Re: move PRU_ACCEPT request to (*pru_accept)()

2022-08-21 Thread Alexander Bluhm
On Mon, Aug 22, 2022 at 02:29:45AM +0300, Vitaliy Makkoveev wrote: > guenther@ said, that sys/protosw.h hunk on top of the diff makes it > easier to review. Is it really helpful? > > ok? OK bluhm@ > Index: sys/sys/protosw.h > === >

move PRU_ACCEPT request to (*pru_accept)()

2022-08-21 Thread Vitaliy Makkoveev
guenther@ said, that sys/protosw.h hunk on top of the diff makes it easier to review. Is it really helpful? ok? Index: sys/sys/protosw.h === RCS file: /cvs/src/sys/sys/protosw.h,v retrieving revision 1.40 diff -u -p -r1.40 protosw.h

Re: move PRU_CONNECT request to (*pru_connect)() handler

2022-08-21 Thread Alexander Bluhm
On Sun, Aug 21, 2022 at 10:08:12PM +0300, Vitaliy Makkoveev wrote: > ok? OK bluhm@ > Index: sys/kern/uipc_usrreq.c > === > RCS file: /cvs/src/sys/kern/uipc_usrreq.c,v > retrieving revision 1.170 > diff -u -p -r1.170 uipc_usrreq.c > -

Re: netlock igmp mld6 timer

2022-08-21 Thread Vitaliy Makkoveev
On Sun, Aug 21, 2022 at 07:19:23PM +0200, Alexander Bluhm wrote: > Hi, > > When testing parallel UDP input, I see contention on the netlock > from the protocol timer. There is a check to avoid loop traversal > if igmp and mld6 are not active. Move the shortcut to avoid grabbing > netlock. This

move PRU_CONNECT request to (*pru_connect)() handler

2022-08-21 Thread Vitaliy Makkoveev
ok? Index: sys/kern/uipc_usrreq.c === RCS file: /cvs/src/sys/kern/uipc_usrreq.c,v retrieving revision 1.170 diff -u -p -r1.170 uipc_usrreq.c --- sys/kern/uipc_usrreq.c 21 Aug 2022 17:30:21 - 1.170 +++ sys/kern/uipc_usrre

Re: move PRU_LISTEN request to (*pru_listen)() handler

2022-08-21 Thread Alexander Bluhm
On Sun, Aug 21, 2022 at 08:07:10PM +0300, Vitaliy Makkoveev wrote: > ok? OK bluhm@ > Index: sys/kern/uipc_usrreq.c > === > RCS file: /cvs/src/sys/kern/uipc_usrreq.c,v > retrieving revision 1.169 > diff -u -p -r1.169 uipc_usrreq.c > -

netlock igmp mld6 timer

2022-08-21 Thread Alexander Bluhm
Hi, When testing parallel UDP input, I see contention on the netlock from the protocol timer. There is a check to avoid loop traversal if igmp and mld6 are not active. Move the shortcut to avoid grabbing netlock. This should be save as variables igmp_timers_are_running and mld_timers_are_runnin

Re: running UDP input in parallel

2022-08-21 Thread Alexander Bluhm
On Fri, Aug 19, 2022 at 10:54:42PM +0200, Alexander Bluhm wrote: > This diff allows to run udp_input() in parallel. It consists of > three major parts. > > - Use PR_MPSAFE flag to protocol deliver loop with shared > netlock. Queue packet and switch to deliver loop with exclusive > netlock, o

move PRU_LISTEN request to (*pru_listen)() handler

2022-08-21 Thread Vitaliy Makkoveev
ok? Index: sys/kern/uipc_usrreq.c === RCS file: /cvs/src/sys/kern/uipc_usrreq.c,v retrieving revision 1.169 diff -u -p -r1.169 uipc_usrreq.c --- sys/kern/uipc_usrreq.c 20 Aug 2022 23:48:57 - 1.169 +++ sys/kern/uipc_usrre

Re: use libkern bzero on i386

2022-08-21 Thread Mike Larkin
On Mon, Aug 22, 2022 at 12:31:45AM +1000, Jonathan Gray wrote: > libkern bzero doesn't have the 486 path but is otherwise the same > ok mlarkin (but didnt test) > diff --git sys/arch/i386/i386/locore.s sys/arch/i386/i386/locore.s > index dba6ce75b81..3055a06812c 100644 > --- sys/arch/i386/i386/lo

Re: regress: vmd: disable on i386

2022-08-21 Thread Mike Larkin
On Sat, Aug 20, 2022 at 09:03:59AM +, Klemens Nanni wrote: > vmd/Makefile filters for amd64 itself but still, no need to enter > on !amd64. > > Index: ../Makefile > === > RCS file: /cvs/src/regress/usr.sbin/Makefile,v > retrieving

Re: comment fixes for i386/locore.s

2022-08-21 Thread Jonathan Gray
On Sun, Aug 21, 2022 at 02:03:54AM -0400, Daniel Dickman wrote: > locore.s rev 1.113 from 2007 removed support for 80386 processors from > copyin and copyout. However, the note about 386 CPUs in the comment for > copyout was missed. Patch below updates the comment to match the code. > > While he

use libkern bzero on i386

2022-08-21 Thread Jonathan Gray
libkern bzero doesn't have the 486 path but is otherwise the same diff --git sys/arch/i386/i386/locore.s sys/arch/i386/i386/locore.s index dba6ce75b81..3055a06812c 100644 --- sys/arch/i386/i386/locore.s +++ sys/arch/i386/i386/locore.s @@ -1518,73 +1518,6 @@ _C_LABEL(doreti_iret): #include #incl

Re: remove ip local

2022-08-21 Thread Vitaliy Makkoveev
ok mvs@ > On 21 Aug 2022, at 16:04, Alexander Bluhm wrote: > > Hi, > > After moving the IPv4 fragment reassembly and IPv6 hob-by-hob header > chain processing out of ip_local() and ip6_local(), we can remove > these almost empty stubs. No functional change intended. > > ok? > > bluhm > > In

remove ip local

2022-08-21 Thread Alexander Bluhm
Hi, After moving the IPv4 fragment reassembly and IPv6 hob-by-hob header chain processing out of ip_local() and ip6_local(), we can remove these almost empty stubs. No functional change intended. ok? bluhm Index: netinet/ip_input.c ==

Re: Change soabort() return type to void

2022-08-21 Thread Alexander Bluhm
On Sun, Aug 21, 2022 at 01:58:12AM +0300, Vitaliy Makkoveev wrote: > We never interesting soabort() return value. Also 4.4BSD does not use the return value. OK bluhm@ > I don't see any reason to change pru_abort() return type right now, > because it calls (*pru_usrreq)() within. I'll modify it w