[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAKv+Gu8jPdCWZrqHq2t7L9Yybg6WCVt48GoCeGRKbnjPnuRmYQ@mail.gmail.com>
Date: Wed, 30 Jan 2019 14:17:23 +0100
From: Ard Biesheuvel <ard.biesheuvel@...aro.org>
To: Qian Cai <cai@....pw>
Cc: linux-efi <linux-efi@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] efi/arm64: add a terminator for ptdump marker
On Tue, 29 Jan 2019 at 18:36, Qian Cai <cai@....pw> wrote:
>
> Read efi_page_tables debugfs triggering an out-of-bounds access here,
>
> arch/arm64/mm/dump.c: 282
> if (addr >= st->marker[1].start_address) {
>
> from,
>
> arch/arm64/mm/dump.c: 331
> note_page(st, addr, 2, pud_val(pud));
>
> because st->marker++ is is called after "UEFI runtime end" which is the
> last element in addr_marker[]. Therefore, add a terminator like the one
> for kernel_page_tables, so it can be skipped to print out non-existent
> markers.
>
> # cat /sys/kernel/debug/efi_page_tables
> ---[ UEFI runtime start ]---
> 0x0000000020000000-0x0000000020010000 64K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x0000000020200000-0x0000000021340000 17664K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x0000000021340000-0x0000000021350000 64K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021350000-0x0000000021370000 128K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x0000000021370000-0x0000000021380000 64K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021380000-0x00000000213b0000 192K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x00000000213b0000-0x0000000021490000 896K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021490000-0x00000000214b0000 128K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x00000000214b0000-0x00000000214e0000 192K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x00000000214e0000-0x00000000214f0000 64K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x00000000214f0000-0x0000000021540000 320K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021540000-0x0000000021560000 128K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x0000000021560000-0x0000000021580000 128K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021580000-0x00000000215a0000 128K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x00000000215a0000-0x00000000215d0000 192K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x00000000215d0000-0x00000000215e0000 64K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x00000000215e0000-0x0000000021600000 128K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021600000-0x0000000021800000 2M PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x0000000021800000-0x0000000021840000 256K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021840000-0x00000000218e0000 640K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x00000000218e0000-0x0000000021900000 128K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021900000-0x0000000021920000 128K PTE RW NX SHD AF NG UXN MEM/NORMAL
> 0x0000000021920000-0x0000000021950000 192K PTE RW x SHD AF NG UXN MEM/NORMAL
> 0x0000000021950000-0x00000000219a0000 320K PTE RW NX SHD AF NG UXN DEVICE/nGnRE
> ---[ UEFI runtime end ]---
> ---[ (null) ]---
> ---[ (null) ]---
>
> [12126.163970] BUG: KASAN: global-out-of-bounds in note_page+0x1f0/0xac0
> [12126.170404] Read of size 8 at addr ffff2000123f2ac0 by task read_all/42464
> [12126.199520] Call trace:
> [12126.201972] dump_backtrace+0x0/0x298
> [12126.205627] show_stack+0x24/0x30
> [12126.208944] dump_stack+0xb0/0xdc
> [12126.212258] print_address_description+0x64/0x2b0
> [12126.216954] kasan_report+0x150/0x1a4
> [12126.220610] __asan_report_load8_noabort+0x30/0x3c
> [12126.225392] note_page+0x1f0/0xac0
> [12126.228786] walk_pgd+0xb4/0x244
> [12126.232005] ptdump_walk_pgd+0xec/0x140
> [12126.235833] ptdump_show+0x40/0x50
> [12126.239237] seq_read+0x3f8/0xad0
> [12126.242548] full_proxy_read+0x9c/0xc0
> [12126.246290] __vfs_read+0xfc/0x4c8
> [12126.249684] vfs_read+0xec/0x208
> [12126.252904] ksys_read+0xd0/0x15c
> [12126.256210] __arm64_sys_read+0x84/0x94
> [12126.260048] el0_svc_handler+0x258/0x304
> [12126.263963] el0_svc+0x8/0xc
> [12126.266834]
> [12126.268317] The buggy address belongs to the variable:
> [12126.273458] __compound_literal.0+0x20/0x800
> [12126.277718]
> [12126.279201] Memory state around the buggy address:
> [12126.283987] ffff2000123f2980: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> [12126.291200] ffff2000123f2a00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 fa
> [12126.298412] >ffff2000123f2a80: fa fa fa fa 00 00 00 00 fa fa fa fa 00 00 00 00
> [12126.305624] ^
> [12126.310927] ffff2000123f2b00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> [12126.318140] ffff2000123f2b80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0
>
> Fixes: 9d80448ac92b ("efi/arm64: Add debugfs node to dump UEFI runtime
> page tables")
> Signed-off-by: Qian Cai <cai@....pw>
Thanks Qian
I will queue this as a fix
> ---
> drivers/firmware/efi/arm-runtime.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/firmware/efi/arm-runtime.c b/drivers/firmware/efi/arm-runtime.c
> index 23ea1ed409d1..1d28fd933297 100644
> --- a/drivers/firmware/efi/arm-runtime.c
> +++ b/drivers/firmware/efi/arm-runtime.c
> @@ -38,7 +38,8 @@ static struct ptdump_info efi_ptdump_info = {
> .mm = &efi_mm,
> .markers = (struct addr_marker[]){
> { 0, "UEFI runtime start" },
> - { DEFAULT_MAP_WINDOW_64, "UEFI runtime end" }
> + { DEFAULT_MAP_WINDOW_64, "UEFI runtime end" },
> + { -1, NULL }
> },
> .base_addr = 0,
> };
> --
> 2.17.2 (Apple Git-113)
>
Powered by blists - more mailing lists