Admin queues almost always have several ASYNC_EVENT_REQUEST outstanding.
They have no timeouts, but their presence in qpair->outstanding_tr caused
useless timeout callout rearming twice a second.
While there, relax timeout callout period from 0.5s to 0.5-1s to improve
aggregation. Command timeouts are measured in seconds, so we don't need
to be precise here.
Reviewed by: imp
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D33781
On large systems even relatively rare callouts may fire many times
per second. This should allow them to aggregate better, since we do
not require any precision when polling for media change, etc.
MFC after: 2 weeks
This file does not build for WITHOUT_PF sometimes. Looking at various
ways to cope (it exposes other issues too), but in the mean time just
add it back here to unbreak WITHOUT_PF in some scenarios.
Sponsored by: Netflix
It was needed for mips only, and only kinda sorta worked for mips. It
can be brought back if we grow another architecture that need it.
Sponsored by: Netflix
libsoft was a transition from the FreeBSD/armv6 10 (and earlier) with
'softfp' ABI (that is, hardware float, but passed as integer registers)
to the 'hardfp' ABI that was in FreeBSD/armv[67] in FreeBSD 11 and
newer. It's been off by default since it was created.
This was mostly used by people that wanted to do a source upgrade of
their system from FreeBSD 10 to 11 or from 11-current before the cutover
to 11-current after. This should not be confused with the full software
floating point implementation (that doesn't use the hardware floating
point instructions at all) that is used out of tree by at least one
company selling armv7 gear that has no FPU.
There's no longer a need for the transition, so retire it like should
likely have happened sometime before FreeBSD 12 was released 3 years
ago.
Sponsored by: Netflix
Suggestions by: jrtc27, jhb
Reviewed by: manu, mmel
Differential Revision: https://reviews.freebsd.org/D33761
Previously algorithms such as AES-CBC would provide an algorithm
without a key size for the smallest key size and additional algorithms
with an explicit key size, e.g. "aes-cbc" (128 bits), "aes-cbc192",
and "aes-cbc256".
Instead, always make the key size name explicit and reuse the
"generic" name to request running tests against all of the key sizes.
For example, for AES-CBC this means "aes-cbc128" is now the name of
the variant with a 128-bit key and "aes-cbc" runs tests of AES-CBC
with all three key sizes.
This makes it easier to run tests on all combinations of ciphers like
AES-GCM or AES-CCM with -z in a single invocation.
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D33759
kib@ reported a problem which was resolved by
reverting commit 867c27c23a, which changed the NFS
client to use direct RPCs to the server for
IO_APPEND writes. He also spotted that the
code only invalidated buffer cache buffers
when they were marked NMODIFIED (had been
written into).
This patch modifies the NFS VOP_WRITE() to
always invalidate the buffer cache buffers
and pages for the file when IO_APPEND is
specified. It also includes some cleanup
suggested by kib@.
Reported by: kib
Tested by: kib
Reviewed by: kib
MFC after: 10 weeks
In 4f6c66cc9c, ifa_ifwithnet() was changed to no longer
ifa_ref() the returned ifaddr, and instead the caller was required
to stay in the net_epoch for as long as they wanted the ifaddr
to remain valid. However, this missed the case where an AF_LINK
lookup would call ifaddr_byindex(), which still does ifa_ref()
the ifaddr. This would cause a refcount leak.
Fix this by inlining the relevant parts of ifaddr_byindex() here,
with the ifa_ref() call removed. This also avoids an unnecessary
entry and exit from the net_epoch for this case.
I've audited all in-tree consumers of ifa_ifwithnet() that could
possibly perform an AF_LINK lookup and confirmed that none of them
will expect the ifaddr to have a reference that they need to
release.
MFC after: 2 months
Sponsored by: Dell Inc
Differential Revision: https://reviews.freebsd.org/D28705
Reviewed by: melifaro
Under some INET/INET6 src.conf configurations sbin/route previously
failed to build due to an unused variable warning. It was functionally
write-only anyway, so just remove it.
Reported by: melifaro
Reviewed by: melifaro
MFC after: 3 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D33717
The arch required two-pages alignment due to single TLB entry caching
two consequtive mappings.
Reviewed by: imp
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D33763
The function makevfslist is only called once in mount.c, but should
be save to be called more than once with different parameters.
The bin/df command links against this file, and will need this
possibility to allow -l and -t to be used together.
MFC after: 3 days
- Move busdma_lock_mutex to subr_bus_dma.c.
- Move _busdma_lock_dflt to subr_bus_dma.c. This function was named a
couple of different things previously. It is not a public API but
an internal helper used in place of a NULL pointer. The prototype
is in <sys/bus_dma.h> as not all backends include
<sys/bus_dma_internal.h>.
Reviewed by: kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D33694
Move mostly duplicated code in various MD bus_dma backends to support
bounce pages into sys/kern/subr_busdma_bounce.c. This file is
currently #include'd into the backends rather than compiled standalone
since it requires access to internal members of opaque bus_dma
structures such as bus_dmamap_t and bus_dma_tag_t.
Reviewed by: kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D33684
It is one of the few remaining Giant-locked callouts. It would be
good to remove it, not mentioning that polling itself is not good.
If this cause keyboard/mouse freezes on some hardware, please set
loader tunable hw.atkbd.hz=1 as workaround and report the issue.
Submitted by: imp, jhb
This eliminates error messages like this from the driver when running at
50Gbps with 100G cables:
[3726] cc0: l1cfg failed: 71
[4407] cc0: l1cfg failed: 71
Note that link comes up anyway with or without this change.
Reported by: Suhas Lokesha @ Chelsio
MFC after: 1 week
Sponsored by: Chelsio Communications
In my understanding this is only needed to workaround lost interrupts.
I was thinking to remove it completely, but the comment about edge-
triggered interrupt may be true and needs deeper investigation. ~1Hz
should be often enough to handle the supposedly rare loss cases, but
rare enough to not appear in top. Add sysctl hw.atkbd.hz to tune it.
MFC after: 1 month
The starting address passed to mprotect was wrong, so in the case where
the last page containing the table is not the last page of the BAR, the
wrong region would be unmapped.
Reported by: Andy Fiddaman <andy@omniosce.org>
Reviewed by: jhb
Fixes: 7fa2335347 ("bhyve: Map the MSI-X table unconditionally for passthrough")
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D33739
The PBA and MSI-X table can reside in different BARs.
Reported by: Andy Fiddaman <andy@omniosce.org>
Reviewed by: jhb
Fixes: 7fa2335347 ("bhyve: Map the MSI-X table unconditionally for passthrough")
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D33739
