Date:   Sun, 15 Sep 2019 19:34:17 +0100
From:   Russell King - ARM Linux admin <linux@...linux.org.uk>
To:     "Eric W. Biederman" <ebiederm@...ssion.com>
Cc:     kstewart@...uxfoundation.org, gustavo@...eddedor.com,
        gregkh@...uxfoundation.org, linux-kernel@...r.kernel.org,
        Jing Xiangfeng <jingxiangfeng@...wei.com>, linux-mm@...ck.org,
        sakari.ailus@...ux.intel.com, bhelgaas@...gle.com,
        tglx@...utronix.de, linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH] arm: fix page faults in do_alignment

On Fri, Sep 06, 2019 at 04:17:59PM +0100, Russell King - ARM Linux admin wrote:
> On Mon, Sep 02, 2019 at 12:36:56PM -0500, Eric W. Biederman wrote:
> > Russell King - ARM Linux admin <linux@...linux.org.uk> writes:
> > 
> > > On Fri, Aug 30, 2019 at 04:02:48PM -0500, Eric W. Biederman wrote:
> > >> Russell King - ARM Linux admin <linux@...linux.org.uk> writes:
> > >> 
> > >> > On Fri, Aug 30, 2019 at 02:45:36PM -0500, Eric W. Biederman wrote:
> > >> >> Russell King - ARM Linux admin <linux@...linux.org.uk> writes:
> > >> >> 
> > >> >> > On Fri, Aug 30, 2019 at 09:31:17PM +0800, Jing Xiangfeng wrote:
> > >> >> >> The function do_alignment can handle misaligned address for user and
> > >> >> >> kernel space. If it is a userspace access, do_alignment may fail on
> > >> >> >> a low-memory situation, because page faults are disabled in
> > >> >> >> probe_kernel_address.
> > >> >> >> 
> > >> >> >> Fix this by using __copy_from_user stead of probe_kernel_address.
> > >> >> >> 
> > >> >> >> Fixes: b255188 ("ARM: fix scheduling while atomic warning in alignment handling code")
> > >> >> >> Signed-off-by: Jing Xiangfeng <jingxiangfeng@...wei.com>
> > >> >> >
> > >> >> > NAK.
> > >> >> >
> > >> >> > The "scheduling while atomic warning in alignment handling code" is
> > >> >> > caused by fixing up the page fault while trying to handle the
> > >> >> > mis-alignment fault generated from an instruction in atomic context.
> > >> >> >
> > >> >> > Your patch re-introduces that bug.
> > >> >> 
> > >> >> And the patch that fixed scheduling while atomic apparently introduced a
> > >> >> regression.  Admittedly a regression that took 6 years to track down but
> > >> >> still.
> > >> >
> > >> > Right, and given the number of years, we are trading one regression for
> > >> > a different regression.  If we revert to the original code where we
> > >> > fix up, we will end up with people complaining about a "new" regression
> > >> > caused by reverting the previous fix.  Follow this policy and we just
> > >> > end up constantly reverting the previous revert.
> > >> >
> > >> > The window is very small - the page in question will have had to have
> > >> > instructions read from it immediately prior to the handler being entered,
> > >> > and would have had to be made "old" before subsequently being unmapped.
> > >> 
> > >> > Rather than excessively complicating the code and making it even more
> > >> > inefficient (as in your patch), we could instead retry executing the
> > >> > instruction when we discover that the page is unavailable, which should
> > >> > cause the page to be paged back in.
> > >> 
> > >> My patch does not introduce any inefficiencies.  It onlys moves the
> > >> check for user_mode up a bit.  My patch did duplicate the code.
> > >> 
> > >> > If the page really is unavailable, the prefetch abort should cause a
> > >> > SEGV to be raised, otherwise the re-execution should replace the page.
> > >> >
> > >> > The danger to that approach is we page it back in, and it gets paged
> > >> > back out before we're able to read the instruction indefinitely.
> > >> 
> > >> I would think either a little code duplication or a function that looks
> > >> at user_mode(regs) and picks the appropriate kind of copy to do would be
> > >> the best way to go.  Because what needs to happen in the two cases for
> > >> reading the instruction are almost completely different.
> > >
> > > That is what I mean.  I'd prefer to avoid that with the large chunk of
> > > code.  How about instead adding a local replacement for
> > > probe_kernel_address() that just sorts out the reading, rather than
> > > duplicating all the code to deal with thumb fixup.
> > 
> > So something like this should be fine?
> > 
> > Jing Xiangfeng can you test this please?  I think this fixes your issue
> > but I don't currently have an arm development box where I could test this.
> 
> Sorry, only just got around to this again.  What I came up with is this:

I've heard nothing, so I've done nothing...

> 8<===
> From: Russell King <rmk+kernel@...linux.org.uk>
> Subject: [PATCH] ARM: mm: fix alignment
> 
> Signed-off-by: Russell King <rmk+kernel@...linux.org.uk>
> ---
>  arch/arm/mm/alignment.c | 44 ++++++++++++++++++++++++++++++++++++--------
>  1 file changed, 36 insertions(+), 8 deletions(-)
> 
> diff --git a/arch/arm/mm/alignment.c b/arch/arm/mm/alignment.c
> index 6067fa4de22b..529f54d94709 100644
> --- a/arch/arm/mm/alignment.c
> +++ b/arch/arm/mm/alignment.c
> @@ -765,6 +765,36 @@ do_alignment_t32_to_handler(unsigned long *pinstr, struct pt_regs *regs,
>  	return NULL;
>  }
>  
> +static int alignment_get_arm(struct pt_regs *regs, u32 *ip, unsigned long *inst)
> +{
> +	u32 instr = 0;
> +	int fault;
> +
> +	if (user_mode(regs))
> +		fault = get_user(instr, ip);
> +	else
> +		fault = probe_kernel_address(ip, instr);
> +
> +	*inst = __mem_to_opcode_arm(instr);
> +
> +	return fault;
> +}
> +
> +static int alignment_get_thumb(struct pt_regs *regs, u16 *ip, u16 *inst)
> +{
> +	u16 instr = 0;
> +	int fault;
> +
> +	if (user_mode(regs))
> +		fault = get_user(instr, ip);
> +	else
> +		fault = probe_kernel_address(ip, instr);
> +
> +	*inst = __mem_to_opcode_thumb16(instr);
> +
> +	return fault;
> +}
> +
>  static int
>  do_alignment(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
>  {
> @@ -772,10 +802,10 @@ do_alignment(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
>  	unsigned long instr = 0, instrptr;
>  	int (*handler)(unsigned long addr, unsigned long instr, struct pt_regs *regs);
>  	unsigned int type;
> -	unsigned int fault;
>  	u16 tinstr = 0;
>  	int isize = 4;
>  	int thumb2_32b = 0;
> +	int fault;
>  
>  	if (interrupts_enabled(regs))
>  		local_irq_enable();
> @@ -784,15 +814,14 @@ do_alignment(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
>  
>  	if (thumb_mode(regs)) {
>  		u16 *ptr = (u16 *)(instrptr & ~1);
> -		fault = probe_kernel_address(ptr, tinstr);
> -		tinstr = __mem_to_opcode_thumb16(tinstr);
> +
> +		fault = alignment_get_thumb(regs, ptr, &tinstr);
>  		if (!fault) {
>  			if (cpu_architecture() >= CPU_ARCH_ARMv7 &&
>  			    IS_T32(tinstr)) {
>  				/* Thumb-2 32-bit */
> -				u16 tinst2 = 0;
> -				fault = probe_kernel_address(ptr + 1, tinst2);
> -				tinst2 = __mem_to_opcode_thumb16(tinst2);
> +				u16 tinst2;
> +				fault = alignment_get_thumb(regs, ptr + 1, &tinst2);
>  				instr = __opcode_thumb32_compose(tinstr, tinst2);
>  				thumb2_32b = 1;
>  			} else {
> @@ -801,8 +830,7 @@ do_alignment(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
>  			}
>  		}
>  	} else {
> -		fault = probe_kernel_address((void *)instrptr, instr);
> -		instr = __mem_to_opcode_arm(instr);
> +		fault = alignment_get_arm(regs, (void *)instrptr, &instr);
>  	}
>  
>  	if (fault) {
> -- 
> 2.7.4
> 
> -- 
> RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
> FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
> According to speedtest.net: 11.9Mbps down 500kbps up
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@...ts.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> 

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up

Powered by blists - more mailing lists