETHOD and destroyed
in device_detach DEVMETHOD.
MFC after: 1 week
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D7574
First, keep a ref count on the stcb after looking it up, as
done in the other lookup cases.
Second, before looking again at sp, ensure that it is not
freed, because the assoc is about to be freed.
MFC after: 3 days
It seems Killer E2200/E2400 has a BIOS misconfiguration or silicon
bug which triggers DMA write errors when driver uses advertised
maximum payload size. Force the maximum payload size to 128 bytes
in DMA configuration.
This change should fix occasional DMA write errors reported on
Killer E2200.
Tested by: <psy0nic@sys-tek.org>
controllers. For Gigabit Ethernet version of AR816x, AR813x/AR815x
except L1D controller, use vendor recommended ASPM parameters.
While here, increase alc_dma_burst array size. Broken H/W can
return bogus value in theory.
so they are memory independent which allows for handling panics
triggered by the keyboard driver itself, typically via CTRL+ALT+ESC
sequences. Or if the USB keyboard driver was processing a key at the
moment of panic. Allow UKBD to be attached while keyboard polling is active.
Tested by: Bruce Evans <bde@freebsd.org>
MFC after: 1 week
everything was broken. The cases that I noticed were Ctrl-PrintScreen
not being mapped to the virtual scancode 0x5c (debug) and Pause not being
mapped to the physical/virtual scancode 0x46 (slock).
These keys are the most complicated ones due to kludges to give some
compatibility back to before AT keyboards.
Alt-PrintScreen must pretend to be a separate key from PrintScreen
even at the "raw" level. The (unique) usb code for it is 0x8a and we
just have to map this to our unique virtual scancode 0x54, but we
mapped it first to the internal code 0x7e and then to 0x79 which is a
key on the Japanese 106/109 keyboard. This fix is under the
UKBD_EMULATE_ATASCANCODE option which shouldn't be used for non-AT
keyboards. If it is, then the syscons Japanese keymaps have nothing
of importance for code 0x79 and can easily be changed. 0x54 is also
unimportant in Japanese and US keymaps.
NonAlt-PrintScreen and NonCtl-Pause/Break had many much larger bugs with
smaller compatibility problems from fixing them. The details are too
ugly to give here. Summary of the changed (hex) codes:
K_RAW PrintScreen (Ctl, Shift, Ctl-Shift): E0-2A-E0-37 -> E0-37
K_RAW Alt-PrintScreen (all shift states): 79 -> 54
K_RAW Pause/Break (unshifted, Shift, Alt, Alt-Shift)): E0-46 -> E1-1D-45
K_CODE ALT-PrintScreen (all shift states): 79 -> 54
That is 15 of 32 shift combinations for 2 keys fixed, with 8 easy cases
from the 79 -> 54 remapping.
The difference is only large and with no workaround using a keymap for
for K_RAW, but this affects other modes when ukbd is layered under kbmux
because kbmux keeps all subdevices in K_RAW mode and translates. Oops.
I used kbdmux to generate the above table of changes.
Essentially, this is a literal copy of the code in sys/compat/cloudabi64,
except that it now makes use of 32-bits datatypes and limits. In
sys/conf/files, we now need to take care to build the code in
sys/compat/cloudabi if either COMPAT_CLOUDABI32 or COMPAT_CLOUDABI64 is
turned on.
This change does not yet include any of the CPU dependent bits. Right
now I have implementations for running i386 binaries both on i386 and
x86-64, which I will send out for review separately.
Right now we're casting uint64_t's to native pointers. This isn't
causing any problems right now, but if we want to provide a 32-bit
compatibility layer that works on 64-bit systems as well, this will
cause problems. Casting a uint32_t to a 64-bit pointer throws a compiler
error.
Introduce a TO_PTR() macro that casts the value to uintptr_t before
casting it to a pointer.
Though uio_resid is of type ssize_t, we need to take into account that
this source file contains an implementation specific to a certain
userspace pointer size. If this file provided 32-bit implementations,
this should have used INT32_MAX, even when running a 64-bit kernel.
This change has no effect, but is simply in preparation for adding
support for running 32-bit CloudABI executables.
On 32-bit platforms, our 64-bit timestamps need to be split up across
two registers. A simple assignment to td_retval[0] will cause the top 32
bits to get lost. By using memcpy(), we will automatically either use 1
or 2 registers depending on the size of register_t.
The reason why the old vDSOs were written in C using inline assembly was
purely because they were embedded in the C library directly as static
inline functions. This was practical during development, because it
meant you could invoke system calls without any library dependencies.
The vDSO was simply a copy of these functions.
Now that we require the use of the vDSO, there is no longer any need for
embedding them in C code directly. Rewriting them in assembly has the
advantage that they are closer to ideal (less useless branching, less
assumptions about registers remaining unclobbered by the kernel, etc).
They are also easier to build, as they no longer depend on the C type
information for CloudABI.
Obtained from: https://github.com/NuxiNL/cloudabi
This was .. an interesting headache.
There are two halves:
* The earlier IRIX stuff (yes, early) occasionally would do dead
code removal and generate multiple consecutive LO16 entries.
If this is done for REL entries then it's fine - there's no
state kept between them. But gcc 5.x seems to do this for
RELA entries.
eg:
HI1 LO1 HI2 LO2 LO3 HI4 LO4
.. in this instance, LO2 should affect HI2, but LO3 doesn't at all
affect anything. The matching HI3 was in code that was deleted
as "dead code".
Then, the next one:
* A "GCC extension" allows for multiple HI entries before a LO entry;
and all of those HI entries use the first LO entry as their basis
for RELA offset calculations.
It does this so GCC can also do dead code deletion without necessarily
having to geneate fake relocation entries for balanced HI/LO RELA
entries.
eg:
HI1 LO1 HI2 HI3 HI4 LO4 LO5 HI6 LO6 LO7
in this instance, HI{2,3,4} are the same relocation as LO4 (eg .bss)
and need to be buffered until LO4 - then the RELA offset is applied
from LO4 to HI{2,3,4} calculations.
/And/, the AHL from HI4 is used during the LO4 relocation calculation,
just like in the normal (ie, before this commit) implementation.
Then, LO5 doesn't trigger anything - the HI "buffer" is empty,
so there are no HI relocations to flush out.
HI6/LO6 are normal, and LO7 doesn't trigger any HI updates.
Tested:
* AR9344 SoC, kernel modules, using gcc-5.3 (mips-gcc-5.3.0 package)
Notes:
* Yes, I do feel dirty having written this code.
Reviewed by: imp (after a handful of "this should be on fire" moments wrt gcc and this code)
The default value of the tunable introduced in r304436 couldn't be
effectively overrided on VIMAGE kernels, because instead of being
accessed via the appropriate VNET() accessor macro, it was accessed
via the VNET_NAME() macro, which resolves to the (should-be) read-only
master template of initial values of per-VNET data. Hence, while the
value of udp_require_l2_bcast could be altered on per-VNET basis, the
code in udp_input() would ignore it as it would always read the default
value (one) from the VNET master template.
Silence from: rstone
The ip6_output routine is missing L2 cache invalication as done
in ip_output. Even with that code, some problems with UDP over
IPv6 have been reported. Diabling L2 cache for that problem works
around the problem for now.
PR: 211872 211926
Reviewed by: gnn
Approved by: gnn (mentor)
MFC after: immediate
(NB: This was likely a mismerge from XNU in audit support, where the
text argument to setlogin(2) is captured -- but as a text token,
whereas this change uses the dedicated login-name field in struct
audit_record.)
MFC after: 2 weeks
Sponsored by: DARPA, AFRL
ftruncate(2) system call. This was not required by the Common
Criteria, which needed only open-time audit.
MFC after: 2 weeks
Sponsored by: DARPA, AFRL
strings provided by user/config files. This update is replacing sprintf with
snprintf for cases the command_errbuf is built from dynamic content.
PR: 211958
Reported by: ecturt@gmail.com
Reviewed by: imp, allanjude
Approved by: imp (mentor), allanjude (mentor)
Differential Revision: https://reviews.freebsd.org/D7563
This driver only supports 10Mb Ethernet using PIO (the hardware supports
DMA, but the driver only does PIO). There are not any PCCard adapters
supported by this driver, only ISA cards. In addition, it does not use
bus_space but instead uses bcopy with volatile pointers triggering a
host of warnings. (if_ie.c is one of 3 files always built with
-Wno-error)
Relnotes: yes
to generate one. This is was U-Boot does to generate a random MAC so we end
up with the same MAC address as if U-Boot did generate it.
MFC after: 1 week
This hardware is not present on any modern systems. The driver is quite
hackish (raw inb/outb instead of bus_space, and raw inb/outb to random
I/O ports to enable ACPI since it predated proper ACPI support).
Relnotes: yes
The wl(4) driver supports pre-802.11 PCCard wireless adapters that
are slower than 802.11b. They do not work with any of the 802.11
framework and the driver hasn't been reported to actually work in a
long time.
Relnotes: yes
The si(4) driver supported multiport serial adapters for ISA, EISA, and
PCI buses. This driver does not use bus_space, instead it depends on
direct use of the pointer returned by rman_get_virtual(). It is also
still locked by Giant and calls for patch testing to convert it to use
bus_space were unanswered.
Relnotes: yes
