Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1345096 > unrolled thread

[PATCH tip/x86/asm] uprobes: __create_xol_area() must nullify xol_mapping.fault

Back to article view | Back to linux.kernel

Contents

  [PATCH tip/x86/asm] uprobes: __create_xol_area() must nullify  xol_mapping.fault Oleg Nesterov <oleg@redhat.com> - 2016-02-27 23:20 +0100
    [tip:x86/asm] uprobes: __create_xol_area() must nullify  xol_mapping.fault tip-bot for Oleg Nesterov <tipbot@zytor.com> - 2016-02-29 12:10 +0100
    Re: [PATCH tip/x86/asm] uprobes: __create_xol_area() must nullify  xol_mapping.fault Peter Zijlstra <peterz@infradead.org> - 2016-02-29 12:50 +0100
      Re: [PATCH tip/x86/asm] uprobes: __create_xol_area() must nullify  xol_mapping.fault Oleg Nesterov <oleg@redhat.com> - 2016-02-29 17:00 +0100
        Re: [PATCH tip/x86/asm] uprobes: __create_xol_area() must nullify  xol_mapping.fault Peter Zijlstra <peterz@infradead.org> - 2016-02-29 17:20 +0100

#1345096 — [PATCH tip/x86/asm] uprobes: __create_xol_area() must nullify xol_mapping.fault

As Jiri pointed out, the recent commit f872f5400cc0 "mm: Add a
vm_special_mapping.fault() method" breaks uprobes. __create_xol_area()
doesn't initialize the new ->fault() method and this obviously leads to
kernel crash when the application tries to execute the probed insn after
bp hit.

We probably want to add uprobes_special_mapping_fault(), this allows to
turn xol_area->xol_mapping into a single instance of vm_special_mapping.
But we need a simple fix, so lets change __create_xol() to nullify the
new member as Jiri suggests.

Reported-by: Jiri Olsa <jolsa@redhat.com>
Suggested-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---
 kernel/events/uprobes.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 0167679..5f6ce93 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -1178,6 +1178,7 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
 		goto free_area;
 
 	area->xol_mapping.name = "[uprobes]";
+	area->xol_mapping.fault = NULL;
 	area->xol_mapping.pages = area->pages;
 	area->pages[0] = alloc_page(GFP_HIGHUSER);
 	if (!area->pages[0])
-- 
2.5.0

[toc] | [next] | [standalone]

#1345668 — [tip:x86/asm] uprobes: __create_xol_area() must nullify xol_mapping.fault

Commit-ID:  869ae76147ffdf21ad24f0e599303cd58a2bb39f
Gitweb:     http://git.kernel.org/tip/869ae76147ffdf21ad24f0e599303cd58a2bb39f
Author:     Oleg Nesterov <oleg@redhat.com>
AuthorDate: Sat, 27 Feb 2016 23:11:28 +0100
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 29 Feb 2016 08:47:51 +0100

uprobes: __create_xol_area() must nullify xol_mapping.fault

As Jiri pointed out, this recent commit:

 f872f5400cc0 ("mm: Add a vm_special_mapping.fault() method")

breaks uprobes: __create_xol_area() doesn't initialize the new ->fault()
method and this obviously leads to kernel crash when the application
tries to execute the probed insn after bp hit.

We probably want to add uprobes_special_mapping_fault(), this allows to
turn xol_area->xol_mapping into a single instance of vm_special_mapping.
But we need a simple fix, so lets change __create_xol() to nullify the
new member as Jiri suggests.

Suggested-by: Jiri Olsa <jolsa@redhat.com>
Reported-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andy Lutomirski <tipbot@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Pratyush Anand <panand@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20160227221128.GA29565@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/events/uprobes.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 0167679..5f6ce93 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -1178,6 +1178,7 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
 		goto free_area;
 
 	area->xol_mapping.name = "[uprobes]";
+	area->xol_mapping.fault = NULL;
 	area->xol_mapping.pages = area->pages;
 	area->pages[0] = alloc_page(GFP_HIGHUSER);
 	if (!area->pages[0])

[toc] | [prev] | [next] | [standalone]

#1345730

On Sat, Feb 27, 2016 at 11:11:28PM +0100, Oleg Nesterov wrote:
> As Jiri pointed out, the recent commit f872f5400cc0 "mm: Add a
> vm_special_mapping.fault() method" breaks uprobes. __create_xol_area()
> doesn't initialize the new ->fault() method and this obviously leads to
> kernel crash when the application tries to execute the probed insn after
> bp hit.
> 
> We probably want to add uprobes_special_mapping_fault(), this allows to
> turn xol_area->xol_mapping into a single instance of vm_special_mapping.
> But we need a simple fix, so lets change __create_xol() to nullify the
> new member as Jiri suggests.
> 
> Reported-by: Jiri Olsa <jolsa@redhat.com>
> Suggested-by: Jiri Olsa <jolsa@redhat.com>
> Signed-off-by: Oleg Nesterov <oleg@redhat.com>
> ---
>  kernel/events/uprobes.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
> index 0167679..5f6ce93 100644
> --- a/kernel/events/uprobes.c
> +++ b/kernel/events/uprobes.c
> @@ -1178,6 +1178,7 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
>  		goto free_area;
>  
>  	area->xol_mapping.name = "[uprobes]";
> +	area->xol_mapping.fault = NULL;
>  	area->xol_mapping.pages = area->pages;

Would not something like:

	area->xol_mapping = (struct vm_special_mapping){
		.name = "[uprobes]",
		.pages = area->pages,
	};

Be a more robust approach? That way, if someone adds more fields, they
at least get initialized (to 0).

>  	area->pages[0] = alloc_page(GFP_HIGHUSER);
>  	if (!area->pages[0])

[toc] | [prev] | [next] | [standalone]

#1345933

On 02/29, Peter Zijlstra wrote:
>
> On Sat, Feb 27, 2016 at 11:11:28PM +0100, Oleg Nesterov wrote:
> > --- a/kernel/events/uprobes.c
> > +++ b/kernel/events/uprobes.c
> > @@ -1178,6 +1178,7 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
> >  		goto free_area;
> >
> >  	area->xol_mapping.name = "[uprobes]";
> > +	area->xol_mapping.fault = NULL;
> >  	area->xol_mapping.pages = area->pages;
>
> Would not something like:
>
> 	area->xol_mapping = (struct vm_special_mapping){
> 		.name = "[uprobes]",
> 		.pages = area->pages,
> 	};
>
> Be a more robust approach? That way, if someone adds more fields, they
> at least get initialized (to 0).

OK, agreed...

Do you want me to send v2? Or incremental patch because this one is already in
-tip tree.

Or do nothing unless you feel strongly about it. area->xol_mapping should go away,
but we need a simple preparation in mm/mmap.c.

Oleg.

[toc] | [prev] | [next] | [standalone]

#1345947

On Mon, Feb 29, 2016 at 04:51:14PM +0100, Oleg Nesterov wrote:
> On 02/29, Peter Zijlstra wrote:
> >
> > On Sat, Feb 27, 2016 at 11:11:28PM +0100, Oleg Nesterov wrote:
> > > --- a/kernel/events/uprobes.c
> > > +++ b/kernel/events/uprobes.c
> > > @@ -1178,6 +1178,7 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
> > >  		goto free_area;
> > >
> > >  	area->xol_mapping.name = "[uprobes]";
> > > +	area->xol_mapping.fault = NULL;
> > >  	area->xol_mapping.pages = area->pages;
> >
> > Would not something like:
> >
> > 	area->xol_mapping = (struct vm_special_mapping){
> > 		.name = "[uprobes]",
> > 		.pages = area->pages,
> > 	};
> >
> > Be a more robust approach? That way, if someone adds more fields, they
> > at least get initialized (to 0).
> 
> OK, agreed...
> 
> Do you want me to send v2? Or incremental patch because this one is already in
> -tip tree.
> 
> Or do nothing unless you feel strongly about it. area->xol_mapping should go away,
> but we need a simple preparation in mm/mmap.c.

Yeah, no real strong feelings. Esp. if you're working on getting it
killed.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web