CVSROOT: /cvs
Module name: src
Changes by: [email protected] 2015/11/24 20:10:00
Modified files:
share/man/man9 : Makefile ifq_enqueue.9
sys/arch/armv7/imx: imxenet.c
sys/arch/armv7/omap: if_cpsw.c
sys/arch/armv7/sunxi: sxie.c
sys/arch/macppc/dev: if_bm.c if_mc.c
sys/arch/octeon/dev: if_cnmac.c
sys/arch/sgi/dev: if_iec.c if_mec.c
sys/arch/sgi/hpc: if_sq.c
sys/arch/socppc/dev: if_tsec.c
sys/arch/sparc/dev: be.c hme.c if_ie.c qe.c
sys/arch/sparc64/dev: vnet.c
sys/arch/vax/if: if_de.c if_qe.c sgec.c
sys/dev/ic : acx.c aic6915.c am7990.c am79900.c an.c
ar5008.c ar9003.c ath.c athn.c atw.c bwi.c dc.c
dp8390.c elink3.c fxp.c gem.c hme.c i82596.c
if_wi.c lance.c lemac.c malo.c mtd8xx.c pgt.c
re.c rt2560.c rt2661.c rt2860.c rtl81x9.c rtw.c
smc83c170.c smc91cxx.c ti.c xl.c
sys/dev/isa : if_ef_isapnp.c if_eg.c if_el.c if_ex.c if_ie.c
sys/dev/pci : if_age.c if_alc.c if_ale.c if_bce.c if_bge.c
if_bnx.c if_cas.c if_de.c if_em.c if_et.c
if_ipw.c if_iwi.c if_iwm.c if_iwn.c if_ix.c
if_ixgb.c if_jme.c if_lge.c if_lii.c if_msk.c
if_myx.c if_nep.c if_nfe.c if_nge.c if_nxe.c
if_oce.c if_pcn.c if_rtwn.c if_se.c if_sis.c
if_sk.c if_ste.c if_stge.c if_tht.c if_tl.c
if_txp.c if_vge.c if_vic.c if_vio.c if_vmx.c
if_vr.c if_vte.c if_wb.c if_wpi.c if_xge.c
sys/dev/pcmcia : if_malo.c if_xe.c
sys/dev/sbus : be.c qe.c
sys/dev/usb : if_athn_usb.c if_atu.c if_aue.c if_axe.c
if_axen.c if_cdce.c if_cdcef.c if_cue.c
if_kue.c if_mos.c if_otus.c if_ral.c if_rsu.c
if_rum.c if_run.c if_smsc.c if_uath.c if_udav.c
if_ugl.c if_upgt.c if_upl.c if_url.c
if_urndis.c if_urtw.c if_urtwn.c if_zyd.c
sys/net : if.c if_var.h
sys/net80211 : ieee80211_pae_output.c
Log message:
replace IFF_OACTIVE manipulation with mpsafe operations.
there are two things shared between the network stack and drivers
in the send path: the send queue and the IFF_OACTIVE flag. the send
queue is now protected by a mutex. this diff makes the oactive
functionality mpsafe too.
IFF_OACTIVE is part of if_flags. there are two problems with that.
firstly, if_flags is a short and we dont have any MI atomic operations
to manipulate a short. secondly, while we could make the IFF_OACTIVE
operates mpsafe, all changes to other flags would have to be made
safe at the same time, otherwise a read-modify-write cycle on their
updates could clobber the oactive change.
instead, this moves the oactive mark into struct ifqueue and provides
an API for changing it. there's ifq_set_oactive, ifq_clr_oactive,
and ifq_is_oactive. these are modelled on ifsq_set_oactive,
ifsq_clr_oactive, and ifsq_is_oactive in dragonflybsd.
this diff includes changes to all the drivers manipulating IFF_OACTIVE
to now use the ifsq_{set,clr_is}_oactive API too.
ok kettenis@ mpi@ jmatthew@ deraadt@
