linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] ARM: mvebu: drop pointless check for coherency_base
@ 2020-09-16 10:56 Ard Biesheuvel
  2020-09-24  8:06 ` Gregory CLEMENT
  0 siblings, 1 reply; 2+ messages in thread
From: Ard Biesheuvel @ 2020-09-16 10:56 UTC (permalink / raw)
  To: linux-arm-kernel
  Cc: Andrew Lunn, Jason Cooper, Gregory CLEMENT, Thomas Petazzoni,
	Ard Biesheuvel, Sebastian Hesselbarth

The MMU off code path in ll_get_coherency_base() attempts to decide
whether the coherency fabric is mapped by testing the value of
coherency_base, which carries its virtual address if its mapped, and
0x0 otherwise.

However, what the code actually does is take the virtual address of
the coherency_base symbol, and compare it with 0x0, which are never
equal, and so the branch is never taken. In fact, with the MMU off,
dereferencing the VA of coherency_base is not possible to begin with,
nor can its value be relied upon with the MMU off since it is not
cleaned to the Dcache as is done with coherency_phys_base in
armada_370_coherency_init().

Instead, the value of coherency_phys_base is returned, which results
in the correct behavior since it will be 0x0 as well if the coherency
fabric is not mapped, and it is accessible with the MMU off. So just
drop the comparison and the branch.

Fixes: 30cdef97107370a7 ("ARM: mvebu: make the coherency_ll.S functions work with no coherency fabric")
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Gregory CLEMENT <gregory.clement@bootlin.com>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
Found by accident. Build tested only.

 arch/arm/mach-mvebu/coherency_ll.S | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/arch/arm/mach-mvebu/coherency_ll.S b/arch/arm/mach-mvebu/coherency_ll.S
index 2d962fe48821..a3a64bf97250 100644
--- a/arch/arm/mach-mvebu/coherency_ll.S
+++ b/arch/arm/mach-mvebu/coherency_ll.S
@@ -35,13 +35,8 @@ ENTRY(ll_get_coherency_base)
 
 	/*
 	 * MMU is disabled, use the physical address of the coherency
-	 * base address. However, if the coherency fabric isn't mapped
-	 * (i.e its virtual address is zero), it means coherency is
-	 * not enabled, so we return 0.
+	 * base address, (or 0x0 if the coherency fabric is not mapped)
 	 */
-	ldr	r1, =coherency_base
-	cmp	r1, #0
-	beq	2f
 	adr	r1, 3f
 	ldr	r3, [r1]
 	ldr	r1, [r1, r3]
-- 
2.17.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] ARM: mvebu: drop pointless check for coherency_base
  2020-09-16 10:56 [PATCH] ARM: mvebu: drop pointless check for coherency_base Ard Biesheuvel
@ 2020-09-24  8:06 ` Gregory CLEMENT
  0 siblings, 0 replies; 2+ messages in thread
From: Gregory CLEMENT @ 2020-09-24  8:06 UTC (permalink / raw)
  To: Ard Biesheuvel, linux-arm-kernel
  Cc: Andrew Lunn, Sebastian Hesselbarth, Ard Biesheuvel,
	Thomas Petazzoni, Jason Cooper

Hi Ard,

> The MMU off code path in ll_get_coherency_base() attempts to decide
> whether the coherency fabric is mapped by testing the value of
> coherency_base, which carries its virtual address if its mapped, and
> 0x0 otherwise.
>
> However, what the code actually does is take the virtual address of
> the coherency_base symbol, and compare it with 0x0, which are never
> equal, and so the branch is never taken. In fact, with the MMU off,
> dereferencing the VA of coherency_base is not possible to begin with,
> nor can its value be relied upon with the MMU off since it is not
> cleaned to the Dcache as is done with coherency_phys_base in
> armada_370_coherency_init().
>
> Instead, the value of coherency_phys_base is returned, which results
> in the correct behavior since it will be 0x0 as well if the coherency
> fabric is not mapped, and it is accessible with the MMU off. So just
> drop the comparison and the branch.

Thanks for this patch. I needed to refresh my memories about this
subject and I agree with your explanation. I applied it on mvebu/fixes.

Gregory

>
> Fixes: 30cdef97107370a7 ("ARM: mvebu: make the coherency_ll.S functions work with no coherency fabric")
> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
> Cc: Jason Cooper <jason@lakedaemon.net>
> Cc: Andrew Lunn <andrew@lunn.ch>
> Cc: Gregory CLEMENT <gregory.clement@bootlin.com>
> Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> ---
> Found by accident. Build tested only.
>
>  arch/arm/mach-mvebu/coherency_ll.S | 7 +------
>  1 file changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/arch/arm/mach-mvebu/coherency_ll.S b/arch/arm/mach-mvebu/coherency_ll.S
> index 2d962fe48821..a3a64bf97250 100644
> --- a/arch/arm/mach-mvebu/coherency_ll.S
> +++ b/arch/arm/mach-mvebu/coherency_ll.S
> @@ -35,13 +35,8 @@ ENTRY(ll_get_coherency_base)
>  
>  	/*
>  	 * MMU is disabled, use the physical address of the coherency
> -	 * base address. However, if the coherency fabric isn't mapped
> -	 * (i.e its virtual address is zero), it means coherency is
> -	 * not enabled, so we return 0.
> +	 * base address, (or 0x0 if the coherency fabric is not mapped)
>  	 */
> -	ldr	r1, =coherency_base
> -	cmp	r1, #0
> -	beq	2f
>  	adr	r1, 3f
>  	ldr	r3, [r1]
>  	ldr	r1, [r1, r3]
> -- 
> 2.17.1
>

-- 
Gregory Clement, Bootlin
Embedded Linux and Kernel engineering
http://bootlin.com

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2020-09-24  8:08 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-09-16 10:56 [PATCH] ARM: mvebu: drop pointless check for coherency_base Ard Biesheuvel
2020-09-24  8:06 ` Gregory CLEMENT

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).