summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2014-09-05Don't bother messing with minphys, we always need to supportStefan Fritsch
transfers of size MAXBSIZE. If the number of DMA segments supported by the device is too low, don't use it. Discussion at http://marc.info/?l=openbsd-tech&m=140978303923966&w=2
2014-09-05syncChris Cappuccio
2014-09-05Another Intel 10GbEChris Cappuccio
2014-09-05KASSERT that the page header pool will use in page headers.David Gwynne
2014-09-04Remove global cp_time[] array; no longer used now that all arches implementMiod Vallat
cpu_info.
2014-09-04Revert previous change, but use curcpu()->ci_schedstate.spc_cp_time[] insteadMiod Vallat
of global cp_time[] for live information, as loadavg is not updated fast enough for hz/16 blinking rate. Prompted by feedback from mickey@
2014-09-04rework how pools with large pages (>PAGE_SIZE) are implemented.David Gwynne
this moves the size of the pool page (not arch page) out of the pool allocator into struct pool. this lets us create only two pools for the automatically determined large page allocations instead of 256 of them. while here support using slack space in large pages for the pool_item_header by requiring km_alloc provide pool page aligned memory. lastly, instead of doing incorrect math to figure how how many arch pages to use for large pool pages, just use powers of two. ok mikeb@
2014-09-03When a route to prefix (connected route) is added, if its associatedMartin Pieuchot
broadcast address is non null, add a broadcast entry flagged with RTF_BROADCAST. Re-use the existing logic to switch a route to prefix from an ifa to another to also move this broadcast entry. Prior to this change broadcast entries were simple clonned ARP entries, that would be deleted once their timer expired since they would always be incomplete. With this change they are now persistant and identifiable with a new flag This version of the diff prevent a corruption reported by millert@ ok mikeb@, florian@
2014-09-03Make sure broadcast entries won't be freed by the arp timer so we canMartin Pieuchot
use them for address lookups. While here do let in_arpinput() overwrite local or broadcast entries. ok mikeb@, florian@
2014-09-03Treat broadcast entries like local ones and give them the highestMartin Pieuchot
route priority. While here document a RTM_ADD message is only send to userland for local entries. ok mikeb@, florian@
2014-09-03Sync amd64 and i386 GENERIC.MP with other arches byBret Lambert
including a commented-out MP_LOCKDEBUG option ok mpi@ phessler@ dlg@ guenther@
2014-09-03tasks dont need a mutex and a semaphore to protect against multiple usesDavid Gwynne
of the task structure like workq tasks did. tested on an fc929
2014-09-02Add Jumbo support for BCM5714 / BCM5780 and BCM5717 / BCM5719 / BCM5720 / ↵Brad Smith
BCM57765 / BCM57766 chipsets. ok mikeb@ "i think it should go in" dlg@
2014-09-01Set the configuration number and get the interface handle in attach(),Martin Pieuchot
like urtwn(4) does, to prevent a null dereference of the configuration descriptor. Issue reported and fix tested by Ludovic Coues on misc@
2014-09-01Revert previous, it puts too much pressure on memory allocation asMartin Pieuchot
reported by Bryan Linton on bugs@.
2014-09-01move mpi(4) from workq to taskqBret Lambert
ok dlg@
2014-09-01Make this consistent with sys/kern/Makefile: require the 'syscalls' targetPhilip Guenther
2014-09-01sync generated filesDoug Hogan
2014-09-01Sync readlink(2) with IEEE Std 1003.1-2008.Doug Hogan
discussion, help and ok guenther@
2014-09-01Add implementations of atomic_{inc,dec,add,sub}_{int,long}() andPhilip Guenther
atomic_{add,sub}_{int,long}_nv(), based on Solaris and translated through the mind of dlg@ Remove the 'memory' constraint from atomic_cas_{uint,ulong,ptr}() now that rw_{enter,exit}*() use membars. ok dlg@
2014-09-01Add explicit membars, so that we can stop requiring rw_cas() toPhilip Guenther
provide the magic. ok matthew@ dlg@
2014-08-31replace LRU bufcache with something originally modelled after 2Q.Ted Unangst
this should provide a degree of scan resistance, and also serves as a midway point for further development of multi queue algorithms. i've tried to minimize the risk and degree of regressions. probably ok beck
2014-08-31Sync enc/dec routines with libc.Miod Vallat
2014-08-31Make the in6addr constant declarations and definitions consistentAlexander Bluhm
in kernel and user land. OK florian@ mpi@
2014-08-31Add uuid support routines to libc. From FreeBSD via NetBSD via Bitrig viaMiod Vallat
Markus Mueller.
2014-08-31Add sockatmark()Philip Guenther
ok millert@ manpage feedback jmc@
2014-08-31regenPhilip Guenther
2014-08-31Add additional kernel interfaces for setting close-on-exec on fdsPhilip Guenther
when creating them: pipe2(), dup3(), accept4(), MSG_CMSG_CLOEXEC, SOCK_CLOEXEC. Includes SOCK_NONBLOCK support. ok matthew@
2014-08-30Fix format strings in STIDEBUG code.Miod Vallat
2014-08-30Recognize EFI protective and system partitions. Do not put disklabel entriesMiod Vallat
for the protective ones when creating a fake label, but do, for the system ones, so that we may eventually copy boot code to them. From Markus Mueller
2014-08-30Make sure we check for GPT in writedisklabel() if kernel is compiled with GPTMiod Vallat
support, for it would risk being overwritten otherwise. From Markus Mueller
2014-08-30Set the MRU to a full size frame instead of basing it on the MTU.Brad Smith
ok dlg@
2014-08-30let the mru always be what the chip can do, not what the mtu implies.David Gwynne
tested by and ok mikeb@
2014-08-30Don't panic when an illegal IRQ is passed to intr_{dis,}establish().Martin Pieuchot
At least the device-tree for the second bge(4) on my PowerMac is not parsed correctly but we can live with that until the bug is fixed.
2014-08-30Bump the fan limit and avoid a panic on my quad G5.Martin Pieuchot
2014-08-30Rewrite how queue heads are inserted and removed from the asynchronousMartin Pieuchot
list. Use a queue head per xfer and no longer per pipe. Since queue heads are now setup when a transfer is submited we do not need to open the control pipe more than once. This also simplify transfer abortion. Tested by many, thanks!
2014-08-30Allow new devices to get an address when XHCI_DEBUG is defined.Martin Pieuchot
2014-08-30Fix profiling (GPROF) build.Martin Pieuchot
2014-08-29Use the load average from averunnable.ldavg[] instead of bogus cp_time[]Miod Vallat
arithmetic to display the load average in leds 5-8 of 8 led systems.
2014-08-29Fix checking sync for old synaptics touchpad (ver 5.9)Alexandr Shadchin
From Thierry Deval <thierry at deval dot be> ok mpi@
2014-08-27Nuke net.inet6.icmp6.rediraccept and allow redirects on interfacesFlorian Obser
with autoconf enabled. If one is doing SLAAC one does already trust link local icmp6 so the policy for icmp6 redirects should be the same. pointed out by & OK bluhm@; OK henning@
2014-08-27rxr ioctl handling.David Gwynne
2014-08-27Remove redundant setting of if_mtu which ether_ifattach() alreadyBrad Smith
takes care of. ok dlg@
2014-08-27deprecate the "item offset" handling. nothing uses it, so we canDavid Gwynne
cut it out of the code to simplify things. ok mikeb@
2014-08-26dont base the mru on the mtu. unconditionally make it what theDavid Gwynne
hardware can do (9k). implement the rxr ioctl while here. ok jsg@
2014-08-26Do not hand-roll all routers link local address.Florian Obser
pointed out & OK bluhm@
2014-08-26Revert the addition of broadcast addresses, it apparently also corruptsMartin Pieuchot
the tree. Found by millert@.
2014-08-26Revert part of the if_rxr diff that incorrectly moves RX ring tailMike Belopuhov
index update code from the buf_get success path to the do it all the time code path. Tested by millert; ok dlg, deraadt
2014-08-25We have never limited the definition of "supported SPF modules"Mike Belopuhov
to the vendor/make whitelist maintained by Intel so there's no reason to start doing it now. When syncing the driver to the FreeBSD codebase I have decided to take this chunk as is but it appears that it breaks cheap chinese SFP+ fiber optics modules that we all love. And while there's still a lot of places where we check for the vendor OUI, most of these checks are not necessary. Issue reported and fix tested by Tony Sarendal. Thanks!
2014-08-25Move sending of router solicitations to the kernel; receiving andFlorian Obser
processing of router advertisements was already in the kernel. With this rtsol{,d}(8) is no longer necessary. The kernel starts sending solicitations with # ifconfig $IF inet6 autoconf or inet6 autoconf in /etc/hostname.$IF. input stsp@ much help & OK mpi@ tweaks & OK bluhm@
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-30 05:55:04 +0000