ni_txseqs is kept as 16-bit counter, but we need to trim the upper four
bits as they may have special meanings for the firmware / hardware.
For instance, bit 15 enables hardware / firmware generation of sequence
numbers that overrides sequence numbers programmed by the driver.
Reviewed by: adrian
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30814
The i386 loader (and hopefully others to come) now passes tslog data
as a "preloaded module". Include this in the data returned by the
debug.tslog sysctl.
Reviewed by: kevans
When the elapsed time of the operation is complete and stored in the QOS
field, set the CAM_QOS_VALID bit. In iosched, test to make sure it's
set before using it.
Sponsored by: Netflix
As pointed out in the bug, the genet driver (RPi4 Ethernet) was
attaching the interrupts before the data structures were fully
initialized, causing a crash if an interrupt came in during the
attach. Fix by reordering code blocks.
PR: 256334
Reported by: < ghuckriede at blackberry.com >
Reviewed by: < ghuckriede at blackberry.com > (informally)
MFC after: 3 days
Every FUSE operation has a unique value in its header. As the name
implies, these values are supposed to be unique among all outstanding
operations. And since FUSE_INTERRUPT is asynchronous and racy, it is
desirable that the unique values be unique among all operations that are
"close in time".
Ensure that they are actually unique by incrementing them whenever we
reuse a fuse_dispatcher object, for example during fsync, write, and
listextattr.
PR: 244686
MFC after: 2 weeks
Reviewed by: pfg
Differential Revision: https://reviews.freebsd.org/D30810
Current data structure is using a hash of unordered lists. Those
unordered lists are quite efficient, because the least recently
inserted entries are most likely to be used again. In order to avoid
long search times in other cases, the lists are hashed into many
buckets. Unfortunatly a search for a miss needs an exhaustive
inspection and a careful definition of the hash.
Splay trees offer a similar feature: Almost O(1) for access of the
least recently used entries, and amortized O(ln(n)) for almost all
other cases. Get rid of the hash.
Now the data structure should able to quickly react to external
packets without eating CPU cycles for breakfast, preventing a DoS.
PR: 192888
Discussed with: Dimitry Luhtionov
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30536
Current data structure is using a hash of unordered lists. Those
unordered lists are quite efficient, because the least recently
inserted entries are most likely to be used again. In order to avoid
long search times in other cases, the lists are hashed into many
buckets. Unfortunatly a search for a miss needs an exhaustive
inspection and a careful definition of the hash.
Splay trees offer a similar feature - almost O(1) for access of the
least recently used entries), and amortized O(ln(n) - for almost all
other cases. Get rid of the hash.
Discussed with: Dimitry Luhtionov
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30516
The entry deleteAllLinks in the struct libalias is only used to signal
a state between internal calls. It's not used between API calls.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30604
Get rid of PORT_BASE, replace by AliasRange. Simplify code.
Factor out the search for a new port. Improves the perfomance a bit.
Discussed with: Dimitry Luhtionov
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30581
Let PPTP use its own data structure.
Regroup and rename other lists, which are not PPTP.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30580
Reorder incoming links by grouping of common search terms.
Significant performance improvement for incoming (missing) flows.
Remove LSNAT from outgoing search.
Slight speedup due to less comparsions in the loop.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30574
Factor out the outgoing search function.
Preparation for a new data structure.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30572
Separate the partially specified links into a separate data structure.
This would causes a major parformance impact, if there are many of
them. Use a (smaller) hash table to speed up the partially link
access.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30570
This completes PRR cwnd reduction in all circumstances
for the base TCP stack (SACK loss recovery, ECN window reduction,
non-SACK loss recovery), preventing the arriving ACKs to
clock out new data at the old, too high rate. This
reduces the chance to induce additional losses while
recovering from loss (during congested network conditions).
For non-SACK loss recovery, each ACK is assumed to have
one MSS delivered. In order to prevent ACK-split attacks,
only one window worth of ACKs is considered to actually
have delivered new data.
MFC after: 6 weeks
Reviewed By: rrs, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D29441
Search fully specified links first. Some performance loss due to need
to revisit the db twice, if not found.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30569
Factor out the common Out and In filter
Slightly better performance due to eager skip of search loop
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30568
Summary:
- Use LibAliasTime as a real global variable for central timekeeping.
- Reduce number of syscalls in user space considerably.
- Dynamically adjust the packet counters to match the second resolution.
- Only check the first few packets after a time increase for expiry.
Discussed with: hselasky
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30566
Stats counters are used as unsigned valued (i.e. printf("%u")) but are
defined as signed int. This causes trouble later, so fix it early.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D30587
jhb@ pointed out an extra plural in this phrase and a gramatical error,
so reword a little to be less awkward to fix both issues.
Sponsored by: Netflix
The original %b description string is wrong.
Sponsored by: The FreeBSD Foundation
Reviewed by: imp, jhb
Differential Revision: https://reviews.freebsd.org/D30805
Fix a last minute change from d4a4960c65
based on review feedback in where a function now gets called before
it is declared which did not fully get merged back to my commit branch.
Noticed by: CI, jkim
MFC after: 10 days
X-MFC with: d4a4960c65
Sponsored-by: The FreeBSD Foundation
Some code manually calls local_bh_disable() and spin_lock() but
then calls spin_unlock_bh() (or vice versa).
Our code then calls local_bh_disable() again from spin_lock()
which means we have the thread pin count increased twice and that
means we get out of synch and are still pinned when returning to
user space.
Avoid this by adding the explicit local_bh_{enable,disable}() to
the spin_[un]lock_bh() versions.
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30711
In sg_pcopy_from_buffer() is an error in that skip can underflow
and lead to bogus page arithmetics which may lead to memory corruption
or more likely panics. Once we found a s/g page to copy into there
is nothing to skip anymore so simply set skip to 0.
Sponsored by: The FreeBSD Foundation
MFC after: 5 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30676
Restructure some code and add support for various "managed" versions
for PCI resource management.
This is beyond of what iwlwifi needs but some was found with other
wireless drivers and it mostly all goes together.
Add one FreeBSD sepcific feature returning the resource rather than
the handle to allow us to use bus_*() functions in drivers directly.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30558
Given we are manually setting up the "device" in PCI in some cases,
we need to initialise the list and lock for device devres here as well
as otherwise we will panic on the uninitialised lock.
Sponsored by: The FreeBSD Foundation
MFC after: 5 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30681
Move request_irq() to an internal function which serves request_irq()
and the newly added request_threaded_irq() and devm_request_threaded_irq().
Likewise factor out parts of free_irq() to also be used with
devm_free_irq(). Add the storage and call to a thread_handler in case
of IRQ_WAKE_THREAD.
This is needed for the iwlwifi driver.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30549
Add dummy functions for dealing with "HotPlug" events which we currently
do not support.
Add pci_dev_get(), pci_find_ext_capability() and pci_pme_capable().
The added pcie_find_root_port() is a bit special as we need to create
another linux pci device; for that make lkpinew_pci_dev() public
which is also helpful for other cases when we want to use the Linux
routines to check for device identifiers only and need a container
for the "bsddev" to use natively. This has proven to avoid basic
checking code for the sake of rewriting it to native field names
elsewhere. Given we cache the newly created "root" we also need to
make sure we clean it up.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30521
dmam_pool_create() is a "managed" version of dma_pool_create() which
will cleanup everything left when the device goes away using the
devres framework. For that add an internal cleanup function to be
called from devres release.
This is used by at least one wireless driver.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30520
Add two new (though untested) functions to linux/device.h which are
dealing with manually managing the device/driver and are used by
at least one wireless driver. We may have to re-fine them in the
future.
Move the devres declarations further up so they can be used earlier
in the file.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D30519
Given all the code does operate on struct ifnet, the last step in this
longer series of changes now is to rename struct net_device to
struct ifnet (that is what it was defined to in the LinuxKPi code).
While mlx4 and OFED are "shared" code the decision was made years ago
to not write it based on the netdevice KPI but the native ifnet KPI
for most of it. This commit simply spells this out and with that
frees "struct netdevice" to be re-done on LinuxKPI to become a more
native/mixed implementation over time as needed by, e.g., wireless
drivers.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30515
While currently the ifp gets cast to a net_device and then returned
and consumers are expecting an ifp again, allow parallel usage now and
in the future by extending and also passing the ifp directly back in
the netdev_notifier_info. Add a function to return the ifp instead of
the net_device.
Sponsored by: The FreeBSD Foundation
MFC after: 10 days
Suggested by: hselasky
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D30522
so that PHYS_TO_VM_PAGE() and consequently physcopyin() work for them
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D30785
Fix a bug in the LPM SEQ benchmark (missing break inside a switch block)
by restructuring the test loop, while introducing additional two
synthetic test options:
ANN: scan only the address space announced in current RIB
REP: repeat lookups over several keys in a sliding window scheme
The total of eight combinations of test options are now available
through dedicated sysctl hooks.
Differential Revision: <https://reviews.freebsd.org/D30311>
Reviewed by: melifaro
MFC after: 3 days
armv6 and armv7 systems already were 1000Hz. The other armv5 were a
mix of 100 and 1000. This changes them to 1000. Should there be
issues, we can add options HZ=100 to the systems that have bad
performance at the drop of a hat.
mips is a lot more complicated. But most of the systems are already
1000HZ. The hardware exceptions are all fast enough to run at
1000Hz. MALTA is our primary emulator, and history has shown emulators
tend to like 100Hz better, so run those systems at 100Hz. As with arm,
any system that shows a huge performance regression can reverted to
100Hz easily.
This was going to be committed well in advance of the 13 branch, but
it was delayed and forgotten til now.
Discussed on: #bsdmips ages ago
Sponsored by: Netflix
