Re: [PATCH] Fix display of Maximum Memory

[Christophe Leroy]

Le 14/01/2020 Ã 22:07, Michael Bringmann a ÃcritÂ:
> Correct overflow problem in calculation+display of Maximum Memory
> value to syscfg where 32bits is insufficient.
>
> Signed-off-by: Michael Bringmann <mwb@xxxxxxxxxxxxx>
> ---
>   arch/powerpc/platforms/pseries/lparcfg.c | 8 ++++----
>   1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/arch/powerpc/platforms/pseries/lparcfg.c b/arch/powerpc/platforms/pseries/lparcfg.c
> index 4ee2594..183aeb7 100644
> --- a/arch/powerpc/platforms/pseries/lparcfg.c
> +++ b/arch/powerpc/platforms/pseries/lparcfg.c
> @@ -435,12 +435,12 @@ static void parse_em_data(struct seq_file *m)
>
>   static void maxmem_data(struct seq_file *m)
>   {
> -       unsigned long maxmem = 0;
> +       unsigned long long maxmem = 0;

What about using u64 instead, for readability ?

> -       maxmem += drmem_info->n_lmbs * drmem_info->lmb_size;
> -       maxmem += hugetlb_total_pages() * PAGE_SIZE;
> +       maxmem += (unsigned long long)drmem_info->n_lmbs * (unsigned long long)drmem_info->lmb_size;

This line is likely too long. You only need to cast one of the two 
operants to force a 64 bits multiply. And using u64 would shorten the line.

Can both multiplications overflow ?

Christophe

> +       maxmem += (unsigned long long)hugetlb_total_pages() * (unsigned long long)PAGE_SIZE;
>
> -       seq_printf(m, "MaxMem=%ld\n", maxmem);
> +       seq_printf(m, "MaxMem=%llu\n", maxmem);
>   }
>
>   static int pseries_lparcfg_data(struct seq_file *m, void *v)