Re: [PATCH] powerpc/fadump: fix integer overflow in MIN_RMA size check

["Christophe Leroy (CS GROUP)" <chleroy@kernel.org>]

Le 20/05/2026 à 07:53, Sayali Patil a écrit :
> The MIN_RMA size checks in fadump_setup_param_area() use
> (MIN_RMA * 1024 * 1024), which is evaluated in int and can
> overflow when MIN_RMA is 2048 or larger. This triggers compiler
> warnings such as:
> 
> warning: integer overflow in expression of type 'int'
> results in '0' [-Woverflow]
> 
> Promote MIN_RMA to u64 before the multiplication so the expression
> is evaluated in 64-bit and matches the surrounding physical address
> and memory size calculations.
> 
> This fixes both the comparison against ppc64_rma_size and the
> assignment to range_start.

How do you create that problem ? MIN_RMA has a fixed value of 768 which 
is defined in asm/prom.h

> 
> Fixes: b7bb46062457 ("powerpc/fadump: fix additional param memory reservation for HASH MMU")
> Signed-off-by: Sayali Patil <sayalip@linux.ibm.com>
> ---
>   arch/powerpc/kernel/fadump.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/kernel/fadump.c b/arch/powerpc/kernel/fadump.c
> index 501d43bf18f3..dea7f7105e42 100644
> --- a/arch/powerpc/kernel/fadump.c
> +++ b/arch/powerpc/kernel/fadump.c
> @@ -1759,10 +1759,10 @@ void __init fadump_setup_param_area(void)
>   		 * 2. The range should be between MIN_RMA and RMA size (ppc64_rma_size)
>   		 * 3. It must not overlap with the fadump reserved area.
>   		 */
> -		if (ppc64_rma_size < MIN_RMA*1024*1024)
> +		if (ppc64_rma_size < (u64)MIN_RMA * 1024 * 1024)

As you are modifying that line, please use SZ_1M instead of keeping 
opencoded 1024 * 1024.

>   			return;
>   
> -		range_start = MIN_RMA * 1024 * 1024;
> +		range_start = (u64)MIN_RMA * 1024 * 1024;

Same.

>   		range_end = min(ppc64_rma_size, fw_dump.boot_mem_top);
>   	}
>   

Maybe the best would be to define MIN_RMA as a bytes value in asm/prom.h 
and to divide the value by SZ_1M in kernel/prom_init.c when initialising 
ibm_architecture_vec_template.

That way you could just define it as SZ_2G instead of the problematic 
2048 you mention.