[PATCH] arm64: mm: fix runtime fallback to 48-bt VA when 52-bit VA is enabled

[Will Deacon]

On Tue, Mar 09, 2021 at 07:32:16PM -0500, Mark Salter wrote:
> I ran into an early boot soft lockup on a Qualcomm Amberwing using a v5.11
> kernel configured for 52-bit VA. This turned into a panic with a v5.12-rc2
> kernel.
> 
> The problem is that when we fall back to 48-bit VA, idmap_t0sz is not
> updated. Later, the kvm hypervisor uses idmap_t0sz to set its tcr_el2 and
> hangs (v5.11). After commit 1401bef703a4 ("arm64: mm: Always update TCR_EL1
> from __cpu_set_tcr_t0sz()"), the kernel panics when trying to use the idmap
> to call idmap_cpu_replace_ttbr1().
> 
> Oddly, other systems (thunderX2 and Ampere eMag) which don't support 52-bit
> VA seem to handle the setting of an unsupported t0sz without any apparent
> problems. Indeed, if one reads back the tcr written with t0sz==12, the
> value read has t0sz==16. Not so with Amberwing.

Nice, you have one of those elusive platforms!

> Fixes: 90ec95cda91a ("arm64: mm: Introduce VA_BITS_MIN")
> Signed-off-by: Mark Salter <msalter at redhat.com>
> ---
>  arch/arm64/kernel/head.S | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/kernel/head.S b/arch/arm64/kernel/head.S
> index 66b0e0b66e31..2bcbbb26292e 100644
> --- a/arch/arm64/kernel/head.S
> +++ b/arch/arm64/kernel/head.S
> @@ -291,6 +291,7 @@ SYM_FUNC_START_LOCAL(__create_page_tables)
>  	 */
>  	adrp	x0, idmap_pg_dir
>  	adrp	x3, __idmap_text_start		// __pa(__idmap_text_start)
> +	mov	x4, TCR_T0SZ(VA_BITS)
>  
>  #ifdef CONFIG_ARM64_VA_BITS_52
>  	mrs_s	x6, SYS_ID_AA64MMFR2_EL1
> @@ -299,6 +300,13 @@ SYM_FUNC_START_LOCAL(__create_page_tables)
>  	cbnz	x6, 1f
>  #endif
>  	mov	x5, #VA_BITS_MIN
> +#ifdef CONFIG_ARM64_VA_BITS_52
> +	mov	x4, TCR_T0SZ(VA_BITS_MIN)
> +	adr_l	x6, idmap_t0sz
> +	str	x4, [x6]
> +	dmb	sy
> +	dc	ivac, x6		// Invalidate potentially stale cache line
> +#endif
>  1:
>  	adr_l	x6, vabits_actual
>  	str	x5, [x6]
> @@ -319,7 +327,7 @@ SYM_FUNC_START_LOCAL(__create_page_tables)
>  	 */
>  	adrp	x5, __idmap_text_end
>  	clz	x5, x5
> -	cmp	x5, TCR_T0SZ(VA_BITS)	// default T0SZ small enough?
> +	cmp	x5, x4			// default T0SZ small enough?
>  	b.ge	1f			// .. then skip VA range extension

Could we instead have the default value be 48-bit, and then avoid having
to update the variable in both cases? e.g. something along the lines of
the entirely untested diff below?

Cheers,

Will

--->8

diff --git a/arch/arm64/kernel/head.S b/arch/arm64/kernel/head.S
index 66b0e0b66e31..fb795123896f 100644
--- a/arch/arm64/kernel/head.S
+++ b/arch/arm64/kernel/head.S
@@ -319,7 +319,7 @@ SYM_FUNC_START_LOCAL(__create_page_tables)
         */
        adrp    x5, __idmap_text_end
        clz     x5, x5
-       cmp     x5, TCR_T0SZ(VA_BITS)   // default T0SZ small enough?
+       cmp     x5, TCR_T0SZ(VA_BITS_MIN)       // default T0SZ small enough?
        b.ge    1f                      // .. then skip VA range extension
 
        adr_l   x6, idmap_t0sz
diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
index 3802cfbdd20d..4c5603c41870 100644
--- a/arch/arm64/mm/mmu.c
+++ b/arch/arm64/mm/mmu.c
@@ -40,7 +40,7 @@
 #define NO_BLOCK_MAPPINGS      BIT(0)
 #define NO_CONT_MAPPINGS       BIT(1)
 
-u64 idmap_t0sz = TCR_T0SZ(VA_BITS);
+u64 idmap_t0sz = TCR_T0SZ(VA_BITS_MIN);
 u64 idmap_ptrs_per_pgd = PTRS_PER_PGD;
 
 u64 __section(".mmuoff.data.write") vabits_actual;