linux-kernel - Re: [PATCH v5 2/2] sh: Restructure setup code to reserve memory regions earlier

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aba64912a7cf700fd281016d4a2ad82d@artur-rojek.eu>
Date: Sun, 29 Sep 2024 23:15:35 +0200
From: Artur Rojek <contact@...ur-rojek.eu>
To: Oreoluwa Babatunde <quic_obabatun@...cinc.com>
Cc: dalias@...c.org, glaubitz@...sik.fu-berlin.de,
 ysato@...rs.sourceforge.jp, kernel@...cinc.com,
 linux-kernel@...r.kernel.org, linux-sh@...r.kernel.org, robh+dt@...nel.org,
 Geert Uytterhoeven <geert+renesas@...der.be>
Subject: Re: [PATCH v5 2/2] sh: Restructure setup code to reserve memory
 regions earlier

On 2024-07-18 04:18, Oreoluwa Babatunde wrote:
> The unflatten_device_tree() function contains a call to
> memblock_alloc(). This is a problem because this allocation is done
> before any of the reserved memory regions are set aside in
> paging_init().
> As a result, there is a possibility for memblock to unknowingly 
> allocate
> from any of the memory regions that are meant to be reserved.
> 
> Hence, restructure the setup code to reserve the memory regions before
> any allocation is done by the unflatten_devicetree*() using memblock.
> 
> Signed-off-by: Oreoluwa Babatunde <quic_obabatun@...cinc.com>

Hi Oreoluwa,

> ---
>  arch/sh/mm/init.c | 15 +++++++--------
>  1 file changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c
> index 643e3617c6a6..857ce8cc84bd 100644
> --- a/arch/sh/mm/init.c
> +++ b/arch/sh/mm/init.c
> @@ -249,6 +249,7 @@ void __init early_reserve_mem(void)
>  	u32 zero_base = (u32)__MEMORY_START + (u32)PHYSICAL_OFFSET;
>  	u32 start = zero_base + (u32)CONFIG_ZERO_PAGE_OFFSET;
> 
> +	sh_mv.mv_mem_init();

One side effect of moving mv_mem_init here is that it makes cache
operations available earlier. But I see no harm in doing that.

Otherwise the patch is looking good. Verified on J2 Turtle Board.

Reviewed-by: Artur Rojek <contact@...ur-rojek.eu>

Cheers,
Artur

>  	/*
>  	 * Partially used pages are not usable - thus
>  	 * we are rounding upwards:
> @@ -274,14 +275,6 @@ void __init early_reserve_mem(void)
>  	 */
>  	check_for_initrd();
>  	reserve_crashkernel();
> -}
> -
> -void __init paging_init(void)
> -{
> -	unsigned long max_zone_pfns[MAX_NR_ZONES];
> -	unsigned long vaddr, end;
> -
> -	sh_mv.mv_mem_init();
> 
>  	/*
>  	 * Once the early reservations are out of the way, give the
> @@ -289,6 +282,12 @@ void __init paging_init(void)
>  	 */
>  	if (sh_mv.mv_mem_reserve)
>  		sh_mv.mv_mem_reserve();
> +}
> +
> +void __init paging_init(void)
> +{
> +	unsigned long max_zone_pfns[MAX_NR_ZONES];
> +	unsigned long vaddr, end;
> 
>  	memblock_enforce_memory_limit(memory_limit);
>  	memblock_allow_resize();