[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <mhng-46a9363c-895f-4cc2-a8e3-3a48f2fee418@palmer-ri-x1c9a>
Date: Thu, 01 Dec 2022 11:24:49 -0800 (PST)
From: Palmer Dabbelt <palmer@...belt.com>
To: alexghiti@...osinc.com
CC: Paul Walmsley <paul.walmsley@...ive.com>, aou@...s.berkeley.edu,
panqinglin2020@...as.ac.cn, linux-mm@...ck.org,
linux-riscv@...ts.infradead.org, linux-kernel@...r.kernel.org,
alexghiti@...osinc.com
Subject: Re: [PATCH 1/1] riscv: Fix P4D_SHIFT definition for 3-level page table mode
On Thu, 01 Dec 2022 05:51:28 PST (-0800), alexghiti@...osinc.com wrote:
> RISC-V kernels support 3,4,5-level page tables at runtime by folding
> upper levels.
>
> In case of a 3-level page table, PGDIR is folded into P4D which in turn
> is folded into PUD: PGDIR_SHIFT value is correctly set to the same value
> as PUD_SHIFT, but P4D_SHIFT is not, then any use of P4D_SHIFT will access
> invalid address bits (all set to 1).
>
> Fix this by dynamically defining P4D_SHIFT value, like we already do for
> PGDIR_SHIFT.
>
> Fixes: d10efa21a937 ("riscv: mm: Control p4d's folding by pgtable_l5_enabled")
> Signed-off-by: Alexandre Ghiti <alexghiti@...osinc.com>
> ---
> arch/riscv/include/asm/pgtable-64.h | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/arch/riscv/include/asm/pgtable-64.h b/arch/riscv/include/asm/pgtable-64.h
> index dc42375c2357..42a042c0e13e 100644
> --- a/arch/riscv/include/asm/pgtable-64.h
> +++ b/arch/riscv/include/asm/pgtable-64.h
> @@ -25,7 +25,11 @@ extern bool pgtable_l5_enabled;
> #define PGDIR_MASK (~(PGDIR_SIZE - 1))
>
> /* p4d is folded into pgd in case of 4-level page table */
> -#define P4D_SHIFT 39
> +#define P4D_SHIFT_L3 30
> +#define P4D_SHIFT_L4 39
> +#define P4D_SHIFT_L5 39
> +#define P4D_SHIFT (pgtable_l5_enabled ? P4D_SHIFT_L5 : \
> + (pgtable_l4_enabled ? P4D_SHIFT_L4 : P4D_SHIFT_L3))
> #define P4D_SIZE (_AC(1, UL) << P4D_SHIFT)
> #define P4D_MASK (~(P4D_SIZE - 1))
Reviewed-by: Palmer Dabbelt <palmer@...osinc.com>
Too late for this week, but if someone's got a concrete regression then
I'm happy to take it next week. Otherwise it'll end up on for-next,
it'll be backported anyway but this way it'll have a touch more time.
Thanks!
Powered by blists - more mailing lists