summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2016-12-06Make sure iwm(4) only calls ieee80211_mira_choose() while in RUN state.Stefan Sperling
Should fix "panic: timeout_add: not initialized" reported by danj@
2016-12-06The kvop is not mandatory to query the type, fixes hostctl -t on vmm(4) VMs.Reyk Floeter
OK mikeb@
2016-12-06avoid a potentially uninitialised return valueJonathan Gray
ok stsp@
2016-12-06Use the sizeof udphdr not the sizeof a pointer to it in the af-toJonathan Gray
specific part of pf_test_state_icmp(). This worked by accident on LP64 archs as the struct is eight bytes long. ok mikeb@ bluhm@ krw@ jca@
2016-12-05Remove recursive splsoftnet()s in ioctl(2) path.Martin Pieuchot
2016-12-05Harmonize battery state thresholds accross platforms.Frederic Cambus
Use the same values hardcoded in sys/dev/acpi/acpi.c where possible. OK visa@
2016-12-05Propagate error type in validation functions that callRafael Zalamena
swofp_validate_action(), because actions have a different class of errors. While there update the error type and error variables type to match the swofp_send_error() prototype.
2016-12-05Remove the flow id from the flow entry struct and the global variable,Rafael Zalamena
since they were not being used and the OpenFlow specification doesn't mention anything about them. ok reyk@
2016-12-05Add missing prototypes for all functions declared in the file.Rafael Zalamena
ok reyk@
2016-12-05Remove unused variable. From Michael W. Bombardieri <mb@ii.net>. Thanks!Alexandre Ratchov
2016-12-04Add handling for the second interrupt bank.Visa Hankala
Needed by an upcoming driver.
2016-12-04Since net/route.c r1.337 interface priority factors into route priorityStefan Sperling
when RTF_CONNECTED routes are added to the routing table. Specify a route priority calculated in the same way when deleting such routes. Makes route add and delete code paths consistent again. ok mpi@
2016-12-03Fix memory leak.Visa Hankala
2016-12-03Implement octeon_intr_disestablish().Visa Hankala
2016-12-03luna88k bootloader needs __ashldi3 to compile sys/lib/libsa/ufs.c 1.26.Kenji Aoyama
2016-12-02Clean up leftovers from r1.442.Vincent Gross
Local var 'up' is never set in ifioctl(). OK mpi@
2016-12-02Limit the amount of flows that can be installed on flow tables.Rafael Zalamena
2016-12-02Fix some spelling errors in the OpenFlow header and update the tcpdump(8)Rafael Zalamena
ofp_map.c file. ok reyk@
2016-12-02Use the right variable for storing the maximum group table size and useRafael Zalamena
it to limit the amount of installable groups.
2016-12-02regenMike Belopuhov
2016-12-028086:154a is not a QSFP model, possibly just a quad portMike Belopuhov
Cleanup trailing whitespaces from the previous change while here.
2016-12-02Don't overwrite the selected flow control settingsMike Belopuhov
procter@ has noticed that flow control settings survive module change when they should be forgotten. It became clear that we're overwriting the selected FC mode with the requested version instead of keeping negotiated settings. Tested by procter@.
2016-12-02Sinc rings are created during attach memory allocations shouldn't sleepMike Belopuhov
2016-12-02Disable the TX laser when interface is going down for all fiber modulesMike Belopuhov
Previously only multi-rate fiber modules would disable the TX laser, but newer Intel driver does it for single rate modules as well. Reminded by kettenis@, tested by procter@ and Hrvoje Popovski. Thanks!
2016-12-02Drop stale prototypes and an outdated comment.Visa Hankala
2016-12-02Simplify interrupt dispatching by processing requests withoutVisa Hankala
prioritization by IPL.
2016-12-02Assert that rt_match() is always called at IPL_SOFTNET rather thanMartin Pieuchot
calling splsoftnet() recursively.
2016-12-02Rremoves 'struct route_in6 *' argument from in6_selectsrc().Martin Pieuchot
Move the corresponding code in in6_pcbselsrc(). This reduces differences with IPv4 and will help us to get rid of 'struct route*'. ok vgross@
2016-12-02enable hvn(4)Jonathan Gray
ok mikeb@
2016-12-02Fix flow-removed OFP header xid value: use htonl() instead of htons() forRafael Zalamena
32bit integers.
2016-12-02regenDavid Gwynne
2016-12-02have a better go at naming xl710 devicesDavid Gwynne
intel use xl710 to refer to 40 gig parts (and 20 for some reason), and x710 to refer to 10g parts. there's allegedly going to be 25g parts called xxv710. i havent included the xl722 parts yet. im naming the devices by the speed and connector rather than going for actual intel product names like XL710-QDA1 because other vendors will use the same chips in product with other names. intel also put the XL710-QDA1 product id on the XL710-QDA2, which is a good argument not to use product names too.
2016-12-01Update comments in struct mbuf_ext to reflect the replacement of ext_free byLawrence Teo
ext_free_fn. ok tedu@
2016-12-01Add noop drm SET_MASTER and DROP_MASTER ioctls. This will allowJonathan Gray
reducing the local diff in libdrm. We only handle a single master as xorg privdrop has already occurred at the point where the ioctls are issued.
2016-11-30Fix another free() with wrong size panic when handling group-mod bucketsRafael Zalamena
size changes and add more sanity checks for group buckets payload.
2016-11-30Update media types upon SFP module changeMike Belopuhov
Tested by Hrvoje Popovski and myself.
2016-11-30Switch the iwm(4) driver to mira rate adaptation in 11n mode.Stefan Sperling
Only the rate adaptation algorithm changes, available data rates do not (yet). Please let me know about any regressions. In 11a/b/g modes the driver still uses AMRR, so forcing one of these modes with ifconfig's 'mode' subcommand will serve as a fallback if necessary. ok tb@ mpi@ kettenis@
2016-11-30Enable mira in kernel builds.Stefan Sperling
For now, only build mira if the iwm(4) or iwn(4) drivers are also built since other wifi drivers don't even have 11n support yet. This limits platforms affected by this change to i386 and amd64. make release on i386/amd64 platforms tested by tb@ (thanks!) ok tb@ mpi@ kettenis@
2016-11-30Add a new implementation of MiRA, a rate scaling algorithm for 802.11n.Stefan Sperling
This algorithm was designed for use with MIMO and Tx aggregation. This is joint work with tb@, who helped with all the tricky math bits. Additional help with testing by phessler@, mpi@, and jmatthew@. I believe this is now ready for wider testing, and for future work to happen in-tree. A paper which explains the algorithm can be found at: http://www.cs.ucla.edu/wing/publication/papers/Pefkianakis.MOBICOM10.pdf Roughly, this algorithm attempts to keep track of the current "goodput" (the effective data rate) for each MCS. It converges towards a rate which gets the most bits per second transmitted with least loss. Occasionally, frames will be steered to different rates to probe for changes. (The algorithm does not send frames on its own. It only advances whenever the driver has sent a frame.) Time-based probing to adjacent MCS rates occurs periodically. This is similar to what AMRR does, except that eventually mira will try out multi-antenna modes as well. Event-based probing happens when a sudden change in goodput is detected. I've chosen to make downwards probing fast, and upwards probing slow. (The paper does not specify such a preference.) This means it should react quickly to worsening conditions and pull the rate down (perhaps to the lowest possible rate). It should then raise upwards slowly on a rate-per-rate basis as conditions improve again. In my testing this works as intended as I keep moving a laptop outside and inside the AP's range. Not linked to the build yet. ok mpi@ kettenis@
2016-11-30Make iwn(4) receive MIMO frames in monitor mode. We can now sniff all 802.11nStefan Sperling
frames the hardware is able to receive. Use an xT3R device for best results. This change has not yet been tested on 1T1R devices due to lack of hardware. ok kettenis@
2016-11-30Do not overlay DMA fragment decriptors with free list handling.Martin Pieuchot
This "cleverness" increase the risk of races due to caching and/or prefetching between the HC and DMA engine. Many of the bug reports on bugs@ involving memory corruptions in usb_allocmem() should be easier to diagnose when not avoided with this change. From Marius Strobl, ok kettenis@
2016-11-29Kill unused 'struct route'.Martin Pieuchot
2016-11-29Stop exposing xen_softc to PV devices directlyMike Belopuhov
2016-11-29Don't expose the xen_softc pointer in the XenStore transaction structMike Belopuhov
2016-11-29Replace the hand-rolled semaphore with a read-write lockMike Belopuhov
This was sitting in my tree for many a month and since the introduction of interrupt threads, the interrupt vs. process context interlock became irrelevant. Taking uncontended write locks while "cold" doesn't look like a big deal as well.
2016-11-29m_free() and m_freem() test for NULL. Simplify callers which had their ownJonathan Gray
NULL tests. ok mpi@
2016-11-29Remove debugging code that was always enabled and printed parsedReyk Floeter
OpenFlow packets: the new DLT_OPENFLOW bpf methods allows to monitor the communication with a switch(4) device with tcpdump now. The remaining debugging messages aren't compiled without DEBUG now. On amd64, this shrinks the switchofp.o object by about 10k by default and about 8k with DEBUG enabled. OK rzalamena@
2016-11-29For virtual Ethernet drivers that don't have a technical limit of theReyk Floeter
hardmtu, pick a value of 65435 that leaves space for some encapsulation and almost a complete max-IP packet. After some discussion we picked this arbitrary value. OK dlg@
2016-11-29Build kernel with DEBUG=-g by default.Martin Pieuchot
This will allow us to extract type informations from DWARF2 sections. It also makes developer life easier as debug information are now included in every object. Resulting kernels will be stripped using strip(1) instead of ld(1). Kernel build time increases by approximately 10%. However it is still possible to disable this by defining DEBUG="". ok kettenis@, bluhm@, natano@, jasper@, reyk@, deraadt@
2016-11-28Path MTU discovery and traceroute did not always work with pf af-to.Alexander Bluhm
If an incoming packet is directly put into the output path, sending the icmp error packet is never done. As this is basically forwarding, calling ip_forward() for such packets does everything that is needed. OK mikeb@
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-10 09:37:34 +0000