From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nir Tzachar Subject: Re: [PATCH 1/1] net: Introduce recvmmsg socket syscall Date: Mon, 12 Oct 2009 19:53:43 +0200 Message-ID: <9b2db90b0910121053h3c422beet487cc9a9b9be2894@mail.gmail.com> References: <1255364440-23271-1-git-send-email-acme@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , netdev@vger.kernel.org, Arnaldo Carvalho de Melo , Caitlin Bestler , Chris Van Hoof , Clark Williams , Neil Horman , Nivedita Singhvi , Paul Moore , =?ISO-8859-1?Q?R=E9mi_Denis=2DCourmont?= , Steven Whitehouse To: Arnaldo Carvalho de Melo Return-path: Received: from mail-bw0-f210.google.com ([209.85.218.210]:53838 "EHLO mail-bw0-f210.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932672AbZJLRyW convert rfc822-to-8bit (ORCPT ); Mon, 12 Oct 2009 13:54:22 -0400 Received: by bwz6 with SMTP id 6so3301063bwz.37 for ; Mon, 12 Oct 2009 10:53:44 -0700 (PDT) In-Reply-To: <1255364440-23271-1-git-send-email-acme@redhat.com> Sender: netdev-owner@vger.kernel.org List-ID: Hi Arnaldo. Do you have any plans on how we can further investigate the delays I have seen with the second part of the patch? I have tried to simply unlock/lock the socket's mutex every couple of iterations inside the loop (to allow the system to process some backlog), but this seems to have little to no effect. Also, a way to enable/disable the no_lock version at runtime will greatly help in testing. Maybe by first introducing a second syscall, recvmmsg_no_lock, for testing purposes?? Cheers, Nir. On Mon, Oct 12, 2009 at 6:20 PM, Arnaldo Carvalho de Melo wrote: > Meaning receive multiple messages, reducing the number of syscalls an= d > net stack entry/exit operations. > > Next patches will introduce mechanisms where protocols that want to > optimize this operation will provide an unlocked_recvmsg operation. > > This takes into account comments made by: > > . Paul Moore: sock_recvmsg is called only for the first datagram, > =A0sock_recvmsg_nosec is used for the rest. > > . Caitlin Bestler: recvmmsg now has a struct timespec timeout, that > =A0works in the same fashion as the ppoll one. > > =A0If the underlying protocol returns a datagram with MSG_OOB set, th= is > =A0will make recvmmsg return right away with as many datagrams (+ the= OOB > =A0one) it has received so far. > > . R=E9mi Denis-Courmont & Steven Whitehouse: If we receive N < vlen > =A0datagrams and then recvmsg returns an error, recvmmsg will return > =A0the successfully received datagrams, store the error and return it > =A0in the next call. > > This paves the way for a subsequent optimization, sk_prot->unlocked_r= ecvmsg, > where we will be able to acquire the lock only at batch start and end= , not at > every underlying recvmsg call. > > Cc: Caitlin Bestler > Cc: Chris Van Hoof > Cc: Clark Williams > Cc: Neil Horman > Cc: Nir Tzachar > Cc: Nivedita Singhvi > Cc: Paul Moore > Cc: R=E9mi Denis-Courmont > Cc: Steven Whitehouse > Signed-off-by: Arnaldo Carvalho de Melo > --- > =A0arch/alpha/kernel/systbls.S =A0 =A0 =A0 =A0 =A0 =A0| =A0 =A01 + > =A0arch/arm/kernel/calls.S =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0 =A01 = + > =A0arch/avr32/kernel/syscall_table.S =A0 =A0 =A0| =A0 =A01 + > =A0arch/blackfin/mach-common/entry.S =A0 =A0 =A0| =A0 =A01 + > =A0arch/ia64/kernel/entry.S =A0 =A0 =A0 =A0 =A0 =A0 =A0 | =A0 =A01 + > =A0arch/microblaze/kernel/syscall_table.S | =A0 =A01 + > =A0arch/mips/kernel/scall32-o32.S =A0 =A0 =A0 =A0 | =A0 =A01 + > =A0arch/mips/kernel/scall64-64.S =A0 =A0 =A0 =A0 =A0| =A0 =A01 + > =A0arch/mips/kernel/scall64-n32.S =A0 =A0 =A0 =A0 | =A0 =A01 + > =A0arch/mips/kernel/scall64-o32.S =A0 =A0 =A0 =A0 | =A0 =A01 + > =A0arch/sh/kernel/syscalls_64.S =A0 =A0 =A0 =A0 =A0 | =A0 =A01 + > =A0arch/sparc/kernel/systbls_64.S =A0 =A0 =A0 =A0 | =A0 =A04 +- > =A0arch/x86/ia32/ia32entry.S =A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0 =A01 + > =A0arch/x86/include/asm/unistd_32.h =A0 =A0 =A0 | =A0 =A03 +- > =A0arch/x86/include/asm/unistd_64.h =A0 =A0 =A0 | =A0 =A02 + > =A0arch/x86/kernel/syscall_table_32.S =A0 =A0 | =A0 =A01 + > =A0arch/xtensa/include/asm/unistd.h =A0 =A0 =A0 | =A0 =A04 +- > =A0include/linux/net.h =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0 =A0= 1 + > =A0include/linux/socket.h =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 | =A0 10 ++ > =A0include/linux/syscalls.h =A0 =A0 =A0 =A0 =A0 =A0 =A0 | =A0 =A04 + > =A0include/net/compat.h =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 | =A0 =A0= 8 + > =A0kernel/sys_ni.c =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0| =A0= =A02 + > =A0net/compat.c =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 |= =A0 33 +++++- > =A0net/socket.c =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 |= =A0225 ++++++++++++++++++++++++++------ > =A024 files changed, 260 insertions(+), 49 deletions(-) > > diff --git a/arch/alpha/kernel/systbls.S b/arch/alpha/kernel/systbls.= S > index 95c9aef..cda6b8b 100644 > --- a/arch/alpha/kernel/systbls.S > +++ b/arch/alpha/kernel/systbls.S > @@ -497,6 +497,7 @@ sys_call_table: > =A0 =A0 =A0 =A0.quad sys_signalfd > =A0 =A0 =A0 =A0.quad sys_ni_syscall > =A0 =A0 =A0 =A0.quad sys_eventfd > + =A0 =A0 =A0 .quad sys_recvmmsg > > =A0 =A0 =A0 =A0.size sys_call_table, . - sys_call_table > =A0 =A0 =A0 =A0.type sys_call_table, @object > diff --git a/arch/arm/kernel/calls.S b/arch/arm/kernel/calls.S > index fafce1b..f58c115 100644 > --- a/arch/arm/kernel/calls.S > +++ b/arch/arm/kernel/calls.S > @@ -374,6 +374,7 @@ > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0CALL(sys_pwritev) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0CALL(sys_rt_tgsigqueueinfo) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0CALL(sys_perf_event_open) > +/* 365 */ =A0 =A0 =A0CALL(sys_recvmmsg) > =A0#ifndef syscalls_counted > =A0.equ syscalls_padding, ((NR_syscalls + 3) & ~3) - NR_syscalls > =A0#define syscalls_counted > diff --git a/arch/avr32/kernel/syscall_table.S b/arch/avr32/kernel/sy= scall_table.S > index 7ee0057..e76bad1 100644 > --- a/arch/avr32/kernel/syscall_table.S > +++ b/arch/avr32/kernel/syscall_table.S > @@ -295,4 +295,5 @@ sys_call_table: > =A0 =A0 =A0 =A0.long =A0 sys_signalfd > =A0 =A0 =A0 =A0.long =A0 sys_ni_syscall =A0 =A0 =A0 =A0 =A0/* 280, wa= s sys_timerfd */ > =A0 =A0 =A0 =A0.long =A0 sys_eventfd > + =A0 =A0 =A0 .long =A0 sys_recvmmsg > =A0 =A0 =A0 =A0.long =A0 sys_ni_syscall =A0 =A0 =A0 =A0 =A0/* r8 is s= aturated at nr_syscalls */ > diff --git a/arch/blackfin/mach-common/entry.S b/arch/blackfin/mach-c= ommon/entry.S > index 1e7cac2..4869272 100644 > --- a/arch/blackfin/mach-common/entry.S > +++ b/arch/blackfin/mach-common/entry.S > @@ -1621,6 +1621,7 @@ ENTRY(_sys_call_table) > =A0 =A0 =A0 =A0.long _sys_pwritev > =A0 =A0 =A0 =A0.long _sys_rt_tgsigqueueinfo > =A0 =A0 =A0 =A0.long _sys_perf_event_open > + =A0 =A0 =A0 .long _sys_recvmmsg =A0 =A0 =A0 =A0 =A0 =A0 /* 370 */ > > =A0 =A0 =A0 =A0.rept NR_syscalls-(.-_sys_call_table)/4 > =A0 =A0 =A0 =A0.long _sys_ni_syscall > diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S > index d0e7d37..d75b872 100644 > --- a/arch/ia64/kernel/entry.S > +++ b/arch/ia64/kernel/entry.S > @@ -1806,6 +1806,7 @@ sys_call_table: > =A0 =A0 =A0 =A0data8 sys_preadv > =A0 =A0 =A0 =A0data8 sys_pwritev =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 = =A0 =A0 // 1320 > =A0 =A0 =A0 =A0data8 sys_rt_tgsigqueueinfo > + =A0 =A0 =A0 data8 sys_recvmmsg > > =A0 =A0 =A0 =A0.org sys_call_table + 8*NR_syscalls =A0 =A0 // guard a= gainst failures to increase NR_syscalls > =A0#endif /* __IA64_ASM_PARAVIRTUALIZED_NATIVE */ > diff --git a/arch/microblaze/kernel/syscall_table.S b/arch/microblaze= /kernel/syscall_table.S > index ecec191..c1ab1dc 100644 > --- a/arch/microblaze/kernel/syscall_table.S > +++ b/arch/microblaze/kernel/syscall_table.S > @@ -371,3 +371,4 @@ ENTRY(sys_call_table) > =A0 =A0 =A0 =A0.long sys_ni_syscall > =A0 =A0 =A0 =A0.long sys_rt_tgsigqueueinfo =A0 =A0 /* 365 */ > =A0 =A0 =A0 =A0.long sys_perf_event_open > + =A0 =A0 =A0 .long sys_recvmmsg > diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall3= 2-o32.S > index fd2a9bb..17202bb 100644 > --- a/arch/mips/kernel/scall32-o32.S > +++ b/arch/mips/kernel/scall32-o32.S > @@ -583,6 +583,7 @@ einval: =A0 =A0 li =A0 =A0 =A0v0, -ENOSYS > =A0 =A0 =A0 =A0sys =A0 =A0 sys_rt_tgsigqueueinfo =A0 4 > =A0 =A0 =A0 =A0sys =A0 =A0 sys_perf_event_open =A0 =A0 5 > =A0 =A0 =A0 =A0sys =A0 =A0 sys_accept4 =A0 =A0 =A0 =A0 =A0 =A0 4 > + =A0 =A0 =A0 sys =A0 =A0 sys_recvmmsg =A0 =A0 =A0 =A0 =A0 =A05 > =A0 =A0 =A0 =A0.endm > > =A0 =A0 =A0 =A0/* We pre-compute the number of _instruction_ bytes ne= eded to > diff --git a/arch/mips/kernel/scall64-64.S b/arch/mips/kernel/scall64= -64.S > index 18bf7f3..a8a6c59 100644 > --- a/arch/mips/kernel/scall64-64.S > +++ b/arch/mips/kernel/scall64-64.S > @@ -420,4 +420,5 @@ sys_call_table: > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_rt_tgsigqueueinfo > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_perf_event_open > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_accept4 > + =A0 =A0 =A0 PTR =A0 =A0 sys_recvmmsg > =A0 =A0 =A0 =A0.size =A0 sys_call_table,.-sys_call_table > diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall6= 4-n32.S > index 6ebc079..5154e64 100644 > --- a/arch/mips/kernel/scall64-n32.S > +++ b/arch/mips/kernel/scall64-n32.S > @@ -418,4 +418,5 @@ EXPORT(sysn32_call_table) > =A0 =A0 =A0 =A0PTR =A0 =A0 compat_sys_rt_tgsigqueueinfo =A0 =A0/* 529= 5 */ > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_perf_event_open > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_accept4 > + =A0 =A0 =A0 PTR =A0 =A0 compat_sys_recvmmsg > =A0 =A0 =A0 =A0.size =A0 sysn32_call_table,.-sysn32_call_table > diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall6= 4-o32.S > index 9bbf977..d0eff53 100644 > --- a/arch/mips/kernel/scall64-o32.S > +++ b/arch/mips/kernel/scall64-o32.S > @@ -538,4 +538,5 @@ sys_call_table: > =A0 =A0 =A0 =A0PTR =A0 =A0 compat_sys_rt_tgsigqueueinfo > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_perf_event_open > =A0 =A0 =A0 =A0PTR =A0 =A0 sys_accept4 > + =A0 =A0 =A0 PTR =A0 =A0 compat_sys_recvmmsg > =A0 =A0 =A0 =A0.size =A0 sys_call_table,.-sys_call_table > diff --git a/arch/sh/kernel/syscalls_64.S b/arch/sh/kernel/syscalls_6= 4.S > index 5bfde6c..07d2aae 100644 > --- a/arch/sh/kernel/syscalls_64.S > +++ b/arch/sh/kernel/syscalls_64.S > @@ -391,3 +391,4 @@ sys_call_table: > =A0 =A0 =A0 =A0.long sys_pwritev > =A0 =A0 =A0 =A0.long sys_rt_tgsigqueueinfo > =A0 =A0 =A0 =A0.long sys_perf_event_open > + =A0 =A0 =A0 .long sys_recvmmsg =A0 =A0 =A0 =A0 =A0 =A0 =A0/* 365 */ > diff --git a/arch/sparc/kernel/systbls_64.S b/arch/sparc/kernel/systb= ls_64.S > index 009825f..f37bef7 100644 > --- a/arch/sparc/kernel/systbls_64.S > +++ b/arch/sparc/kernel/systbls_64.S > @@ -83,7 +83,7 @@ sys_call_table32: > =A0/*310*/ =A0 =A0 =A0 =A0.word compat_sys_utimensat, compat_sys_sign= alfd, sys_timerfd_create, sys_eventfd, compat_sys_fallocate > =A0 =A0 =A0 =A0.word compat_sys_timerfd_settime, compat_sys_timerfd_g= ettime, compat_sys_signalfd4, sys_eventfd2, sys_epoll_create1 > =A0/*320*/ =A0 =A0 =A0 =A0.word sys_dup3, sys_pipe2, sys_inotify_init= 1, sys_accept4, compat_sys_preadv > - =A0 =A0 =A0 .word compat_sys_pwritev, compat_sys_rt_tgsigqueueinfo,= sys_perf_event_open > + =A0 =A0 =A0 .word compat_sys_pwritev, compat_sys_rt_tgsigqueueinfo,= sys_perf_event_open, compat_sys_recvmmsg > > =A0#endif /* CONFIG_COMPAT */ > > @@ -158,4 +158,4 @@ sys_call_table: > =A0/*310*/ =A0 =A0 =A0 =A0.word sys_utimensat, sys_signalfd, sys_time= rfd_create, sys_eventfd, sys_fallocate > =A0 =A0 =A0 =A0.word sys_timerfd_settime, sys_timerfd_gettime, sys_si= gnalfd4, sys_eventfd2, sys_epoll_create1 > =A0/*320*/ =A0 =A0 =A0 =A0.word sys_dup3, sys_pipe2, sys_inotify_init= 1, sys_accept4, sys_preadv > - =A0 =A0 =A0 .word sys_pwritev, sys_rt_tgsigqueueinfo, sys_perf_even= t_open > + =A0 =A0 =A0 .word sys_pwritev, sys_rt_tgsigqueueinfo, sys_perf_even= t_open, sys_recvmmsg > diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S > index 74619c4..11a6c79 100644 > --- a/arch/x86/ia32/ia32entry.S > +++ b/arch/x86/ia32/ia32entry.S > @@ -832,4 +832,5 @@ ia32_sys_call_table: > =A0 =A0 =A0 =A0.quad compat_sys_pwritev > =A0 =A0 =A0 =A0.quad compat_sys_rt_tgsigqueueinfo =A0 =A0 =A0/* 335 *= / > =A0 =A0 =A0 =A0.quad sys_perf_event_open > + =A0 =A0 =A0 .quad compat_sys_recvmmsg > =A0ia32_syscall_end: > diff --git a/arch/x86/include/asm/unistd_32.h b/arch/x86/include/asm/= unistd_32.h > index 6fb3c20..3baf379 100644 > --- a/arch/x86/include/asm/unistd_32.h > +++ b/arch/x86/include/asm/unistd_32.h > @@ -342,10 +342,11 @@ > =A0#define __NR_pwritev =A0 =A0 =A0 =A0 =A0 334 > =A0#define __NR_rt_tgsigqueueinfo 335 > =A0#define __NR_perf_event_open =A0 336 > +#define __NR_recvmmsg =A0 =A0 =A0 =A0 =A0337 > > =A0#ifdef __KERNEL__ > > -#define NR_syscalls 337 > +#define NR_syscalls 338 > > =A0#define __ARCH_WANT_IPC_PARSE_VERSION > =A0#define __ARCH_WANT_OLD_READDIR > diff --git a/arch/x86/include/asm/unistd_64.h b/arch/x86/include/asm/= unistd_64.h > index 8d3ad0a..4843f7b 100644 > --- a/arch/x86/include/asm/unistd_64.h > +++ b/arch/x86/include/asm/unistd_64.h > @@ -661,6 +661,8 @@ __SYSCALL(__NR_pwritev, sys_pwritev) > =A0__SYSCALL(__NR_rt_tgsigqueueinfo, sys_rt_tgsigqueueinfo) > =A0#define __NR_perf_event_open =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 2= 98 > =A0__SYSCALL(__NR_perf_event_open, sys_perf_event_open) > +#define __NR_recvmmsg =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0299 > +__SYSCALL(__NR_recvmmsg, sys_recvmmsg) > > =A0#ifndef __NO_STUBS > =A0#define __ARCH_WANT_OLD_READDIR > diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/sys= call_table_32.S > index 0157cd2..70c2125 100644 > --- a/arch/x86/kernel/syscall_table_32.S > +++ b/arch/x86/kernel/syscall_table_32.S > @@ -336,3 +336,4 @@ ENTRY(sys_call_table) > =A0 =A0 =A0 =A0.long sys_pwritev > =A0 =A0 =A0 =A0.long sys_rt_tgsigqueueinfo =A0 =A0 /* 335 */ > =A0 =A0 =A0 =A0.long sys_perf_event_open > + =A0 =A0 =A0 .long sys_recvmmsg > diff --git a/arch/xtensa/include/asm/unistd.h b/arch/xtensa/include/a= sm/unistd.h > index c092c8f..4e55dc7 100644 > --- a/arch/xtensa/include/asm/unistd.h > +++ b/arch/xtensa/include/asm/unistd.h > @@ -681,8 +681,10 @@ __SYSCALL(304, sys_signalfd, 3) > =A0__SYSCALL(305, sys_ni_syscall, 0) > =A0#define __NR_eventfd =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 306 > =A0__SYSCALL(306, sys_eventfd, 1) > +#define __NR_recvmmsg =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0307 > +__SYSCALL(307, sys_recvmmsg, 5) > > -#define __NR_syscall_count =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 3= 07 > +#define __NR_syscall_count =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 3= 08 > > =A0/* > =A0* sysxtensa syscall handler > diff --git a/include/linux/net.h b/include/linux/net.h > index 529a093..b42bb60 100644 > --- a/include/linux/net.h > +++ b/include/linux/net.h > @@ -41,6 +41,7 @@ > =A0#define SYS_SENDMSG =A0 =A016 =A0 =A0 =A0 =A0 =A0 =A0 =A0/* sys_se= ndmsg(2) =A0 =A0 =A0 =A0 =A0 =A0 =A0 */ > =A0#define SYS_RECVMSG =A0 =A017 =A0 =A0 =A0 =A0 =A0 =A0 =A0/* sys_re= cvmsg(2) =A0 =A0 =A0 =A0 =A0 =A0 =A0 */ > =A0#define SYS_ACCEPT4 =A0 =A018 =A0 =A0 =A0 =A0 =A0 =A0 =A0/* sys_ac= cept4(2) =A0 =A0 =A0 =A0 =A0 =A0 =A0 */ > +#define SYS_RECVMMSG =A0 19 =A0 =A0 =A0 =A0 =A0 =A0 =A0/* sys_recvmm= sg(2) =A0 =A0 =A0 =A0 =A0 =A0 =A0*/ > > =A0typedef enum { > =A0 =A0 =A0 =A0SS_FREE =3D 0, =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= /* not allocated =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/ > diff --git a/include/linux/socket.h b/include/linux/socket.h > index 3273a0c..59966f1 100644 > --- a/include/linux/socket.h > +++ b/include/linux/socket.h > @@ -65,6 +65,12 @@ struct msghdr { > =A0 =A0 =A0 =A0unsigned =A0 =A0 =A0 =A0msg_flags; > =A0}; > > +/* For recvmmsg/sendmmsg */ > +struct mmsghdr { > + =A0 =A0 =A0 struct msghdr =A0 msg_hdr; > + =A0 =A0 =A0 unsigned =A0 =A0 =A0 =A0msg_len; > +}; > + > =A0/* > =A0* =A0 =A0 POSIX 1003.1g - ancillary data object information > =A0* =A0 =A0 Ancillary data consits of a sequence of pairs of > @@ -312,6 +318,10 @@ extern int move_addr_to_user(struct sockaddr *ka= ddr, int klen, void __user *uadd > =A0extern int move_addr_to_kernel(void __user *uaddr, int ulen, struc= t sockaddr *kaddr); > =A0extern int put_cmsg(struct msghdr*, int level, int type, int len, = void *data); > > +struct timespec; > + > +extern int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsig= ned int vlen, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 unsigned int flags,= struct timespec *timeout); > =A0#endif > =A0#endif /* not kernel and not glibc */ > =A0#endif /* _LINUX_SOCKET_H */ > diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h > index a990ace..714f063 100644 > --- a/include/linux/syscalls.h > +++ b/include/linux/syscalls.h > @@ -25,6 +25,7 @@ struct linux_dirent64; > =A0struct list_head; > =A0struct msgbuf; > =A0struct msghdr; > +struct mmsghdr; > =A0struct msqid_ds; > =A0struct new_utsname; > =A0struct nfsctl_arg; > @@ -677,6 +678,9 @@ asmlinkage long sys_recv(int, void __user *, size= _t, unsigned); > =A0asmlinkage long sys_recvfrom(int, void __user *, size_t, unsigned, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0struct= sockaddr __user *, int __user *); > =A0asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, uns= igned flags); > +asmlinkage long sys_recvmmsg(int fd, struct mmsghdr __user *msg, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0unsigned int= vlen, unsigned flags, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0struct times= pec __user *timeout); > =A0asmlinkage long sys_socket(int, int, int); > =A0asmlinkage long sys_socketpair(int, int, int, int __user *); > =A0asmlinkage long sys_socketcall(int call, unsigned long __user *arg= s); > diff --git a/include/net/compat.h b/include/net/compat.h > index 7c30028..9679f05 100644 > --- a/include/net/compat.h > +++ b/include/net/compat.h > @@ -18,6 +18,11 @@ struct compat_msghdr { > =A0 =A0 =A0 =A0compat_uint_t =A0 msg_flags; > =A0}; > > +struct compat_mmsghdr { > + =A0 =A0 =A0 struct compat_msghdr msg_hdr; > + =A0 =A0 =A0 compat_uint_t =A0 =A0 =A0 =A0msg_len; > +}; > + > =A0struct compat_cmsghdr { > =A0 =A0 =A0 =A0compat_size_t =A0 cmsg_len; > =A0 =A0 =A0 =A0compat_int_t =A0 =A0cmsg_level; > @@ -35,6 +40,9 @@ extern int get_compat_msghdr(struct msghdr *, struc= t compat_msghdr __user *); > =A0extern int verify_compat_iovec(struct msghdr *, struct iovec *, st= ruct sockaddr *, int); > =A0extern asmlinkage long compat_sys_sendmsg(int,struct compat_msghdr= __user *,unsigned); > =A0extern asmlinkage long compat_sys_recvmsg(int,struct compat_msghdr= __user *,unsigned); > +extern asmlinkage long compat_sys_recvmmsg(int, struct compat_mmsghd= r __user *, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0 =A0unsigned, unsigned, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0 =A0struct timespec __user *); > =A0extern asmlinkage long compat_sys_getsockopt(int, int, int, char _= _user *, int __user *); > =A0extern int put_cmsg_compat(struct msghdr*, int, int, int, void *); > > diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c > index e06d0b8..f050ba8 100644 > --- a/kernel/sys_ni.c > +++ b/kernel/sys_ni.c > @@ -48,8 +48,10 @@ cond_syscall(sys_shutdown); > =A0cond_syscall(sys_sendmsg); > =A0cond_syscall(compat_sys_sendmsg); > =A0cond_syscall(sys_recvmsg); > +cond_syscall(sys_recvmmsg); > =A0cond_syscall(compat_sys_recvmsg); > =A0cond_syscall(compat_sys_recvfrom); > +cond_syscall(compat_sys_recvmmsg); > =A0cond_syscall(sys_socketcall); > =A0cond_syscall(sys_futex); > =A0cond_syscall(compat_sys_futex); > diff --git a/net/compat.c b/net/compat.c > index a407c3a..e13f525 100644 > --- a/net/compat.c > +++ b/net/compat.c > @@ -727,10 +727,10 @@ EXPORT_SYMBOL(compat_mc_getsockopt); > > =A0/* Argument list sizes for compat_sys_socketcall */ > =A0#define AL(x) ((x) * sizeof(u32)) > -static unsigned char nas[19]=3D{AL(0),AL(3),AL(3),AL(3),AL(2),AL(3), > +static unsigned char nas[20]=3D{AL(0),AL(3),AL(3),AL(3),AL(2),AL(3), > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0AL(3),= AL(3),AL(4),AL(4),AL(4),AL(6), > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0AL(6),= AL(2),AL(5),AL(5),AL(3),AL(3), > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 AL(4)}; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 AL(4),A= L(5)}; > =A0#undef AL > > =A0asmlinkage long compat_sys_sendmsg(int fd, struct compat_msghdr __= user *msg, unsigned flags) > @@ -755,13 +755,36 @@ asmlinkage long compat_sys_recvfrom(int fd, voi= d __user *buf, size_t len, > =A0 =A0 =A0 =A0return sys_recvfrom(fd, buf, len, flags | MSG_CMSG_COM= PAT, addr, addrlen); > =A0} > > +asmlinkage long compat_sys_recvmmsg(int fd, struct compat_mmsghdr __= user *mmsg, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= unsigned vlen, unsigned int flags, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= struct timespec __user *timeout) > +{ > + =A0 =A0 =A0 int datagrams; > + =A0 =A0 =A0 struct timespec ktspec; > + =A0 =A0 =A0 struct compat_timespec __user *utspec =3D > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 (struct compat_timespec= __user *)timeout; > + > + =A0 =A0 =A0 if (get_user(ktspec.tv_sec, &utspec->tv_sec) || > + =A0 =A0 =A0 =A0 =A0 get_user(ktspec.tv_nsec, &utspec->tv_nsec)) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -EFAULT; > + > + =A0 =A0 =A0 datagrams =3D __sys_recvmmsg(fd, (struct mmsghdr __user= *)mmsg, vlen, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= flags | MSG_CMSG_COMPAT, &ktspec); > + =A0 =A0 =A0 if (datagrams > 0 && > + =A0 =A0 =A0 =A0 =A0 (put_user(ktspec.tv_sec, &utspec->tv_sec) || > + =A0 =A0 =A0 =A0 =A0 =A0put_user(ktspec.tv_nsec, &utspec->tv_nsec))) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 datagrams =3D -EFAULT; > + > + =A0 =A0 =A0 return datagrams; > +} > + > =A0asmlinkage long compat_sys_socketcall(int call, u32 __user *args) > =A0{ > =A0 =A0 =A0 =A0int ret; > =A0 =A0 =A0 =A0u32 a[6]; > =A0 =A0 =A0 =A0u32 a0, a1; > > - =A0 =A0 =A0 if (call < SYS_SOCKET || call > SYS_ACCEPT4) > + =A0 =A0 =A0 if (call < SYS_SOCKET || call > SYS_RECVMMSG) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0return -EINVAL; > =A0 =A0 =A0 =A0if (copy_from_user(a, args, nas[call])) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0return -EFAULT; > @@ -823,6 +846,10 @@ asmlinkage long compat_sys_socketcall(int call, = u32 __user *args) > =A0 =A0 =A0 =A0case SYS_RECVMSG: > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0ret =3D compat_sys_recvmsg(a0, compat_= ptr(a1), a[2]); > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0break; > + =A0 =A0 =A0 case SYS_RECVMMSG: > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 ret =3D compat_sys_recvmmsg(a0, compat_= ptr(a1), a[2], a[3], > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0 compat_ptr(a[4])); > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > =A0 =A0 =A0 =A0case SYS_ACCEPT4: > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0ret =3D sys_accept4(a0, compat_ptr(a1)= , compat_ptr(a[2]), a[3]); > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0break; > diff --git a/net/socket.c b/net/socket.c > index 954f338..3dd03df 100644 > --- a/net/socket.c > +++ b/net/socket.c > @@ -668,10 +668,9 @@ void __sock_recv_timestamp(struct msghdr *msg, s= truct sock *sk, > > =A0EXPORT_SYMBOL_GPL(__sock_recv_timestamp); > > -static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *= sock, > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0stru= ct msghdr *msg, size_t size, int flags) > +static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct so= cket *sock, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0struct msghdr *msg, size_t size, int flags) > =A0{ > - =A0 =A0 =A0 int err; > =A0 =A0 =A0 =A0struct sock_iocb *si =3D kiocb_to_siocb(iocb); > > =A0 =A0 =A0 =A0si->sock =3D sock; > @@ -680,13 +679,17 @@ static inline int __sock_recvmsg(struct kiocb *= iocb, struct socket *sock, > =A0 =A0 =A0 =A0si->size =3D size; > =A0 =A0 =A0 =A0si->flags =3D flags; > > - =A0 =A0 =A0 err =3D security_socket_recvmsg(sock, msg, size, flags)= ; > - =A0 =A0 =A0 if (err) > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 return err; > - > =A0 =A0 =A0 =A0return sock->ops->recvmsg(iocb, sock, msg, size, flags= ); > =A0} > > +static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *= sock, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0stru= ct msghdr *msg, size_t size, int flags) > +{ > + =A0 =A0 =A0 int err =3D security_socket_recvmsg(sock, msg, size, fl= ags); > + > + =A0 =A0 =A0 return err ?: __sock_recvmsg_nosec(iocb, sock, msg, siz= e, flags); > +} > + > =A0int sock_recvmsg(struct socket *sock, struct msghdr *msg, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 size_t size, int flags) > =A0{ > @@ -702,6 +705,21 @@ int sock_recvmsg(struct socket *sock, struct msg= hdr *msg, > =A0 =A0 =A0 =A0return ret; > =A0} > > +static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *ms= g, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 size_t size= , int flags) > +{ > + =A0 =A0 =A0 struct kiocb iocb; > + =A0 =A0 =A0 struct sock_iocb siocb; > + =A0 =A0 =A0 int ret; > + > + =A0 =A0 =A0 init_sync_kiocb(&iocb, NULL); > + =A0 =A0 =A0 iocb.private =3D &siocb; > + =A0 =A0 =A0 ret =3D __sock_recvmsg_nosec(&iocb, sock, msg, size, fl= ags); > + =A0 =A0 =A0 if (-EIOCBQUEUED =3D=3D ret) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 ret =3D wait_on_sync_kiocb(&iocb); > + =A0 =A0 =A0 return ret; > +} > + > =A0int kernel_recvmsg(struct socket *sock, struct msghdr *msg, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 struct kvec *vec, size_t num, siz= e_t size, int flags) > =A0{ > @@ -1968,22 +1986,15 @@ out: > =A0 =A0 =A0 =A0return err; > =A0} > > -/* > - * =A0 =A0 BSD recvmsg interface > - */ > - > -SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg, > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 unsigned int, flags) > +static int __sys_recvmsg(struct socket *sock, struct msghdr __user *= msg, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0struct msghdr *msg_s= ys, unsigned flags, int nosec) > =A0{ > =A0 =A0 =A0 =A0struct compat_msghdr __user *msg_compat =3D > =A0 =A0 =A0 =A0 =A0 =A0(struct compat_msghdr __user *)msg; > - =A0 =A0 =A0 struct socket *sock; > =A0 =A0 =A0 =A0struct iovec iovstack[UIO_FASTIOV]; > =A0 =A0 =A0 =A0struct iovec *iov =3D iovstack; > - =A0 =A0 =A0 struct msghdr msg_sys; > =A0 =A0 =A0 =A0unsigned long cmsg_ptr; > =A0 =A0 =A0 =A0int err, iov_size, total_len, len; > - =A0 =A0 =A0 int fput_needed; > > =A0 =A0 =A0 =A0/* kernel mode address */ > =A0 =A0 =A0 =A0struct sockaddr_storage addr; > @@ -1993,27 +2004,23 @@ SYSCALL_DEFINE3(recvmsg, int, fd, struct msgh= dr __user *, msg, > =A0 =A0 =A0 =A0int __user *uaddr_len; > > =A0 =A0 =A0 =A0if (MSG_CMSG_COMPAT & flags) { > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (get_compat_msghdr(&msg_sys, msg_com= pat)) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (get_compat_msghdr(msg_sys, msg_comp= at)) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0return -EFAULT; > =A0 =A0 =A0 =A0} > - =A0 =A0 =A0 else if (copy_from_user(&msg_sys, msg, sizeof(struct ms= ghdr))) > + =A0 =A0 =A0 else if (copy_from_user(msg_sys, msg, sizeof(struct msg= hdr))) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0return -EFAULT; > > - =A0 =A0 =A0 sock =3D sockfd_lookup_light(fd, &err, &fput_needed); > - =A0 =A0 =A0 if (!sock) > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out; > - > =A0 =A0 =A0 =A0err =3D -EMSGSIZE; > - =A0 =A0 =A0 if (msg_sys.msg_iovlen > UIO_MAXIOV) > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out_put; > + =A0 =A0 =A0 if (msg_sys->msg_iovlen > UIO_MAXIOV) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out; > > =A0 =A0 =A0 =A0/* Check whether to allocate the iovec area */ > =A0 =A0 =A0 =A0err =3D -ENOMEM; > - =A0 =A0 =A0 iov_size =3D msg_sys.msg_iovlen * sizeof(struct iovec); > - =A0 =A0 =A0 if (msg_sys.msg_iovlen > UIO_FASTIOV) { > + =A0 =A0 =A0 iov_size =3D msg_sys->msg_iovlen * sizeof(struct iovec)= ; > + =A0 =A0 =A0 if (msg_sys->msg_iovlen > UIO_FASTIOV) { > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0iov =3D sock_kmalloc(sock->sk, iov_siz= e, GFP_KERNEL); > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0if (!iov) > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out_put; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out; > =A0 =A0 =A0 =A0} > > =A0 =A0 =A0 =A0/* > @@ -2021,46 +2028,47 @@ SYSCALL_DEFINE3(recvmsg, int, fd, struct msgh= dr __user *, msg, > =A0 =A0 =A0 =A0 * =A0 =A0 =A0kernel msghdr to use the kernel address = space) > =A0 =A0 =A0 =A0 */ > > - =A0 =A0 =A0 uaddr =3D (__force void __user *)msg_sys.msg_name; > + =A0 =A0 =A0 uaddr =3D (__force void __user *)msg_sys->msg_name; > =A0 =A0 =A0 =A0uaddr_len =3D COMPAT_NAMELEN(msg); > =A0 =A0 =A0 =A0if (MSG_CMSG_COMPAT & flags) { > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D verify_compat_iovec(&msg_sys, i= ov, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D verify_compat_iovec(msg_sys, io= v, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0(struct sockaddr *)&addr, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0VERIFY_WRITE); > =A0 =A0 =A0 =A0} else > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D verify_iovec(&msg_sys, iov, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D verify_iovec(msg_sys, iov, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 (= struct sockaddr *)&addr, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 V= ERIFY_WRITE); > =A0 =A0 =A0 =A0if (err < 0) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto out_freeiov; > =A0 =A0 =A0 =A0total_len =3D err; > > - =A0 =A0 =A0 cmsg_ptr =3D (unsigned long)msg_sys.msg_control; > - =A0 =A0 =A0 msg_sys.msg_flags =3D flags & (MSG_CMSG_CLOEXEC|MSG_CMS= G_COMPAT); > + =A0 =A0 =A0 cmsg_ptr =3D (unsigned long)msg_sys->msg_control; > + =A0 =A0 =A0 msg_sys->msg_flags =3D flags & (MSG_CMSG_CLOEXEC|MSG_CM= SG_COMPAT); > > =A0 =A0 =A0 =A0if (sock->file->f_flags & O_NONBLOCK) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0flags |=3D MSG_DONTWAIT; > - =A0 =A0 =A0 err =3D sock_recvmsg(sock, &msg_sys, total_len, flags); > + =A0 =A0 =A0 err =3D (nosec ? sock_recvmsg_nosec : sock_recvmsg)(soc= k, msg_sys, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 total_len, flags); > =A0 =A0 =A0 =A0if (err < 0) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto out_freeiov; > =A0 =A0 =A0 =A0len =3D err; > > =A0 =A0 =A0 =A0if (uaddr !=3D NULL) { > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0err =3D move_addr_to_user((struct sock= addr *)&addr, > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 msg_sys.msg_namelen, uaddr, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0 msg_sys->msg_namelen, uaddr, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= =A0 =A0uaddr_len); > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0if (err < 0) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto out_freeiov; > =A0 =A0 =A0 =A0} > - =A0 =A0 =A0 err =3D __put_user((msg_sys.msg_flags & ~MSG_CMSG_COMPA= T), > + =A0 =A0 =A0 err =3D __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMP= AT), > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 COMPAT_FLAGS(msg)); > =A0 =A0 =A0 =A0if (err) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto out_freeiov; > =A0 =A0 =A0 =A0if (MSG_CMSG_COMPAT & flags) > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D __put_user((unsigned long)msg_s= ys.msg_control - cmsg_ptr, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D __put_user((unsigned long)msg_s= ys->msg_control - cmsg_ptr, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 &msg_= compat->msg_controllen); > =A0 =A0 =A0 =A0else > - =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D __put_user((unsigned long)msg_s= ys.msg_control - cmsg_ptr, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D __put_user((unsigned long)msg_s= ys->msg_control - cmsg_ptr, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 &msg-= >msg_controllen); > =A0 =A0 =A0 =A0if (err) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto out_freeiov; > @@ -2069,21 +2077,150 @@ SYSCALL_DEFINE3(recvmsg, int, fd, struct msg= hdr __user *, msg, > =A0out_freeiov: > =A0 =A0 =A0 =A0if (iov !=3D iovstack) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0sock_kfree_s(sock->sk, iov, iov_size); > -out_put: > +out: > + =A0 =A0 =A0 return err; > +} > + > +/* > + * =A0 =A0 BSD recvmsg interface > + */ > + > +SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 unsigned int, flags) > +{ > + =A0 =A0 =A0 int fput_needed, err; > + =A0 =A0 =A0 struct msghdr msg_sys; > + =A0 =A0 =A0 struct socket *sock =3D sockfd_lookup_light(fd, &err, &= fput_needed); > + > + =A0 =A0 =A0 if (!sock) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out; > + > + =A0 =A0 =A0 err =3D __sys_recvmsg(sock, msg, &msg_sys, flags, 0); > + > =A0 =A0 =A0 =A0fput_light(sock->file, fput_needed); > =A0out: > =A0 =A0 =A0 =A0return err; > =A0} > > -#ifdef __ARCH_WANT_SYS_SOCKETCALL > +/* > + * =A0 =A0 Linux recvmmsg interface > + */ > + > +int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int= vlen, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0unsigned int flags, struct times= pec *timeout) > +{ > + =A0 =A0 =A0 int fput_needed, err, datagrams; > + =A0 =A0 =A0 struct socket *sock; > + =A0 =A0 =A0 struct mmsghdr __user *entry; > + =A0 =A0 =A0 struct msghdr msg_sys; > + =A0 =A0 =A0 struct timespec end_time; > + > + =A0 =A0 =A0 if (timeout && > + =A0 =A0 =A0 =A0 =A0 poll_select_set_timeout(&end_time, timeout->tv_= sec, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= timeout->tv_nsec)) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -EINVAL; > + > + =A0 =A0 =A0 datagrams =3D 0; > + > + =A0 =A0 =A0 sock =3D sockfd_lookup_light(fd, &err, &fput_needed); > + =A0 =A0 =A0 if (!sock) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return err; > + > + =A0 =A0 =A0 err =3D sock_error(sock->sk); > + =A0 =A0 =A0 if (err) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto out_put; > + > + =A0 =A0 =A0 entry =3D mmsg; > + > + =A0 =A0 =A0 while (datagrams < vlen) { > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* No need to ask LSM for more than t= he first datagram. > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/ > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D __sys_recvmsg(sock, (struct msg= hdr __user *)entry, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= &msg_sys, flags, datagrams); > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (err < 0) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D put_user(err, &entry->msg_len); > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (err) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 ++entry; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 ++datagrams; > + > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (timeout) { > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 ktime_get_ts(timeout); > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 *timeout =3D timespec_s= ub(end_time, *timeout); > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (timeout->tv_sec < 0= ) { > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 timeout= ->tv_sec =3D timeout->tv_nsec =3D 0; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 } > + > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* Timeout, return less= than vlen datagrams */ > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (timeout->tv_nsec =3D= =3D 0 && timeout->tv_sec =3D=3D 0) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 } > + > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* Out of band data, return right away = */ > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (msg_sys.msg_flags & MSG_OOB) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > + =A0 =A0 =A0 } > + > +out_put: > + =A0 =A0 =A0 fput_light(sock->file, fput_needed); > > + =A0 =A0 =A0 if (err =3D=3D 0) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return datagrams; > + > + =A0 =A0 =A0 if (datagrams !=3D 0) { > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* We may return less entries than re= quested (vlen) if the > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* sock is non block and there aren't= enough datagrams... > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/ > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (err !=3D -EAGAIN) { > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* ... or =A0if recvm= sg returns an error after we > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* received some data= grams, where we record the > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* error to return on= the next call or if the > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* app asks about it = using getsockopt(SO_ERROR). > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/ > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 sock->sk->sk_err =3D -e= rr; > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 } > + > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return datagrams; > + =A0 =A0 =A0 } > + > + =A0 =A0 =A0 return err; > +} > + > +SYSCALL_DEFINE5(recvmmsg, int, fd, struct mmsghdr __user *, mmsg, > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 unsigned int, vlen, unsigned int, flags= , > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 struct timespec __user *, timeout) > +{ > + =A0 =A0 =A0 int datagrams; > + =A0 =A0 =A0 struct timespec timeout_sys; > + > + =A0 =A0 =A0 if (!timeout) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return __sys_recvmmsg(fd, mmsg, vlen, f= lags, NULL); > + > + =A0 =A0 =A0 if (copy_from_user(&timeout_sys, timeout, sizeof(timeou= t_sys))) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -EFAULT; > + > + =A0 =A0 =A0 datagrams =3D __sys_recvmmsg(fd, mmsg, vlen, flags, &ti= meout_sys); > + > + =A0 =A0 =A0 if (datagrams > 0 && > + =A0 =A0 =A0 =A0 =A0 copy_to_user(timeout, &timeout_sys, sizeof(time= out_sys))) > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 datagrams =3D -EFAULT; > + > + =A0 =A0 =A0 return datagrams; > +} > + > +#ifdef __ARCH_WANT_SYS_SOCKETCALL > =A0/* Argument list sizes for sys_socketcall */ > =A0#define AL(x) ((x) * sizeof(unsigned long)) > -static const unsigned char nargs[19]=3D{ > +static const unsigned char nargs[20] =3D { > =A0 =A0 =A0 =A0AL(0),AL(3),AL(3),AL(3),AL(2),AL(3), > =A0 =A0 =A0 =A0AL(3),AL(3),AL(4),AL(4),AL(4),AL(6), > =A0 =A0 =A0 =A0AL(6),AL(2),AL(5),AL(5),AL(3),AL(3), > - =A0 =A0 =A0 AL(4) > + =A0 =A0 =A0 AL(4),AL(5) > =A0}; > > =A0#undef AL > @@ -2103,7 +2240,7 @@ SYSCALL_DEFINE2(socketcall, int, call, unsigned= long __user *, args) > =A0 =A0 =A0 =A0int err; > =A0 =A0 =A0 =A0unsigned int len; > > - =A0 =A0 =A0 if (call < 1 || call > SYS_ACCEPT4) > + =A0 =A0 =A0 if (call < 1 || call > SYS_RECVMMSG) > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0return -EINVAL; > > =A0 =A0 =A0 =A0len =3D nargs[call]; > @@ -2181,6 +2318,10 @@ SYSCALL_DEFINE2(socketcall, int, call, unsigne= d long __user *, args) > =A0 =A0 =A0 =A0case SYS_RECVMSG: > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0err =3D sys_recvmsg(a0, (struct msghdr= __user *)a1, a[2]); > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0break; > + =A0 =A0 =A0 case SYS_RECVMMSG: > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 err =3D sys_recvmmsg(a0, (struct mmsghd= r __user *)a1, a[2], a[3], > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0= (struct timespec __user *)a[4]); > + =A0 =A0 =A0 =A0 =A0 =A0 =A0 break; > =A0 =A0 =A0 =A0case SYS_ACCEPT4: > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0err =3D sys_accept4(a0, (struct sockad= dr __user *)a1, > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(i= nt __user *)a[2], a[3]); > -- > 1.5.5.1 > >