summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2015-02-11Save OpenFirmware mappings in save_ofw_mapping(), currently unusedMartin Pieuchot
but required for upcoming firmware callback improvement. ok miod@
2015-02-11deprecate use of sys/lock.h and replace it with sys/atomic.h orDavid Gwynne
machine/lock.h as appropriate.
2015-02-11Direct config glue for octdwctwo(4).Masao Uebayashi
2015-02-11Explicitely include <sys/atomic.h>Miod Vallat
2015-02-11Extend struct kinfo_file a bit so that netstat has all the info it needs.Claudio Jeker
OK guenther@
2015-02-11Don't rely in <sys/lock.h> bringing <sys/rwlock.h>.Miod Vallat
2015-02-11dont need lockmgr locks here.David Gwynne
2015-02-11Until carp(4) is converted to call ether_ifattach() and ether_ifdetach(),Martin Pieuchot
reroll the loop to free its Ethernet multicast addresses when destroying an interface. Prevent a panic found the hard way by phessler@ ok henning@, pelikan@, phessler@
2015-02-11Eliminate sc_tx_cnt. Instead use an unsigned integer for sc_tx_prod andMark Kettenis
sc_tx_cont and let them wrap around. Make sure we don't fill the last descriptor on the ring such that we don't confuse a completely filled ring with a completely empty one. Also make sure we don't post the same descriptors twice.
2015-02-11Prefer arg != 0 over arg for non-boolean. nitted kettenis@Philip Guenther
2015-02-11the only user of __cpu_cas is lock_machdep, so move it there.David Gwynne
machine/lock.h only provides SPINLOCK_SPIN_HOOK which is included directly by the three files that use it.
2015-02-11Backout non-wanted changed introduced by accident as part of 1.24.Miod Vallat
2015-02-11sysctl({CTL_KERN, KERN_FILE, KERN_FILE_BYFILE}) previously requiredPhilip Guenther
the extra argument to be zero; instead, make it filter on the file type (DTYPE_*) when non-zero to make claudio's netstat work easier. ok claudio@
2015-02-11sys/lock.h and machine/lock.h dont provide anything this code uses.David Gwynne
2015-02-11intr.c needs atomic.h for atomic_setbits_int to say softints are pending.David Gwynne
2015-02-11Fix DIAGNOSTIC build. Trim blank lines.Masao Uebayashi
2015-02-11this doesnt need anything sys/lock.h providesDavid Gwynne
2015-02-11Needs <sys/atomic.h> now.Miod Vallat
2015-02-11Catch up with the new taskq_create(9) API.Masao Uebayashi
2015-02-11get SPINLOCK bits from machine/lock.h instead of sys/lock.hDavid Gwynne
2015-02-11get the SPINLOCK bits from machine/lock.hDavid Gwynne
2015-02-11Convert to uiomove(). ok kettenis@Miod Vallat
2015-02-11alpha (nor any other arch) uses any types or api provided by sys/lock.h.David Gwynne
2015-02-11g/c sys/cdefs.h header.Brad Smith
ok stsp@
2015-02-11slight knf, and __inline is really spelled inline.David Gwynne
2015-02-11Prevent processes from entering iwm_ioctl while another process isStefan Sperling
tsleep'ing in it. From iwn(4).
2015-02-11need sys/atomic.h to get atomic_{set,clear}bits_intDavid Gwynne
2015-02-11need sys/atomic.h for atomic_{set,clear}bits_int.David Gwynne
2015-02-11Show error code in DPRINTF error messages within iwm_send_phy_db_data.Stefan Sperling
2015-02-11we dont need sys/lock.h because we dont use lockmgr, but we do needDavid Gwynne
sys/atomic.h for atomic_setbits_int.
2015-02-11zero ecx before cpuid(1), for some Cyrix processors do not reallyMiod Vallat
support cpuid with other values than zero, and leave the ecx register unchanged. ok kettenis@
2015-02-11Minimal attachment to build-test dwc2(4).Masao Uebayashi
2015-02-11make the rwlock implementation MI.David Gwynne
each arch used to have to provide an rw_cas operation, but now we have the rwlock code build its own version. on smp machines it uses atomic_cas_ulong. on uniproc machines it avoids interlocked instructions by using straight loads and stores. this is safe because rwlocks are only used from process context and processes are currently not preemptible in our kernel. so alpha/ppc/etc might get a benefit. ok miod@ kettenis@ deraadt@
2015-02-10atomic_{cas,swap}_ptr takes a volatile void *, not a volatile void **.David Gwynne
this makes mips64 consistent with the api and all other archs. testing and ok miod@
2015-02-10Convert to uiomove()Miod Vallat
2015-02-10Adjust usb(9) API differences to at least compile.Masao Uebayashi
2015-02-10Testing indicates TX is broken for 11a in iwm(4) so stop advertising 11aStefan Sperling
support to the net80211 layer for now. Discussed with phessler.
2015-02-10Don't use an uninitialized variable when a PT_LOAD segment withPhilip Guenther
alignment 0 or 1 is encountered. The result before was just a spurious failure by execve(), though I had to manually mangle a binary to hit this case: segments are all long-aligned or better in practice. uninitialized variable noted by Maxime Villard (rustyBSD (at) gmx.fr) ok and prod jsg@
2015-02-10Adjust various kernel API differences.Masao Uebayashi
2015-02-10Wireless drivers call if_input() via ieee80211_input() which set `rcvif'Martin Pieuchot
on every received mbuf, so there's no need to initialize this pointer in the drivers. Tested by and ok phessler@
2015-02-10Convert tsleep(9)/wakeup(9) usages.Masao Uebayashi
2015-02-10Replace hand rolled code with m_defrag().Brad Smith
ok pelikan@
2015-02-10Whitespace.Masao Uebayashi
2015-02-10Convert timeout(9)/taskq(9) usages.Masao Uebayashi
2015-02-10Trivial conversion to uiomove()Miod Vallat
2015-02-10Convert sun4 eeprom access to uiomove()Miod Vallat
2015-02-10Convert to uiomove().Miod Vallat
ok guenther@
2015-02-10Convert to uiomove().Miod Vallat
2015-02-10In iwm(4), call ieee80211_media_init() after overriding net80211 stateStefan Sperling
machine methods, like other drivers do.
2015-02-10The attach procedure of iwm(4) was setting up the MAC address correctlyStefan Sperling
but forgot about other information obtained from firmware, like the map of supported channels. Completely re-attach the net80211 layer after first successful firmware load. Fixes output of 'ifconfig iwm0 media'.
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-03 00:58:33 +0000