Re: [PATCH 28/42] x86, boot: Allow 64bit EFI kernel to be loaded above 4G

From: Kees Cook
Date: Tue Jul 07 2015 - 19:12:37 EST

Next message: Kees Cook: "Re: [PATCH 25/42] x86, boot: print compression suffix in decompress stage"
Previous message: Toshi Kani: "Re: [PATCH v5 2/6] arch: unify ioremap prototypes and macro aliases"
In reply to: Yinghai Lu: "[PATCH 28/42] x86, boot: Allow 64bit EFI kernel to be loaded above 4G"
Next in thread: Yinghai Lu: "[PATCH 09/42] x86, boot: Split kernel_ident_mapping_init to another file"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue, Jul 7, 2015 at 1:20 PM, Yinghai Lu <yinghai@xxxxxxxxxx> wrote:
> Now could use kexec to place kernel/boot_params/cmd_line/initrd
> above 4G, but that is with legacy interface with startup_64 directly.
>
> This patch will allow 64bit EFI kernel to be loaded above 4G
> and use EFI HANDOVER PROTOCOL to start the kernel.
>
> Current 32bit code32_start is used for passing around load address,
> so it will overflow when kernel is loaded abover 4G.
>
> The patch mainly add ext_code32_start to take load address high 32bits.
>
> After this patch, could use patched grub2-x86_64.efi to place
> kernel/boot_params/cmd_line/initrd all above 4G and execute the kernel
> above 4G.
>
> bootlog like:
>
> kernel: done [ linux 9.25MiB 100% 6.66MiB/s ]
> params: [1618fc000,1618fffff]
> cmdline: [1618fb000,1618fb7fe]
> kernel: [15e000000,161385fff]
> initrd: [15bcbe000,15dffffbb]
> initrd: 1 file done [ initrd.img 35.26MiB 100% 11.93MiB/s ]
> early console in decompress_kernel
> decompress_kernel:
> input: [0x15fd0b3b4-0x16063c803], output: 0x15e000000, heap: [0x160645b00-0x16064daff]
>
> Decompressing Linux... xz... Parsing ELF... done.
> Booting the kernel.
> [ 0.000000] bootconsole [uart0] enabled
> [ 0.000000] real_mode_data : phys 00000001618fc000
> [ 0.000000] real_mode_data : virt ffff8801618fc000
> [ 0.000000] Kernel Layout:
> [ 0.000000] .text: [0x15e000000-0x15f08f72c]
> [ 0.000000] .rodata: [0x15f200000-0x15fa44fff]
> [ 0.000000] .data: [0x15fc00000-0x15fe545ff]
> [ 0.000000] .init: [0x15fe56000-0x16021afff]
> [ 0.000000] .bss: [0x160229000-0x16135ffff]
> [ 0.000000] .brk: [0x161360000-0x161385fff]
> [ 0.000000] memblock_reserve: [0x0000000009f000-0x000000000fffff] flags 0x0 * BIOS reserved
> ...
> [ 0.000000] memblock_reserve: [0x0000015e000000-0x0000016135ffff] flags 0x0 TEXT DATA BSS
> [ 0.000000] memblock_reserve: [0x0000015bcbe000-0x0000015dffffff] flags 0x0 RAMDISK
>
> -v2: add cast to avoid warning with 32bit, also update description for
> ext_code32_start in boot.txt
> -v3: change to 4.0 from 3.20.
>
> Signed-off-by: Yinghai Lu <yinghai@xxxxxxxxxx>
> ---
> Documentation/x86/boot.txt | 19 +++++++++++++++++++
> arch/x86/boot/compressed/eboot.c | 15 ++++++++++-----
> arch/x86/boot/compressed/head_64.S | 7 ++++++-
> arch/x86/boot/header.S | 3 ++-
> arch/x86/include/uapi/asm/bootparam.h | 1 +
> arch/x86/kernel/asm-offsets.c | 1 +
> 6 files changed, 39 insertions(+), 7 deletions(-)
>
> diff --git a/Documentation/x86/boot.txt b/Documentation/x86/boot.txt
> index 9da6f35..90efaa2 100644
> --- a/Documentation/x86/boot.txt
> +++ b/Documentation/x86/boot.txt
> @@ -61,6 +61,9 @@ Protocol 2.12: (Kernel 3.8) Added the xloadflags field and extension fields
> to struct boot_params for loading bzImage and ramdisk
> above 4G in 64bit.
>
> +Protocol 2.14: (Kernel 4.0) Added the ext_code32_start to support 64bit
> + EFI kernel to be loaded above 4G.
> +

Should be at least kernel 4.3.

> **** MEMORY LAYOUT
>
> The traditional memory map for the kernel loader, used for Image or
> @@ -197,6 +200,7 @@ Offset Proto Name Meaning
> 0258/8 2.10+ pref_address Preferred loading address
> 0260/4 2.10+ init_size Linear memory required during initialization
> 0264/4 2.11+ handover_offset Offset of handover entry point
> +0268/4 2.14+ ext_code32_start Extended part for code32_start
>
> (1) For backwards compatibility, if the setup_sects field contains 0, the
> real value is 4.
> @@ -744,6 +748,14 @@ Offset/size: 0x264/4
>
> See EFI HANDOVER PROTOCOL below for more details.
>
> +Field name: ext_code32_start
> +Type: modify (optional, reloc)
> +Offset/size: 0x268/4
> +Protocol: 2.14+
> +
> + This field is the upper 32bits of load address when EFI 64bit kernel
> + is loaded above 4G. And it is used with code32_start to compare to
> + pref_address to decide if kernel need to be relocated further.
>
> **** THE IMAGE CHECKSUM
>
> @@ -1127,4 +1139,11 @@ The boot loader *must* fill out the following fields in bp,
> o hdr.ramdisk_image (if applicable)
> o hdr.ramdisk_size (if applicable)
>
> +for 64bit, when loading above 4G, *must* fill out the following fields,
> +
> + o hdr.ext_code32_start
> + o ext_cmd_line_ptr
> + o ext_ramdisk_image (if applicable)
> + o ext_ramdisk_size (if applicable)
> +
> All other fields should be zero.
> diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c
> index 2c82bd1..05d77a5 100644
> --- a/arch/x86/boot/compressed/eboot.c
> +++ b/arch/x86/boot/compressed/eboot.c
> @@ -1394,6 +1394,7 @@ struct boot_params *efi_main(struct efi_config *c,
> void *handle;
> efi_system_table_t *_table;
> bool is64;
> + unsigned long loaded_addr;
>
> efi_early = c;
>
> @@ -1435,9 +1436,12 @@ struct boot_params *efi_main(struct efi_config *c,
> * If the kernel isn't already loaded at the preferred load
> * address, relocate it.
> */
> - if (hdr->pref_address != hdr->code32_start) {
> - unsigned long bzimage_addr = hdr->code32_start;
> - status = efi_relocate_kernel(sys_table, &bzimage_addr,
> + loaded_addr = hdr->code32_start;
> + loaded_addr |= (unsigned long)((u64)hdr->ext_code32_start << 32);
> + if (hdr->pref_address != loaded_addr) {
> + unsigned long loaded_addr_orig = loaded_addr;
> +
> + status = efi_relocate_kernel(sys_table, &loaded_addr,
> hdr->init_size, hdr->init_size,
> hdr->pref_address,
> hdr->kernel_alignment);
> @@ -1446,8 +1450,9 @@ struct boot_params *efi_main(struct efi_config *c,
> goto fail;
> }
>
> - hdr->pref_address = hdr->code32_start;
> - hdr->code32_start = bzimage_addr;
> + hdr->pref_address = loaded_addr_orig;
> + hdr->code32_start = loaded_addr & 0xffffffff;
> + hdr->ext_code32_start = (unsigned long)((u64)loaded_addr >> 32);
> }
>
> status = exit_boot(boot_params, handle, is64);
> diff --git a/arch/x86/boot/compressed/head_64.S b/arch/x86/boot/compressed/head_64.S
> index 075bb15..ab52d2c 100644
> --- a/arch/x86/boot/compressed/head_64.S
> +++ b/arch/x86/boot/compressed/head_64.S
> @@ -266,6 +266,8 @@ ENTRY(efi_pe_entry)
> mov %rax, %rsi
> leaq startup_32(%rip), %rax
> movl %eax, BP_code32_start(%rsi)
> + shr $32, %rax
> + movl %eax, BP_ext_code32_start(%rsi)
> jmp 2f /* Skip the relocation */
>
> handover_entry:
> @@ -289,7 +291,10 @@ fail:
> hlt
> jmp fail
> 2:
> - movl BP_code32_start(%esi), %eax
> + movl BP_code32_start(%rsi), %eax
> + movl BP_ext_code32_start(%rsi), %ebx
> + shl $32, %rbx
> + orq %rbx, %rax
> leaq preferred_addr(%rax), %rax
> jmp *%rax
>
> diff --git a/arch/x86/boot/header.S b/arch/x86/boot/header.S
> index 99204e5..09e7c69 100644
> --- a/arch/x86/boot/header.S
> +++ b/arch/x86/boot/header.S
> @@ -301,7 +301,7 @@ _start:
> # Part 2 of the header, from the old setup.S
>
> .ascii "HdrS" # header signature
> - .word 0x020d # header version number (>= 0x0105)
> + .word 0x020e # header version number (>= 0x0105)
> # or else old loadlin-1.5 will fail)
> .globl realmode_swtch
> realmode_swtch: .word 0, 0 # default_switch, SETUPSEG
> @@ -478,6 +478,7 @@ pref_address: .quad LOAD_PHYSICAL_ADDR # preferred load addr
> #endif
> init_size: .long INIT_SIZE # kernel initialization size
> handover_offset: .long 0 # Filled in by build.c
> +ext_code32_start: .long 0 # werid one!

Comment is a typo?

>
> # End of setup header #####################################################
>
> diff --git a/arch/x86/include/uapi/asm/bootparam.h b/arch/x86/include/uapi/asm/bootparam.h
> index ab456dc..bb9973d 100644
> --- a/arch/x86/include/uapi/asm/bootparam.h
> +++ b/arch/x86/include/uapi/asm/bootparam.h
> @@ -84,6 +84,7 @@ struct setup_header {
> __u64 pref_address;
> __u32 init_size;
> __u32 handover_offset;
> + __u32 ext_code32_start;
> } __attribute__((packed));
>
> struct sys_desc_table {
> diff --git a/arch/x86/kernel/asm-offsets.c b/arch/x86/kernel/asm-offsets.c
> index d2e00bc..3f9789f 100644
> --- a/arch/x86/kernel/asm-offsets.c
> +++ b/arch/x86/kernel/asm-offsets.c
> @@ -90,6 +90,7 @@ void common(void) {
> OFFSET(BP_init_size, boot_params, hdr.init_size);
> OFFSET(BP_pref_address, boot_params, hdr.pref_address);
> OFFSET(BP_code32_start, boot_params, hdr.code32_start);
> + OFFSET(BP_ext_code32_start, boot_params, hdr.ext_code32_start);
>
> BLANK();
> DEFINE(PTREGS_SIZE, sizeof(struct pt_regs));
> --
> 1.8.4.5
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

-Kees

--
Kees Cook
Chrome OS Security
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Kees Cook: "Re: [PATCH 25/42] x86, boot: print compression suffix in decompress stage"
Previous message: Toshi Kani: "Re: [PATCH v5 2/6] arch: unify ioremap prototypes and macro aliases"
In reply to: Yinghai Lu: "[PATCH 28/42] x86, boot: Allow 64bit EFI kernel to be loaded above 4G"
Next in thread: Yinghai Lu: "[PATCH 09/42] x86, boot: Split kernel_ident_mapping_init to another file"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]