- Send IPI wakeups once SMP is started even if cold is true.
- Permit preemptions when cold is true.
These changes are needed for EARLY_AP_STARTUP.
MFC after: 2 weeks
Sponsored by: Netflix
The other CPU might resume and see a still-empty runq and go back to
sleep before sched_add() adds the thread to the runq. This results
in a lost wakeup and a potential hang if the system is otherwise
completely idle.
The race originated due to a micro-optimization (my fault) in 4BSD in
that it avoided putting a thread on the run queue if the scheduler was
going to preempt to the new thread. To avoid complexity while fixing
this race, just drop this optimization. 4BSD now always sets the
"owepreempt" flag when a preemption is warranted and defers the actual
preemption to the thread_unlock of the caller the same as ULE.
MFC after: 2 weeks
Sponsored by: Netflix
As in the gnu/lib/libgcc Makefile:
libgcc is linked in last and thus cannot depend on ssp
symbols coming from earlier libraries. Disable stack protection
for this library.
Reviewed by: dim
Sponsored by: The FreeBSD Foundation
In a git worktree, the gitdir is in an entirely different location.
In arcgit, use git rev-parse --git-dir to get the correct path to it
always.
When running git from outside of the work tree, as in importgit,
the path provided by git rev-parse --git-dir can be either a
relative or absolute path depending on the work tree. Rather
than trying to deal with that, just use git -C.
Differential Revision: https://reviews.freebsd.org/D8501
Reviewed by: markj
AArch64: only try to use scaled fcvt ops on legal vector types.
Before we ended up calling getSimpleVectorType on a <3 x float>, which
asserted.
This fixes an assertion when building the print/ghostscript9-agpl-base
port for AArch64.
PR: 213865
MFC after: 3 days
In the case where a hardware error is detected during
ioat_process_events, hardware may advance (by one descriptor, probably)
and a subsequent ioat_process_events may race the intended ioat_reset_hw
followup. In that case, the second process_events would observe a
completion update that does not match the software "last_seen" status,
and attempt to successfully complete already-failed descriptors.
Guard against this race with the resetting_cleanup flag.
Reviewed by: bdrewery, markj
Sponsored by: Dell EMC Isilon
Newer CPUs (SkyLakes) have updates of 100K size, which is bigger than
current limit 32K. Increase it to 4M but leave the check around to
prevent kernel memory allocator abuse. Some time ago, the memory for
update was allocated by contigmalloc(9), and it was reasonable to be
conservative as much as possible. Since all uses of contigmalloc(9)
appear to be either misunderstanding or too cautious, and were
removed, provide more slack than strictly neccessary.
Submitted by: Oliver Pinter
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D8486
/loader.rc) to specify a Forth file to read from the pkgfs tarball and
process by Ficl.
This allows for the tarball to do runtime things like load a
platform-specific FDT blob, among other things.
Reviewed by: imp
Approved by: sjg (mentor)
MFC after: 2 weeks
Sponsored by: Juniper Networks, Inc.
Differential Revision: https://reviews.freebsd.org/D8494
After removal of SMB_TRANS some information in the description of
SMB_BWRITE has become stale. E.g., the maximum block size has been
restored to 32.
Also, the descriptions of SMB_BREAD and SMB_BWRITE had some
incorrect information on the SMBus protocol details.
MFC after: 1 week
X-MFC with: r308242
Differential Revision: https://reviews.freebsd.org/D8431
The constant was set to the correct value in r308242.
While there, fix iicsmb_bread() to not use a value of an out parameter
'count'.
MFC after: 3 weeks
X-MFC after: r308242
The hardware does not implement SMBus Process Call command, so remove
ifdef-ed out code from intsmb_pcall. The code used exactly the same
start sequence as for Write Word command.
intsmb_bread code used to access an in value of the count parameter,
but that parameter is supposed to be an out only parameter.
For example, smb(4) does not initialize it before calling smbus_bread.
MFC after: 3 weeks
Developers writing code for accessing /dev/smb may use this base utility
as an example. Now that SMB_READB, SMB_READW, SMB_PCALL behave as
documented, wwe can use them in a more convenient way than before.
MFC after: 4 weeks
X-MFC after: r308527
Previously, those ioctls were defined as 'in' only, so rdata.byte and
rdata.word were never updated in the userland. The read data went only
to rbuf if it was provided. Thus, consumers were forced to always use it.
Now the ioctls are marked as in-out.
Compatibility handlers are provided for old ioctls.
PR: 213481
Reported by: Lewis Donzis <lew@perftech.com>
MFC after: 2 weeks
Relnotes: maybe
Differential Revision: https://reviews.freebsd.org/D8430
Used extensively on my network over the past month.
Reviewed by: pfg, brooks
Suggested by: pfg
Obtained from: ftp://ftp.am-utils.org/pub/am-utils/
MFC after: 6 weeks
Relnotes: yes
Differential Revision: D8405
[AArch64] Don't blindly lower f16/f128 FCCMPs.
Instead, extend f16 (like we do when lowering a standalone SETCC),
and let f128 be legalized to the RT calls.
Fixes PR26803.
This fixes a fatal "Cannot select" backend error when building the
net/freerdp port for AArch64.
PR: 214380
MFC after: 3 days
Unfortunately (sigh) some firmware doesn't provide the RX BA starting point,
so we need to cope and set a "close enough" sequence number so we (hopefully!)
don't discard frames as duplicates.
Tested:
* QCA9880v2, athp driver (under development), STA mode
I see the fllowing panic on AMD when exiting pmcstat:
panic: [pmc,1473] pp_pmcval outside of expected range cpu=2 ri=17
pp_pmcval=fffffffffa529f5b pm_reloadcount=10000
It seems that at least on AMD a performance counter keeps counting after
overflowing. When pmcstat exits it sets counters that it used to
PMC_STATE_DELETED and waits until their use count goes to zero.
amd_intr() wouldn't reload a counter in that state and, thus, a counter
would be allowed to overflow. That means that the counter's value would
be allowed to go outside the expected range.
MFC after: 2 weeks
The code assumed that 'timeout' and 'timeout_sec' are in sync
which they weren't if no '-t' option was passed to watchdogd.
Reported by: Olivier Smedts <olivier@gid0.org>,
Alex Deiter <alex.deiter@gmail.com>
Tested by: Olivier Smedts <olivier@gid0.org>,
Alex Deiter <alex.deiter@gmail.com>
MFC after: 5 days
X-MFC with: r308040
To avoid have warning for services that are using oomprotect, oomprotect
will only be applied on services that won't run inside jails.
Reported by: allanjude
MFC after: 2 weeks.
Always define boot.netif.server in kenv in pxeboot
Add "boot.tftproot.server" to kenv when pxeboot uses tftpfs
Change the code order when setting env for TFTP or NFS to be the same as
common/dev_net.c
Reported by: tsoome
pages, specificially, dirty pages that have passed once through the inactive
queue. A new, dedicated thread is responsible for both deciding when to
launder pages and actually laundering them. The new policy uses the
relative sizes of the inactive and laundry queues to determine whether to
launder pages at a given point in time. In general, this leads to more
intelligent swapping behavior, since the laundry thread will avoid pageouts
when the marginal benefit of doing so is low. Previously, without a
dedicated queue for dirty pages, the page daemon didn't have the information
to determine whether pageout provides any benefit to the system. Thus, the
previous policy often resulted in small but steadily increasing amounts of
swap usage when the system is under memory pressure, even when the inactive
queue consisted mostly of clean pages. This change addresses that issue,
and also paves the way for some future virtual memory system improvements by
removing the last source of object-cached clean pages, i.e., PG_CACHE pages.
The new laundry thread sleeps while waiting for a request from the page
daemon thread(s). A request is raised by setting the variable
vm_laundry_request and waking the laundry thread. We request launderings
for two reasons: to try and balance the inactive and laundry queue sizes
("background laundering"), and to quickly make up for a shortage of free
pages and clean inactive pages ("shortfall laundering"). When background
laundering is requested, the laundry thread computes the number of page
daemon wakeups that have taken place since the last laundering. If this
number is large enough relative to the ratio of the laundry and (global)
inactive queue sizes, we will launder vm_background_launder_target pages at
vm_background_launder_rate KB/s. Otherwise, the laundry thread goes back
to sleep without doing any work. When scanning the laundry queue during
background laundering, reactivated pages are counted towards the laundry
thread's target.
In contrast, shortfall laundering is requested when an inactive queue scan
fails to meet its target. In this case, the laundry thread attempts to
launder enough pages to meet v_free_target within 0.5s, which is the
inactive queue scan period.
A laundry request can be latched while another is currently being
serviced. In particular, a shortfall request will immediately preempt a
background laundering.
This change also redefines the meaning of vm_cnt.v_reactivated and removes
the functions vm_page_cache() and vm_page_try_to_cache(). The new meaning
of vm_cnt.v_reactivated now better reflects its name. It represents the
number of inactive or laundry pages that are returned to the active queue
on account of a reference.
In collaboration with: markj
Reviewed by: kib
Tested by: pho
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D8302
Some tools spawn c++filt and pass it a single line at a time for
demangling. This is akin to r276689 for addr2line.
Sponsored by: The FreeBSD Foundation
print. Also add a new flag -s that add blocks size to statistics.
PR: 198347, 212726
Submitted by: Ben RUBSON <ben.rubson@gmail.com>
Tested by: pi
MFC After: 2 weeks.
