public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Frederik Deweerdt <deweerdt@free.fr>
To: Andrew Morton <akpm@osdl.org>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>, linux-kernel@vger.kernel.org
Subject: Re: [-mm patch] ptrace: make {put,get}reg work again for gs and fs
Date: Fri, 22 Dec 2006 06:06:18 +0000	[thread overview]
Message-ID: <20061222060618.GD18827@slug> (raw)
In-Reply-To: <20061221181108.6cede9ba.akpm@osdl.org>

On Thu, Dec 21, 2006 at 06:11:08PM -0800, Andrew Morton wrote:
> On Thu, 21 Dec 2006 18:00:49 -0800
> Jeremy Fitzhardinge <jeremy@goop.org> wrote:
> 
> > Frederik Deweerdt wrote:
> > > This is a -mm1 kernel + your efl_offset fix + the attached patch.
> > > So the problem came from putreg still saving %gs to the stack where
> > > there's no slot for it, whereas getreg got things right.
> > >   
> > 
> > That patch looks good, but I think it is already effectively in Andrew's
> > queue, because I noticed some problems in there when I reviewed  the
> > convert-to-%fs patch.
> > 
> 
> The below is what I have queued for urgent mainlining to address these
> problems.
> 
> Is it sufficient?
> 
No, it's not. The patch below fixes the place where we get eflags, this
triggered the "BUG while gdb'ing" reports.
The one I sent was to fix a problem that only I reported, AFAIK: when
you use gdb/ptrace to modify %fs, the value gets written in the wrong
place (see gdb sessions). So, unless you have another patch fixing the
way putreg() writes %fs, the patch[1] I sent should also be queued for
mainline.

Regards,
Frederik

[1] http://lkml.org/lkml/2006/12/21/267
> 
> 
> 
> From: Jeremy Fitzhardinge <jeremy@goop.org>
> 
> The PDA patches introduced a bug in ptrace: it reads eflags from the wrong
> place on the target's stack, but writes it back to the correct place.  The
> result is a corrupted eflags, which is most visible when it turns interrupts
> off unexpectedly.
> 
> This patch fixes this by making the ptrace code a little less fragile.  It
> changes [gs]et_stack_long to take a straightforward byte offset into struct
> pt_regs, rather than requiring all callers to do a sizeof(struct pt_regs)
> offset adjustment.  This means that the eflag's offset (EFL_OFFSET) on the
> target stack can be simply computed with offsetof().
> 
> Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
> Cc: Frederik Deweerdt <deweerdt@free.fr>
> Cc: Andi Kleen <ak@suse.de>
> Signed-off-by: Andrew Morton <akpm@osdl.org>
> ---
> 
>  arch/i386/kernel/ptrace.c |   21 ++++++++++-----------
>  1 file changed, 10 insertions(+), 11 deletions(-)
> 
> diff -puN arch/i386/kernel/ptrace.c~ptrace-fix-efl_offset-value-according-to-i386-pda-changes arch/i386/kernel/ptrace.c
> --- a/arch/i386/kernel/ptrace.c~ptrace-fix-efl_offset-value-according-to-i386-pda-changes
> +++ a/arch/i386/kernel/ptrace.c
> @@ -45,7 +45,7 @@
>  /*
>   * Offset of eflags on child stack..
>   */
> -#define EFL_OFFSET ((EFL-2)*4-sizeof(struct pt_regs))
> +#define EFL_OFFSET offsetof(struct pt_regs, eflags)
>  
>  static inline struct pt_regs *get_child_regs(struct task_struct *task)
>  {
> @@ -54,24 +54,24 @@ static inline struct pt_regs *get_child_
>  }
>  
>  /*
> - * this routine will get a word off of the processes privileged stack. 
> - * the offset is how far from the base addr as stored in the TSS.  
> - * this routine assumes that all the privileged stacks are in our
> + * This routine will get a word off of the processes privileged stack.
> + * the offset is bytes into the pt_regs structure on the stack.
> + * This routine assumes that all the privileged stacks are in our
>   * data space.
>   */   
>  static inline int get_stack_long(struct task_struct *task, int offset)
>  {
>  	unsigned char *stack;
>  
> -	stack = (unsigned char *)task->thread.esp0;
> +	stack = (unsigned char *)task->thread.esp0 - sizeof(struct pt_regs);
>  	stack += offset;
>  	return (*((int *)stack));
>  }
>  
>  /*
> - * this routine will put a word on the processes privileged stack. 
> - * the offset is how far from the base addr as stored in the TSS.  
> - * this routine assumes that all the privileged stacks are in our
> + * This routine will put a word on the processes privileged stack.
> + * the offset is bytes into the pt_regs structure on the stack.
> + * This routine assumes that all the privileged stacks are in our
>   * data space.
>   */
>  static inline int put_stack_long(struct task_struct *task, int offset,
> @@ -79,7 +79,7 @@ static inline int put_stack_long(struct 
>  {
>  	unsigned char * stack;
>  
> -	stack = (unsigned char *) task->thread.esp0;
> +	stack = (unsigned char *)task->thread.esp0 - sizeof(struct pt_regs);
>  	stack += offset;
>  	*(unsigned long *) stack = data;
>  	return 0;
> @@ -114,7 +114,7 @@ static int putreg(struct task_struct *ch
>  	}
>  	if (regno > ES*4)
>  		regno -= 1*4;
> -	put_stack_long(child, regno - sizeof(struct pt_regs), value);
> +	put_stack_long(child, regno, value);
>  	return 0;
>  }
>  
> @@ -137,7 +137,6 @@ static unsigned long getreg(struct task_
>  		default:
>  			if (regno > ES*4)
>  				regno -= 1*4;
> -			regno = regno - sizeof(struct pt_regs);
>  			retval &= get_stack_long(child, regno);
>  	}
>  	return retval;
> _
> 
> 

  parent reply	other threads:[~2006-12-22  6:08 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-12-15  6:59 2.6.20-rc1-mm1 Andrew Morton
2006-12-15  6:59 ` BUG: NMI Watchdog detected LOCKUP (was: 2.6.20-rc1-mm1) Tilman Schmidt
2006-12-19 18:52   ` Thomas Gleixner
2006-12-19 19:56   ` [patch] hrtimers: add state tracking, fix Ingo Molnar
2006-12-20  1:38     ` Tilman Schmidt
2006-12-20 20:05     ` Tilman Schmidt
2006-12-15 14:46 ` OOPS: deref 0x14 at pdc_port_start+0x82 [Was: 2.6.20-rc1-mm1] Jiri Slaby
2006-12-15 19:24   ` Andrew Morton
2006-12-15 22:49     ` Jiri Slaby
2006-12-15 20:39 ` 2.6.20-rc1-mm1 Damien Wyart
2006-12-15 21:01   ` 2.6.20-rc1-mm1 Andrew Morton
2006-12-17 11:07     ` 2.6.20-rc1-mm1 Damien Wyart
2006-12-18  8:03       ` 2.6.20-rc1-mm1 Laurent Riffard
2006-12-18 18:35         ` 2.6.20-rc1-mm1 Damien Wyart
2006-12-19 23:29           ` 2.6.20-rc1-mm1 Luben Tuikov
2006-12-18  7:44     ` 2.6.20-rc1-mm1 Jens Axboe
2006-12-15 23:26 ` WARNING (1) at .../arch/i386/mm/highmem.c:49 [Was: 2.6.20-rc1-mm1] Jiri Slaby
2006-12-16  0:16   ` Andrew Morton
2006-12-16  0:04 ` 2.6.20-rc1-mm1: unused sysrq_timer_list_show() Adrian Bunk
2006-12-16  7:56   ` [patch] debugging feature: SysRq-Q to print timers Ingo Molnar
2006-12-18 23:31     ` Andrew Morton
2006-12-18 23:45       ` Dave Jones
2006-12-19  0:00         ` Andrew Morton
2006-12-19 12:01           ` Ingo Molnar
2006-12-16 12:37 ` (Cross) compiling fails on first try (was Re: 2.6.20-rc1-mm1) Jan Dittmer
2006-12-16 13:56 ` [-mm patch] drivers/ide/pci/tc86c001.c: make a function static Adrian Bunk
2006-12-16 14:18   ` Alan
2006-12-17 18:08   ` Sergei Shtylyov
2006-12-17 20:52   ` Sergei Shtylyov
2006-12-16 13:56 ` [-mm patch] make uio_irq_handler() static Adrian Bunk
2006-12-20  6:09   ` Greg KH
2006-12-16 13:56 ` [-mm patch] drivers/video/{s3fb,svgalib}.c: possible cleanups Adrian Bunk
2006-12-16 17:39   ` Randy Dunlap
2006-12-16 18:16   ` Ondrej Zajicek
2006-12-16 13:57 ` [-mm patch] mm/vmscan.c: make a function static Adrian Bunk
2006-12-16 19:30 ` [-mm patch] noinitramfs cleanup Frederik Deweerdt
2006-12-18 13:38 ` [-mm patch] kill pxa2xx Kconfig warning Frederik Deweerdt
2006-12-18 20:06 ` 2.6.20-rc1-mm1 Bartlomiej Zolnierkiewicz
2006-12-19  0:29 ` 2.6.20-rc1-mm1 Randy Dunlap
2006-12-19  0:42   ` 2.6.20-rc1-mm1 Andrew Morton
2006-12-21 18:35 ` [-mm patch] ptrace: make {put,get}reg work again for gs and fs Frederik Deweerdt
2006-12-21 19:22   ` Jeremy Fitzhardinge
2006-12-21 20:53     ` Frederik Deweerdt
2006-12-21 21:59     ` Frederik Deweerdt
2006-12-22  2:00       ` Jeremy Fitzhardinge
     [not found]         ` <20061221181108.6cede9ba.akpm@osdl.org>
2006-12-22  6:06           ` Frederik Deweerdt [this message]
     [not found]             ` <20061221225414.de09c7df.akpm@osdl.org>
2006-12-22  7:00               ` Jeremy Fitzhardinge
2006-12-22  8:05               ` Frederik Deweerdt
2006-12-22  6:52           ` Jeremy Fitzhardinge
2006-12-22  6:55       ` Jeremy Fitzhardinge

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20061222060618.GD18827@slug \
    --to=deweerdt@free.fr \
    --cc=akpm@osdl.org \
    --cc=jeremy@goop.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox