[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1458762438.23434.120.camel@redhat.com>
Date: Wed, 23 Mar 2016 15:47:18 -0400
From: Mark Salter <msalter@...hat.com>
To: Catalin Marinas <catalin.marinas@....com>,
Will Deacon <will.deacon@....com>
Cc: Mark Langsdorf <mlangsdo@...hat.com>, linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org,
Ard Biesheuvel <ard.biesheuvel@...aro.org>
Subject: Re: [PATCH] arm64: handle unmapped pages in initrd relocation
On Mon, 2016-02-01 at 19:30 -0500, Mark Salter wrote:
> Commit 4dffbfc48d65 ("arm64/efi: mark UEFI reserved regions as
> MEMBLOCK_NOMAP") causes a potential problem in arm64 initrd relocation
> code. If the kernel uses a pagesize greater than the 4k pagesize used
> by UEFI, pagesize rounding may lead to one or both ends of the initrd
> image to be marked unmapped. This leads to a panic when the kernel goes
> to unpack it. This patch looks for unmapped pages at beginning and end
> of the initrd image and if seen, relocated the initrd to a new area
> completely covered by the kernel linear map.
>
> Signed-off-by: Mark Salter <msalter@...hat.com>
> Cc: Ard Biesheuvel <ard.biesheuvel@...aro.org>
> ---
The Fedora folks have run into this problem with a certain kernel build. What ever
happened to Ard's suggested fix. The MEMBLOCK_NOMAP patch caused a regression which
should be fixed. Whether this patch, Ard's patch, or something else.
https://bugzilla.redhat.com/show_bug.cgi?id=1309147
> arch/arm64/kernel/setup.c | 46 +++++++++++++++++++++++++++++++++++++---------
> 1 file changed, 37 insertions(+), 9 deletions(-)
>
> diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c
> index d22c5fc..849566e 100644
> --- a/arch/arm64/kernel/setup.c
> +++ b/arch/arm64/kernel/setup.c
> @@ -235,24 +235,29 @@ static void __init relocate_initrd(void)
> phys_addr_t ram_end = memblock_end_of_DRAM();
> phys_addr_t new_start;
> unsigned long size, to_free = 0;
> + unsigned long unmapped_start = 0, unmapped_end = 0;
> void *dest;
>
> - if (orig_end <= ram_end)
> + size = orig_end - orig_start;
> + if (!size)
> return;
>
> /*
> - * Any of the original initrd which overlaps the linear map should
> - * be freed after relocating.
> + * If kernel pagesize > 4K, pagesize rounding may have placed
> + * part of either end of initrd in an unmapped page.
> + *
> + * Find any unmapped bytes at start or end of initrd.
> */
> - if (orig_start < ram_end)
> - to_free = ram_end - orig_start;
> + if (!memblock_is_map_memory(orig_start))
> + unmapped_start = PAGE_SIZE - (orig_start & (PAGE_SIZE - 1));
> + if (!memblock_is_map_memory(orig_end - 1))
> + unmapped_end = ((orig_end - 1) & (PAGE_SIZE - 1)) + 1;
>
> - size = orig_end - orig_start;
> - if (!size)
> + if (unmapped_start == 0 && unmapped_end == 0 && orig_end <= ram_end)
> return;
>
> /* initrd needs to be relocated completely inside linear mapping */
> - new_start = memblock_find_in_range(0, PFN_PHYS(max_pfn),
> + new_start = memblock_find_in_range(0, ram_end,
> size, PAGE_SIZE);
> if (!new_start)
> panic("Cannot relocate initrd of size %ld\n", size);
> @@ -267,7 +272,30 @@ static void __init relocate_initrd(void)
>
> dest = (void *)initrd_start;
>
> - if (to_free) {
> + if (unmapped_end) {
> + copy_from_early_mem(dest + size - unmapped_end,
> + orig_start + size - unmapped_end,
> + unmapped_end);
> + size -= unmapped_end;
> + if (size == 0)
> + return;
> + }
> +
> + if (unmapped_start) {
> + copy_from_early_mem(dest, orig_start, unmapped_start);
> + dest += unmapped_start;
> + orig_start += unmapped_start;
> + size -= unmapped_start;
> + if (size == 0)
> + return;
> + }
> +
> + /*
> + * Any of the remaining original initrd which overlaps the linear map
> + * should be freed after relocating.
> + */
> + if (orig_start < ram_end) {
> + to_free = min(size, (unsigned long)(ram_end - orig_start));
> memcpy(dest, (void *)__phys_to_virt(orig_start), to_free);
> dest += to_free;
> }
Powered by blists - more mailing lists