Re: [tip:x86/boot] x86/boot: Early parse RSDP and save it in boot_params

From: Ard Biesheuvel <ard.biesheuvel@linaro.org>
To: Borislav Petkov <bp@alien8.de>
Cc: Chao Fan <fanc.fnst@cn.fujitsu.com>,
	Guenter Roeck <linux@roeck-us.net>,
	Thomas Gleixner <tglx@linutronix.de>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
	Ingo Molnar <mingo@redhat.com>,
	"Lendacky, Thomas" <thomas.lendacky@amd.com>,
	Masahiro Yamada <yamada.masahiro@socionext.com>,
	caoj.fnst@cn.fujitsu.com, Juergen Gross <jgross@suse.com>,
	Ingo Molnar <mingo@kernel.org>, Kees Cook <keescook@chromium.org>,
	"the arch/x86 maintainers" <x86@kernel.org>,
	"H. Peter Anvin" <hpa@zytor.com>,
	linux-tip-commits@vger.kernel.org,
	Matt Fleming <matt@codeblueprint.co.uk>
Subject: Re: [tip:x86/boot] x86/boot: Early parse RSDP and save it in boot_params
Date: Mon, 11 Feb 2019 10:33:02 +0000	[thread overview]
Message-ID: <CAKv+Gu-Fk+_uDvRz7nbceaHAWS_mevm0BuDQd1sSYAQpiPgmNw@mail.gmail.com> (raw)
In-Reply-To: <20190211102426.GA19618@zn.tnic>

On Mon, 11 Feb 2019 at 11:24, Borislav Petkov <bp@alien8.de> wrote:
>
> On Mon, Feb 11, 2019 at 10:17:36AM +0000, Ard Biesheuvel wrote:
> > That still does not explain how 'table' can assume a value > 4 GB
> > after assigning the contents of a u32 to it.
>
> See efi_get_rsdp_addr() in tip/master and especially that systable address
> computation:
>
> #ifdef CONFIG_X86_64
>         systab = (efi_system_table_t *)(ei->efi_systab | ((__u64)ei->efi_systab_hi<<32));
> #else
>         if (ei->efi_systab_hi || ei->efi_memmap_hi) {
>                 debug_putstr("Error getting RSDP address: EFI system table located above 4GB.\n");
>                 return 0;
>         }
>         systab = (efi_system_table_t *)ei->efi_systab;
> #endif
>
>         ...
>
>         config_tables = (void *)(systab->tables + size * i);
>
> It is hard to debug that early but I managed to singlestep it last night
> to this deref:
>
>               } else {
>                         efi_config_table_32_t *tmp_table;
>
>                         tmp_table = config_tables;
>                         guid = tmp_table->guid;
>                                 ^^^^^^^^^^^^^^
>
>                         table = tmp_table->table;
>
> which in asm is:
>
> # arch/x86/boot/compressed/acpi.c:114:                  guid = tmp_table->guid;
> #NO_APP
>         movq    (%rdi), %rax    # MEM[(struct efi_config_table_32_t *)config_tables_37].guid, guid
>         movq    8(%rdi), %rsi   # MEM[(struct efi_config_table_32_t *)config_tables_37].guid, guid
> # arch/x86/boot/compressed/acpi.c:115:                  table = tmp_table->table;
>         movl    16(%rdi), %r10d # MEM[(struct efi_config_table_32_t *)config_tables_37].table, table
>
> and %rdi has 0x630646870 so either we got the wrong address or qemu
> mapped it above 4G...
>
> It is only an observation for now though...
>

I think the problem is that efi_system_table_t is native, so if you
want to access a 32-bit EFI system table from a 64-bit build, you will
need to use efi_system_table_32_t explicitly.