NetBSD Problem Report #50985
From martin@duskware.de Sun Mar 20 12:45:01 2016
Return-Path: <martin@duskware.de>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
(Client CN "mail.netbsd.org", Issuer "Postmaster NetBSD.org" (verified OK))
by mollari.NetBSD.org (Postfix) with ESMTPS id 6D2717ABDF
for <gnats-bugs@gnats.NetBSD.org>; Sun, 20 Mar 2016 12:45:01 +0000 (UTC)
Date: Sun, 20 Mar 2016 13:43:50 CET
From: martin@NetBSD.org
Reply-To: martin@NetBSD.org
To: gnats-bugs@NetBSD.org
Subject: uvm_mmap.c::range_test should use runtime limits of the vmspace?
X-Send-Pr-Version: 3.95
>Number: 50985
>Category: kern
>Synopsis: uvm_mmap.c::range_test should use runtime limits of the vmspace?
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: martin
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Sun Mar 20 12:50:00 +0000 2016
>Closed-Date: Tue May 24 20:22:15 +0000 2016
>Last-Modified: Tue May 24 20:25:01 +0000 2016
>Originator: Martin Husemann
>Release: NetBSD 7.99.26
>Organization:
The NetBSD Foundation, Inc.
>Environment:
System: NetBSD night-owl.duskware.de 7.99.26 NetBSD 7.99.26 (NIGHT-OWL) #393: Fri Mar 11 14:19:47 CET 2016 martin@night-owl.duskware.de:/usr/src/sys/arch/amd64/compile/NIGHT-OWL amd64
Architecture: x86_64
Machine: amd64
>Description:
When creating a vmspace, we take various things (like the binary we are going
to execute) into account. This may result in different min/max values for
the virtual address in the new vmspace.
However, in uvm_mmap.c::range_test() we do not check this (potential differing)
values, but compare against the compile time defaults VM_MIN_ADDRESS
and VM_MAXUSER_ADDRESS.
Differences might be seen on some emulations, or for example with old
sparc64 binaries compiled with a restricted compiler memory model.
>How-To-Repeat:
code inspection
>Fix:
Pass the vmspace to the function and use its vm_map headers .end/.start members
instead? Maybe create accessor macros symetric to vm_map_setmin/vm_map_setmax.
>Release-Note:
>Audit-Trail:
From: Martin Husemann <martin@duskware.de>
To: gnats-bugs@NetBSD.org
Cc: Chuck Silvers <chs@NetBSD.org>
Subject: Re: kern/50985: uvm_mmap.c::range_test should use runtime limits of the vmspace?
Date: Mon, 21 Mar 2016 21:32:32 +0100
An (untestested) simple/mechanical fix would be the change below.
Chuck, care to review?
Martin
Index: uvm_mmap.c
===================================================================
RCS file: /cvsroot/src/sys/uvm/uvm_mmap.c,v
retrieving revision 1.154
diff -u -p -r1.154 uvm_mmap.c
--- uvm_mmap.c 26 Nov 2015 13:15:34 -0000 1.154
+++ uvm_mmap.c 21 Mar 2016 20:27:12 -0000
@@ -70,10 +70,10 @@ static int uvm_mmap(struct vm_map *, vad
int, int, struct uvm_object *, voff_t, vsize_t);
static int
-range_test(vaddr_t addr, vsize_t size, bool ismmap)
+range_test(struct vm_map *map, vaddr_t addr, vsize_t size, bool ismmap)
{
- vaddr_t vm_min_address = VM_MIN_ADDRESS;
- vaddr_t vm_max_address = VM_MAXUSER_ADDRESS;
+ vaddr_t vm_min_address = vm_map_min(map);
+ vaddr_t vm_max_address = vm_map_max(map);
vaddr_t eaddr = addr + size;
int res = 0;
@@ -356,7 +356,7 @@ sys_mmap(struct lwp *l, const struct sys
if (addr & PAGE_MASK)
return (EINVAL);
- error = range_test(addr, size, true);
+ error = range_test(&p->p_vmspace->vm_map, addr, size, true);
if (error) {
return error;
}
@@ -487,16 +487,16 @@ sys___msync13(struct lwp *l, const struc
size += pageoff;
size = (vsize_t)round_page(size);
- error = range_test(addr, size, false);
- if (error)
- return error;
/*
* get map
*/
-
map = &p->p_vmspace->vm_map;
+ error = range_test(map, addr, size, false);
+ if (error)
+ return error;
+
/*
* XXXCDC: do we really need this semantic?
*
@@ -573,12 +573,12 @@ sys_munmap(struct lwp *l, const struct s
if (size == 0)
return (0);
- error = range_test(addr, size, false);
+ map = &p->p_vmspace->vm_map;
+
+ error = range_test(map, addr, size, false);
if (error)
return error;
- map = &p->p_vmspace->vm_map;
-
/*
* interesting system call semantic: make sure entire range is
* allocated before allowing an unmap.
@@ -634,7 +634,7 @@ sys_mprotect(struct lwp *l, const struct
size += pageoff;
size = round_page(size);
- error = range_test(addr, size, false);
+ error = range_test(&p->p_vmspace->vm_map, addr, size, false);
if (error)
return error;
@@ -675,7 +675,7 @@ sys_minherit(struct lwp *l, const struct
size += pageoff;
size = (vsize_t)round_page(size);
- error = range_test(addr, size, false);
+ error = range_test(&p->p_vmspace->vm_map, addr, size, false);
if (error)
return error;
@@ -716,7 +716,7 @@ sys_madvise(struct lwp *l, const struct
size += pageoff;
size = (vsize_t)round_page(size);
- error = range_test(addr, size, false);
+ error = range_test(&p->p_vmspace->vm_map, addr, size, false);
if (error)
return error;
@@ -816,7 +816,7 @@ sys_mlock(struct lwp *l, const struct sy
size += pageoff;
size = (vsize_t)round_page(size);
- error = range_test(addr, size, false);
+ error = range_test(&p->p_vmspace->vm_map, addr, size, false);
if (error)
return error;
@@ -867,7 +867,7 @@ sys_munlock(struct lwp *l, const struct
size += pageoff;
size = (vsize_t)round_page(size);
- error = range_test(addr, size, false);
+ error = range_test(&p->p_vmspace->vm_map, addr, size, false);
if (error)
return error;
From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@NetBSD.org
Cc: kern-bug-people@netbsd.org, gnats-admin@netbsd.org,
netbsd-bugs@netbsd.org, martin@NetBSD.org
Subject: re: kern/50985: uvm_mmap.c::range_test should use runtime limits of the vmspace?
Date: Tue, 22 Mar 2016 07:48:56 +1100
> - vaddr_t vm_min_address = VM_MIN_ADDRESS;
> - vaddr_t vm_max_address = VM_MAXUSER_ADDRESS;
> + vaddr_t vm_min_address = vm_map_min(map);
> + vaddr_t vm_max_address = vm_map_max(map);
the max address now is in the kernel space on some platforms, i think.
.mrg.
From: Martin Husemann <martin@duskware.de>
To: matthew green <mrg@eterna.com.au>
Cc: gnats-bugs@NetBSD.org
Subject: Re: kern/50985: uvm_mmap.c::range_test should use runtime limits of the vmspace?
Date: Mon, 21 Mar 2016 22:13:34 +0100
On Tue, Mar 22, 2016 at 07:48:56AM +1100, matthew green wrote:
> > - vaddr_t vm_min_address = VM_MIN_ADDRESS;
> > - vaddr_t vm_max_address = VM_MAXUSER_ADDRESS;
> > + vaddr_t vm_min_address = vm_map_min(map);
> > + vaddr_t vm_max_address = vm_map_max(map);
>
> the max address now is in the kernel space on some platforms, i think.
It comes from exec paccks ep_vm_maxaddr, which usually is VM_MAXUSER_ADDRESS,
I think.
We can min()/max() it with VM_MAXUSER_ADDRESS/VM_MIN_ADDRESS though, or
KASSERT that the limit is within the static limit.
Martin
From: Chuck Silvers <chuq@chuq.com>
To: Martin Husemann <martin@duskware.de>
Cc: gnats-bugs@NetBSD.org
Subject: Re: kern/50985: uvm_mmap.c::range_test should use runtime limits of
the vmspace?
Date: Mon, 28 Mar 2016 08:51:35 -0700
On Mon, Mar 21, 2016 at 09:32:32PM +0100, Martin Husemann wrote:
> An (untestested) simple/mechanical fix would be the change below.
>
> Chuck, care to review?
looks fine to me.
-Chuck
Responsible-Changed-From-To: kern-bug-people->martin
Responsible-Changed-By: martin@NetBSD.org
Responsible-Changed-When: Tue, 24 May 2016 20:22:15 +0000
Responsible-Changed-Why:
Fixed
State-Changed-From-To: open->closed
State-Changed-By: martin@NetBSD.org
State-Changed-When: Tue, 24 May 2016 20:22:15 +0000
State-Changed-Why:
Fixed
From: "Martin Husemann" <martin@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/50985 CVS commit: src/sys/uvm
Date: Tue, 24 May 2016 20:20:57 +0000
Module Name: src
Committed By: martin
Date: Tue May 24 20:20:57 UTC 2016
Modified Files:
src/sys/uvm: uvm_mmap.c
Log Message:
PR kern/50985: use the runtime limits of the vmspace in range_test()
instead of the compile time defaults for it.
To generate a diff of this commit:
cvs rdiff -u -r1.157 -r1.158 src/sys/uvm/uvm_mmap.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.39 2013/11/01 18:47:49 spz Exp $
$NetBSD: gnats_config.sh,v 1.8 2006/05/07 09:23:38 tsutsui Exp $
Copyright © 1994-2014
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.
Home