summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2016-11-28Remove NULL checks before m_free{m,}().Martin Pieuchot
ok reyk@, rzalamena@
2016-11-28Remove simple recursive splsoftnet() calls inside ifioctl().Martin Pieuchot
2016-11-28Assert that every slow/fast timeout routine is called at IPL_SOFTNET.Martin Pieuchot
This removes multipe recursive splsoftnet()/splx() dances.
2016-11-28Allow to build kernels without IPSEC but with PIPEX.Martin Pieuchot
2016-11-28Make sure the descriptor has been removed from the interface listMartin Pieuchot
before we call ifpromisc() and possibly sleep. ok bluhm@
2016-11-28Kill a micro optimization that no longer make sense since the two routingMartin Pieuchot
blocks have been merged in r1.292. ok claudio@
2016-11-28Rename "flowmax" to "maxflow" and give each switch(4) ioctl aReyk Floeter
dedicated number. Both changes for consistency. OK rzalamena@
2016-11-28Merge two "#ifdef MROUTING" blocks.Martin Pieuchot
It's one more step towards splitting ip6_input() in two and it reduces differences with v4. ok bluhm@
2016-11-27Add missing OpenBSD CVS tagsReyk Floeter
2016-11-27Fix signedness warnings with careful casts and replace a re-defined variable.Reyk Floeter
(Missed this chunk as part of previous commit to libsa/ufs.c) OK krw@
2016-11-27Match on "allwinner,sun7i-a20-mmc". In linux 4.9 the device trees forJonathan Gray
a31/a20/a23/a33/h3 mmc devices set this compatible string to denote the presence of sample clocks and no longer set "allwinner,sun5i-a13-mmc". ok kettenis@
2016-11-27ashldi3 is now neededTheo de Raadt
2016-11-26Allow rootdev to be specified without /dev/ prefix. The full path isn'tMartijn van Duren
needed down the road anyway. Both forms are supported. e.g. rootdev=/dev/sd0 and rootdev=sd0 OK visa@
2016-11-26Start process_bootargs at 0 instead of 1.Martijn van Duren
When using endbootargs in U-Boot arguments start at position 0. OK visa@
2016-11-26Remove unnecessary function cn30xxgmx_submatch and call config_foundMartijn van Duren
instead of config_found_sm. OK visa@
2016-11-25Add a new DRM_IOCTL_GET_PCIINFO ioctl to have a non driver specific wayJonathan Gray
to get pci device ids from a drm fd as an ordinary user. This will be used to provide non-sysfs implementations of functions in libdrm. These functions are required for Mesa >= 13. The ioctl number used is the same as the DRM_IOCTL_ADD_MAP ioctl which was removed when we removed DRI1 support. This was chosen to not collide with new ioctls from linux. ok kettenis@ ok mpi@ on an earlier version
2016-11-25syncMike Belopuhov
2016-11-25Fixup product names for Intel X550 family of controllersMike Belopuhov
X550 moniker is used only for 10GbaseT chips, while what's known internally as X550EM_X seems to be marketed under the X557 name. The X552 series is an upcoming X550EM_A chip. However the Xeon-D embedded model is called X552/X557.
2016-11-25Fix signedness warnings with careful casts and replace a re-defined variable.Reyk Floeter
OK krw@
2016-11-25The NCM encoding allows to aggregate multiple segments in one singleGerhard Roth
transfer. Using wMaxSegmentSize for bulk-in could break the rx-path because xfer length can be smaller than the messages sent by the device. And using some constant value for rx/tx size is also a bad idea, because we might be sending messages to the device that are too big for it to handle them. Therefore use the NCM GET_NTB_PARAMETERS request to query the device's configuration. Found, tested, and ok'ed by otto@
2016-11-24Add MID for arm64.Patrick Wildt
ok deraadt@
2016-11-24Enable support for the X550 family of 10 Gigabit controllersMike Belopuhov
Code was obtained from FreeBSD. Make release testing by tb@ on i386 and mikeb@ on amd64 and sparc64. X552 SFP tested by Hrvoje Popovski, HUGE thanks! X550T tested by mikeb@ on amd64 and sparc64. ok kettenis, deraadt
2016-11-24Fixup active media reporting for multi-speed fiber modulesMike Belopuhov
A gigabit fiber connection was mistakenly reported as 1000baseT when a mutli-speed 10GbaseSR/1000baseSX fiber optics module was set to the gigabit mode. Reported by and fix tested by Hrvoje Popovski <hrvoje at srce ! hr>, HUGE thanks!
2016-11-23Some socket splicing tests on loopback hang with large mbufs andAlexander Bluhm
reduced buffer size. If the send buffer size is less than the size of a single mbuf, it will never fit. So if the send buffer is empty, split the large mbuf and move only a part. OK claudio@
2016-11-23Move as much code out of a startup hook as possibleMike Belopuhov
To be able to perform asynchronous operations early we need to have a robust polling mechanism with flexible condition check semantics. A new hv_wait interface provides such facilities. Right now hvn(4) submits NVS commands and sleeps waiting for a interrupt to fire and run the completion code and it will take some effort to convert this code to be able to poll instead of tsleep. But VMBus attachment, channel scanning, etc can done at an earlier stage. Lets see if this sticks.
2016-11-23Explicitly forbid to combine af-to with route-to in pfctl. TheAlexander Bluhm
parser cannot handle that correctly and is is unclear wether the kernel code would work. Remove the feature until someone needs it and properly implements and tests it. OK mike@ sashan@ mpi@
2016-11-23M_*WAIT* flags cannot be treated as boolean valuesMike Belopuhov
2016-11-23Keep checks for local delivery close to in_ouraddr().Martin Pieuchot
ok claudio@
2016-11-23Disable debugging outputMike Belopuhov
2016-11-22Fix panic on detach hook when interfaces get destroyed.Rafael Zalamena
ok reyk@
2016-11-22Fold union pf_headers buffer into struct pf_pdesc (enabled by pfvar_priv.h).Richard Procter
Prevent pf_socket_lookup() reading uninitialised header buffers on fragments. OK blum@ sashan@
2016-11-22Fix CPU info struct.Visa Hankala
2016-11-22Use CPUID flags to determine working componentsMike Belopuhov
This adds a few checks to make sure we're not trying to use features that are not advertised with CPUID feature flags and avoid attaching when Xen viridium emulation is turned on. Prompted by a report from Kirill Miazine <km at krot ! org>, thanks! Discussed with reyk@.
2016-11-22Fix OFP spelling for version bitmap define in ofp header and tcpdump, alsoRafael Zalamena
change the truncation message in hello with the standard one. ok reyk@
2016-11-22Enforce that ifioctl() is called at IPL_SOFTNET.Martin Pieuchot
This will allow us to keep locking simple as soon as we trade splsoftnet() for a rwlock. ok bluhm@
2016-11-22Add support for the sun9i USB PHYs and attach to the sun9i EHCI controllers.Mark Kettenis
Doesn't work yet, but it is a step in the right direction.
2016-11-22Enforce that pr_ctlinput, pr_slowtimo and pr_fasttimo are calledMartin Pieuchot
at IPL_SOFTNET. This will allow us to keep locking simple as soon as we trade splsoftnet() for a rwlock. ok bluhm@
2016-11-22Enforce that pr_ctloutput is called at IPL_SOFTNET.Martin Pieuchot
This will allow us to keep locking simple as soon as we trade splsoftnet() for a rwlock. ok bluhm@
2016-11-21Add support for "allwinner,sun9i-a80-usb-mod-clk" andMark Kettenis
"allwinner,sun9i-a80-usb-phy-clk" compatile clocks. Enable parent clock of clock gates.
2016-11-21Simplify pad calculation for flow match on swofp_put_flow().Rafael Zalamena
ok reyk@
2016-11-21Follow RFC 5722 more strictly when handling overlapping fragmentsAlexander Bluhm
in pf. Drop the whole fragment state if IPv6 fragments appear which have invalid length or fragment-offset or more-fragment-bit. In IPv4 they are considered invalid and just dropped like before. Found by Antonios Atlasis; OK sashan@ sthen@
2016-11-21Add ability to change media typeMike Belopuhov
Tested with a X540 interconnected with a X550 via a CAT6 twisted pair cable, but is expected to work on multi-speed fiber modules as well to select between 10GbaseLR and 1000baseLX or 10GbaseSR and 1000baseSX, etc. This is largely required because X550 doesn't provide support for auto-negotiation and requires manual configuration. Obtained from FreeBSD.
2016-11-21Fixup a PCIe transaction completion timeout issueMike Belopuhov
Mirrors the change 71bde60191a9fa44d33b582c5e3713ffe99b0fef in Linux; taken as a part of the 280182 commit to FreeBSD.
2016-11-21Factor out RSS initialization into a separate functionMike Belopuhov
No functional change.
2016-11-21In pf_route() and pf_route6() the !r->rt case was only used byAlexander Bluhm
af-to. pf_route6() called ip6_output() to do the work while pf_route() had some custom implementation for that. It is simpler to call ip_output() or ip6_output() from pf_test() directly. OK procter@ sashan@
2016-11-21Enabling Loongson 3A bits turned on a code path that uses a MIPS64r2Visa Hankala
register on Loongson 2. This causes a boot failure on LS2 because the CPU does not implement the register. Disable the code on LS2 similarly to mips64 pte.h. Yeelong boot hang reported and fix tested by matthieu@
2016-11-21Correctly set the default value for the FCRTHMike Belopuhov
X540 datasheet specifies that Flow Control Receive Threshold High should be set to the Rx Buffer Size minus the delay value which is different for different modes of operation, however the minimum is 0x6000 (24576). Mirrors the bc1fc64fd2d9093496e5b04c6d94d26bfa629c9c commit to the Linux source code, but picked up from the 280182 commit to FreeBSD.
2016-11-21Correct the test for requiring a 64 bit mem bar.Jonathan Gray
ok mikeb@ who tested on 82599, x540 and x550.
2016-11-21Turn on the PHY power during attachMike Belopuhov
After a cold boot the PHY power might be disabled by another OS: https://svnweb.freebsd.org/base?view=revision&revision=295093
2016-11-21Assert that defrtrlist_update() is always called at IPL_SOFTNET.Martin Pieuchot