trasz [Sat, 3 Oct 2020 12:03:08 +0000 (12:03 +0000)]
Move KTRUSERRET() from userret() to ast(). It's a really long
detour - it writes ktrace entries to the filesystem - so the overhead
of ast() won't make any difference.
melifaro [Sat, 3 Oct 2020 10:47:17 +0000 (10:47 +0000)]
Introduce scalable route multipath.
This change is based on the nexthop objects landed in D24232.
The change introduces the concept of nexthop groups.
Each group contains the collection of nexthops with their
relative weights and a dataplane-optimized structure to enable
efficient nexthop selection.
Simular to the nexthops, nexthop groups are immutable. Dataplane part
gets compiled during group creation and is basically an array of
nexthop pointers, compiled w.r.t their weights.
With this change, `rt_nhop` field of `struct rtentry` contains either
nexthop or nexthop group. They are distinguished by the presense of
NHF_MULTIPATH flag.
All dataplane lookup functions returns pointer to the nexthop object,
leaving nexhop groups details inside routing subsystem.
User-visible changes:
The change is intended to be backward-compatible: all non-mpath operations
should work as before with ROUTE_MPATH and net.route.multipath=1.
All routes now comes with weight, default weight is 1, maximum is 2^24-1.
Current maximum multipath group width is statically set to 64.
This will become sysctl-tunable in the followup changes.
Using functionality:
* Recompile kernel with ROUTE_MPATH
* set net.route.multipath to 1
Next steps:
* Land outbound hashing for locally-originated routes ( D26523 ).
* Fix net/bird multipath (net/frr seems to work fine)
* Add ROUTE_MPATH to GENERIC
* Set net.route.multipath=1 by default
markj [Fri, 2 Oct 2020 19:16:06 +0000 (19:16 +0000)]
vm_pageout: Avoid rounding down the inactive scan target
With helper page daemon threads, enabled by default in r364786, we
divide the inactive target by the number of threads, rounding down, and
sum the total number of pages freed by the threads. This sum is
compared with the original target, but by rounding down we might lose
pages, causing the page daemon control loop to conclude that inactive
queue scanning isn't keeping up with demand for free pages. Typically
this results in excessive swapping.
Fix the problem by accounting for the error in the main pagedaemon
thread's target. Note that by default the problem will manifest only in
systems with >16 CPUs in a NUMA domain.
Reviewed by: cem
Discussed with: dougm
Reported and tested by: dhw, glebius
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26610
markj [Fri, 2 Oct 2020 19:04:29 +0000 (19:04 +0000)]
uma: Use the bucket cache for cross-domain allocations
uma_zalloc_domain() allocates from the requested domain instead of
following a first-touch policy (the default for most zones). Currently
it is only used by malloc_domainset(), and consumers free returned items
with free(9) since r363834.
Previously uma_zalloc_domain() worked by always going to the keg for an
item. As a result, the use of UMA zone caches was unbalanced: we free
items to the caches, but always allocate from the keg, skipping the
caches.
Make some effort to allocate from the UMA caches when performing a
cross-domain allocation. This avoids blowing up the caches when
something is performing many transient allocations with
malloc_domainset().
Reported and tested by: dhw, glebius
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26427
markj [Fri, 2 Oct 2020 19:04:09 +0000 (19:04 +0000)]
uma: Use LIFO for non-SMR bucket caches
When SMR was introduced, zone_put_bucket() was changed to always place
full buckets at the end of the queue. However, it is generally
preferable to use recently used buckets since their items are more
likely to be resident in cache. So, for buckets that have no constraint
on item reuse, use a last-in-first-out ordering as we did before.
Reviewed by: rlibby
Tested by: dhw, glebius
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26426
markj [Fri, 2 Oct 2020 18:35:55 +0000 (18:35 +0000)]
newlocale(3): Fix a memory leak.
newlocale() optionally takes a "base" locale, from which components not
specified in the mask are inherited. POSIX says that newlocale() may
modify "base" and return it, or free "base" and return a newly allocated
locale. We were not doing either, so applications which use newlocale()
to modify an existing base locale end up leaking memory on FreeBSD.
This diff fixes the leak by releasing a reference to the base locale
before returning. This is less efficient than modifying "base"
directly, but is simpler for an initial bug fix. Also, update the man
page to clarify behaviour with respect to "base".
PR: 249416
MFC after: 3 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26522
manu [Fri, 2 Oct 2020 18:26:41 +0000 (18:26 +0000)]
linuxkpi: Add backlight support
Add backlight function to linuxkpi.
Graphics drivers expose the backlight of the panel directly so allow them to use the backlight subsystem so
user can use backlight(8) to configure them.
Reviewed by: hselasky
Relnotes: yes
Differential Revision: The FreeBSD Foundation
manu [Fri, 2 Oct 2020 18:18:01 +0000 (18:18 +0000)]
Add backlight subsystem
This is a simple subsystem that allow drivers to register as a backlight.
Each backlight creates a device node under /dev/backlight/backlightX and
an alias based on the name provided.
Relnotes: yes
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26250
markj [Fri, 2 Oct 2020 17:50:22 +0000 (17:50 +0000)]
Implement sparse core dumps
Currently we allocate and map zero-filled anonymous pages when dumping
core. This can result in lots of needless disk I/O and page
allocations. This change tries to make the core dumper more clever and
represent unbacked ranges of virtual memory by holes in the core dump
file.
Add a new page fault type, VM_FAULT_NOFILL, which causes vm_fault() to
clean up and return an error when it would otherwise map a zero-filled
page. Then, in the core dumper code, prefault all user pages and handle
errors by simply extending the size of the core file. This also fixes a
bug related to the fact that vn_io_fault1() does not attempt partial I/O
in the face of errors from vm_fault_quick_hold_pages(): if a truncated
file is mapped into a user process, an attempt to dump beyond the end of
the file results in an error, but this means that valid pages
immediately preceding the end of the file might not have been dumped
either.
The change reduces the core dump size of trivial programs by a factor of
ten simply by excluding unaccessed libc.so pages.
PR: 249067
Reviewed by: kib
Tested by: pho
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26590
markj [Fri, 2 Oct 2020 17:49:13 +0000 (17:49 +0000)]
Simplify the check for non-dumpable VM object types
OBJT_DEFAULT, _SWAP, _VNODE and _PHYS is exactly the set of
non-fictitious object types, so just check for OBJ_FICTITIOUS. The
check no longer excludes dead objects, but such objects have to be
handled regardless.
No functional change intended.
Reviewed by: alc, dougm, kib
Tested by: pho
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26589
emaste [Fri, 2 Oct 2020 14:00:52 +0000 (14:00 +0000)]
libmd: add dependency workaround for r366344
r366344 fixed and reenabled the assembly optimized skein implementation,
but skein_block objects were not being rebuilt in no-clean builds. This
resulted in failing no-clean builds. SKEIN_USE_ASM controls which
routines come from C vs assembly, and with no explicit dependency
r366344's change to SKEIN_USE_ASM did not cause skein_block.{o,pico}
to be rebuilt.
Add a dependency on this Makefile for the skein_block objects. This
dependency is broader in scope than absolutely required (that is, the
skein_block objects will now be rebuilt on any change to this Makefile).
There are ways this could be addressed, but it is probably not worth the
additional effort or testing time to pursue them.
PR: 248221
Reported by: kevans, Jeremy Faulkner
Discussed with: kevans
Sponsored by: The FreeBSD Foundation
Access faults in user mode are treated like TLB misses, which leads to an
endless loop of faults. It's less serious than the same fault in kernel mode,
because we can just terminate the process, but that's not ideal.
cxgbe(4): validate largest_rx_cluster and safest_rx_cluster.
These tunables can only be set to a valid cluster size (2K, 4K, 9K, or
16K) as documented in the man page. Anything else could lead to a
panic on interface up.
mmacy [Thu, 1 Oct 2020 23:28:21 +0000 (23:28 +0000)]
OpenZFS: MFV 2.0-rc3-gfc5966
- Annotate FreeBSD sysctls with CTLFLAG_MPSAFE
- Reduce stack usage of Lua
- Don't save user FPU context in kernel threads
- Add support for procfs_list
- Code cleanup in zio_crypt
- Add DB_RF_NOPREFETCH to dbuf_read()s in dnode.c
- Drop references when skipping dmu_send due to EXDEV
- Eliminate gratuitous bzeroing in dbuf_stats_hash_table_data
- Fix legacy compat for platform IOCs
vangyzen [Thu, 1 Oct 2020 21:48:22 +0000 (21:48 +0000)]
zgrep: fix exit status with multiple files
zgrep should exit with success when given multiple files and the
pattern is found in at least one file. Prior to this change,
it would exit with success only if the pattern was found in _every_ file.
The assembly implementation incorrectly used logical AND instead of
bitwise AND. Fix, and re-enable in libmd.
Submitted by: Yang Zhong <yzhong@freebsdfoundation.org>
Reviewed by: cem (earlier)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26614
kevans [Thu, 1 Oct 2020 19:56:38 +0000 (19:56 +0000)]
auxv: partially revert r366207, cast buflen to unsigned int as needed
The warning generated pre-r366207 is actually a sign comparison warning:
error: comparison of integers of different signs: 'unsigned long' and 'int'
if (strlcpy(buf, execpath, buflen) >= buflen)
Revert parts that affected other lines and just cast this to unsigned int.
The buflen < 0 -> EINVAL has been kept despite no longer serving any
purposes w.r.t. sign-extension because I do believe it's the right thing to
do: "The provided buffer was not the right size for the requested item."
The original warning is confirmed to still be gone with an:
env WARNS=6 make WITHOUT_TESTS=yes.
jhb [Thu, 1 Oct 2020 16:45:11 +0000 (16:45 +0000)]
Clear the upper 32-bits of registers in x86_emulate_cpuid().
Per the Intel manuals, CPUID is supposed to unconditionally zero the
upper 32 bits of the involved (rax/rbx/rcx/rdx) registers.
Previously, the emulation would cast pointers to the 64-bit register
values down to `uint32_t`, which while properly manipulating the lower
bits, would leave any garbage in the upper bits uncleared. While no
existing guest OSes seem to stumble over this in practice, the bhyve
emulation should match x86 expectations.
This was discovered through alignment warnings emitted by gcc9, while
testing it against SmartOS/bhyve.
ngie [Thu, 1 Oct 2020 16:37:49 +0000 (16:37 +0000)]
Eliminate duplicate `afterinstallconfigs` target
Define separate dependent targets which `afterinstallconfigs` relies on, in
order to modify `${DESTDIR}/etc/master.passwd` and
`${DESTDIR}/etc/nsswitch.conf`.
Mark these targets .PHONY, since they manipulate configurations on the fly and
the generation logic isn't 100% defined in terms of the source files/logic,
and is variable, based on MK_foo flags.
netchild [Thu, 1 Oct 2020 08:57:36 +0000 (08:57 +0000)]
Remove nfsstat. Running nfsstat in crashinfo will give the stats of the
running kernel instead of the stats of the crashed kernel. The current
version uses sysctls to query the stats and does not work at all (anymore)
on crash dumps.
This version incorporates many fixes in particular a fix for vi -w
Another approach was proposed to merge those fixes (see review), I find
it easier to track changes if we keep importing snapshot on regular
basis
kevans [Thu, 1 Oct 2020 01:10:51 +0000 (01:10 +0000)]
Do a sweep and remove most WARNS=6 settings
Repeating the default WARNS here makes it slightly more difficult to
experiment with default WARNS changes, e.g. if we did something absolutely
bananas and introduced a WARNS=7 and wanted to try lifting the default to
that.
Drop most of them; there is one in the blake2 kernel module, but I suspect
it should be dropped -- the default WARNS in the rest of the build doesn't
currently apply to kernel modules, and I haven't put too much thought into
whether it makes sense to make it so.
rmacklem [Thu, 1 Oct 2020 00:47:35 +0000 (00:47 +0000)]
Modify the NFSv4.2 VOP_COPY_FILE_RANGE() client call to return after one
successful RPC.
Without this patch, the NFSv4.2 VOP_COPY_FILE_RANGE() client call would
loop until the copy "len" was completed. The problem with doing this is
that it might take a considerable time to complete for a large "len".
By returning after a single successful Copy RPC that copied some of the
data, the application that did the copy_file_range(2) syscall will be
more responsive to signal delivery for large "len" copies.
rmacklem [Thu, 1 Oct 2020 00:33:44 +0000 (00:33 +0000)]
Clip the "len" argument to vn_generic_copy_file_range() at a
hole size boundary.
By clipping the len argument of vn_generic_copy_file_range() to end at
an exact multiple of hole size, holes are more likely to be maintained
during the copy.
A hole can still straddle the boundary at the end of the
copy range, resulting in a block being allocated in the
output file as it is being grown in size, but this will reduce the
likelyhood of this happening.
While here, also modify setting of blksize to better handle the
case where _PC_MIN_HOLE_SIZE is returned as 1.
Avoid a dubious assignment to bio_data in aio_qbio().
A user pointer is not a suitable value for bio_data and the next block
of code always overwrites bio_data anyway. Just use cb->aio_buf
directly in the call to vm_fault_quick_hold_pages().
cem [Wed, 30 Sep 2020 14:55:54 +0000 (14:55 +0000)]
gdb(4): Don't escape GDB special characters at application layer
In r351368, we introduced this XML- and GDB-encoded data. The protocol
'offset' should reflex the logical XML data offset, but unfortunately we
counted the GDB escapes as well.
In fact, we cannot safely do GDB character escaping at this layer at
all, because we don't know what will be flushed in a packet. It is
bogus to send only the first character of a two-character escape
sequence.
This patch "corrects" the problem by squashing these characters in the
transmitted XML document. It would be nice to transmit the characters
faithfully, but that is a more complicated change. Thread names are a
nice convenience feature for the GDB client, but one can always inspect
td_name or p_comm directly to find the true name.
Reported by: Ka Ho Ng <khng300 AT gmail.com>
Tested by: Ka Ho Ng
Reviewed by: emaste, markj, rlibby
Differential Revision: https://reviews.freebsd.org/D26599
Load/store/fetch access exceptions always indicate a violation of a PMP
rule. We can't treat those as page faults, because updating the page
table and trying again will only result in exactly the same access
exception recurring. This leaves us in an endless exception loop.
We cannot recover from these exceptions, so panic instead.
Every other architecture defines this and this is required for
interrupts to work when using QEMU's PCI VirtIO devices (which all
report an interrupt line of 0) for two reasons.
Firstly, interrupt line 0 is wrong; they use one of 0x20-0x23 with the
lines being cycled across devices like normal. Moreover, RISC-V uses
INTRNG, whose IRQs are virtual as indices into its irq_map, so even if
we have the right interrupt line we still need to try and route the
interrupt in order to ultimately call into intr_map_irq and get back a
unique index into the map for the given line, otherwise we will use
whatever happens to be in irq_map[line] (which for QEMU where the line
is initialised to 0 results in using the first allocated interrupt,
namely the RTC on IRQ 11 at time of commit).
Note that pci_assign_interrupt will still do the wrong thing for INTRNG
when using a tunable, as it will bypass INTRNG entirely and use the
tunable's value as the index into irq_map, when it should instead
(indirectly) call intr_map_irq to allocate a new entry for the given
IRQ and treat the tunable as stating the physical line in use, which is
what one would expect. This, however, is a problem shared by all INTRNG
architectures, and not exclusive to RISC-V.
Make copy_file_range(2) Linux compatible for overflow of offset + len.
Without this patch, if a call to copy_file_range(2) specifies an input file
offset + len that would wrap around, EINVAL is returned.
I thought that was the Linux behaviour, but recent testing showed that
Linux accepts this case and does the copy_file_range() to EOF.
This patch changes the FreeBSD code to exhibit the same behaviour as
Linux for this case.
This appears to be a typo. The AdvSIMD field encodes support for
half-precision floating point SIMD instructions, which corresponds to
HWCAP_ASIMDHP, not HWCAP_ASIMDDP.
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Fallback to software for more GCM and CCM requests.
ccr(4) uses software to handle GCM and CCM requests not supported by
the crypto engine (e.g. with only AAD and no payload). This change
adds a fallback for a few more requests such as those with more SGL
entries than can fit in a work request (this can happen for GCM when
decrypting a TLS record split across 15 or more packets).
Rather than placing the epoch around the entire receive loop which
might call into rtwn_rx_frame() and USB and sleep, split the loop
into two[1] and leave us with one unlock/lock cycle as well.
PR: 249925
Reported by: thj, (rkoberman gmail.com)
Tested by: thj
Suggested by: adrian [1]
Reviewed by: adrian
MFC after: 3 days
Sponsored by: The FreeBSD Foundation (initially, paniced my iwl lab host)
Differential Revision: https://reviews.freebsd.org/D26554
Add EXAMPLES section to the man page showing the use of all flags except for
-S.
While here, clarify -f description. It not only suppresses diagnostic messages
but it also affects the exit status of the command itself. This is shown in two
of the examples.
OpenZFS will start using some of the kernel timekeeping bits
shortly. This implements the bare minimum of that which currently
is just the time_seconds variable.
o Rename acpi_iommu_get_dma_tag() -> iommu_get_dma_tag().
This function isn't ACPI dependent and we may use it on FDT systems
as well.
o Don't repeat the function declaration, include iommu.h instead.
LLVM 11 changed the meaning of '-O' from '-O2' to '-O1', which resulted
in debug kernels (with 'makeoptions DEBUG=-g') being built with inlining
disabled, causing severe performance hit.
The -O2 was already being used for building amd64, powerpc, and powerpcspe.
Improve the input validation and processing of cookies.
This avoids setting the association in an inconsistent
state, which could result in a use-after-free situation.
This can be triggered by a malicious peer, if the peer
can modify the cookie without the local endpoint recognizing
it.
Thanks to Ned Williamson for reporting the issue.
cxgbe(4): Avoid unnecessary work in the firmware during netmap tx.
Bind the netmap tx queues to a special '0xff' scheduling class which
makes the firmware skip some processing related to rate limiting on the
outgoing traffic. Future firmwares will do this automatically.
cxgbe(4): fixes for netmap operation with only some queues active.
- Only active netmap receive queues should be in the RSS lookup table.
- The RSS table should be restored for NIC operation when the last
active netmap queue is switched off, not the first one.
- Support repeated netmap ON/OFF on a subset of the queues. This works
whether the the queues being enabled and disabled are the only ones
active or not. Some kring indexes have to be reset in the driver for
the second case.
projects / FreeBSD / FreeBSD.git / log