Date: Fri, 13 Apr 2012 11:37:44 -0700 (PDT) From: Sushanth Rai <sushanth_rai@yahoo.com> To: Konstantin Belousov <kostikbel@gmail.com> Cc: freebsd-hackers@freebsd.org Subject: Re: mlockall() on freebsd 7.2 + amd64 returns EAGAIN Message-ID: <1334342264.8672.YahooMailClassic@web180006.mail.gq1.yahoo.com>
next in thread | raw e-mail | index | archive | help
---729476438-1186198209-1334342264=:8672 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable I've attached the simple program that creates 5 threads. Following is the o= /p of /proc/<pid>/map when this program is running. Note that I modified = =0Asys/fs/procfs/procfs_map.c to print whether a region is wired. As you ca= n see from this o/p, none of stack areas get wired. =0A=0A0x400000 0x401000= 1 0 0xffffff002d943bd0 r-x 1 0 0x1000 COW NC wired vnode /var/tmp/thread1= =0A0x500000 0x501000 1 0 0xffffff002dd13e58 rw- 2 0 0x3100 NCOW NNC wired d= efault -=0A0x501000 0x600000 255 0 0xffffff002dd13e58 rwx 2 0 0x3100 NCOW N= NC wired default -=0A0x800500000 0x800526000 38 0 0xffffff0025574000 r-x 19= 2 46 0x1004 COW NC wired vnode /libexec/ld-elf.so.1=0A0x800526000 0x8005370= 00 17 0 0xffffff002d9f81b0 rw- 1 0 0x3100 NCOW NNC wired default -=0A0x8006= 26000 0x80062d000 7 0 0xffffff002dd13bd0 rw- 1 0 0x3100 COW NNC wired vnode= /libexec/ld-elf.so.1=0A0x80062d000 0x800633000 6 0 0xffffff002dd145e8 rw- = 1 0 0x3100 NCOW NNC wired default -=0A0x800633000 0x800645000 18 0 0xffffff= 00256d71b0 r-x 63 42 0x4 COW NC wired vnode /lib/libthr.so.3=0A0x800645000 = 0x800646000 1 0 0xffffff002d975510 r-x 1 0 0x3100 COW NNC wired vnode /lib/= libthr.so.3=0A0x800646000 0x800746000 0 0 0xffffff002dc5cca8 --- 4 0 0x3100= NCOW NNC not-wired default -=0A0x800746000 0x80074a000 4 0 0xffffff002572a= 288 rw- 1 0 0x3100 COW NNC wired vnode /lib/libthr.so.3=0A0x80074a000 0x800= 74c000 2 0 0xffffff002dc5cca8 rw- 4 0 0x3100 NCOW NNC wired default -=0A0x8= 0074c000 0x80083e000 242 0 0xffffff001cd226c0 r-x 238 92 0x1004 COW NC wire= d vnode /lib/libc.so.7=0A0x80083e000 0x80083f000 1 0 0xffffff002dd12000 r-x= 1 0 0x3100 COW NNC wired vnode /lib/libc.so.7=0A0x80083f000 0x80093e000 0 = 0 0xffffff002dc5cca8 --- 4 0 0x3100 NCOW NNC not-wired default -=0A0x80093e= 000 0x80095d000 31 0 0xffffff002dddc360 rw- 1 0 0x3100 COW NNC wired vnode = /lib/libc.so.7=0A0x80095d000 0x800974000 23 0 0xffffff002dc5cca8 rw- 4 0 0x= 3100 NCOW NNC wired default -=0A0x800a00000 0x800b00000 256 0 0xffffff002db= d1798 rw- 1 0 0x3100 NCOW NNC wired default -=0A0x800b00000 0x800c00000 256= 0 0xffffff002dd14948 rw- 1 0 0x3100 NCOW NNC wired default -=0A0x7fffff3db= 000 0x7fffff3fb000 1 0 0xffffff002dbb4360 rw- 1 0 0x3100 NCOW NNC not-wired= default -=0A0x7fffff5dc000 0x7fffff5fc000 1 0 0xffffff002dc66af8 rw- 1 0 0= x3100 NCOW NNC not-wired default -=0A0x7fffff7dd000 0x7fffff7fd000 1 0 0xff= ffff002dbea438 rw- 1 0 0x3100 NCOW NNC not-wired default -=0A0x7fffff9de000= 0x7fffff9fe000 1 0 0xffffff002dd7fd80 rw- 1 0 0x3100 NCOW NNC not-wired de= fault -=0A0x7fffffbdf000 0x7fffffbff000 1 0 0xffffff002dbe9438 rw- 1 0 0x31= 00 NCOW NNC not-wired default -=0A0x7fffffbff000 0x7fffffc00000 0 0 0 --- 0= 0 0x0 NCOW NNC not-wired none -=0A0x7ffffffe0000 0x800000000000 32 0 0xfff= fff002dd125e8 rwx 1 0 0x3100 NCOW NNC wired default -=0A=0A--- On Fri, 4/13= /12, Konstantin Belousov <kostikbel@gmail.com> wrote:=0A=0A> From: Konstant= in Belousov <kostikbel@gmail.com>=0A> Subject: Re: mlockall() on freebsd 7.= 2 + amd64 returns EAGAIN=0A> To: "Sushanth Rai" <sushanth_rai@yahoo.com>=0A= > Cc: freebsd-hackers@freebsd.org=0A> Date: Friday, April 13, 2012, 1:11 AM= =0A> On Thu, Apr 12, 2012 at 08:10:26PM=0A> -0700, Sushanth Rai wrote:=0A> = > > =0A> > > Then it should be fixed in r190885.=0A> > > =0A> > =0A> > Than= ks. That works like a charm. =0A> > =0A> > mlockall() mostly works now. The= re is still a, issue in=0A> wiring the stacks of multithreaded program when= the program=0A> uses default stack allocation scheme. Thread library=0A> a= llocates stack for each thread by calling mmap() and=0A> sending address an= d size to be mapped. The kernel adjusts=0A> the start address to sgrowsz in= =A0 vm_map_stack() and=0A> maps at the adjusted address. But the subsequent= wiring is=0A> done using the original address, which fails. =0A> > =0A> Ca= n you, please provide stand-alone example which=0A> demostrates the issue ?= =0A> I suspect this should have not changed in HEAD.=0A> ---729476438-1186198209-1334342264=:8672--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1334342264.8672.YahooMailClassic>