[PATCH u-boot v2019.04-aspeed-openbmc v3] aspeed: Disable backdoor interfaces
Zev Weiss
zev at bewilderbeest.net
Wed May 4 10:21:08 AEST 2022
On Mon, May 02, 2022 at 01:11:00AM PDT, Joel Stanley wrote:
>On Tue, 19 Apr 2022 at 23:43, Zev Weiss <zev at bewilderbeest.net> wrote:
>>
>> On ast2400 and ast2500 we now disable the various hardware backdoor
>> interfaces as is done on ast2600. Two Kconfig options can selectively
>> re-enable some of these interfaces: CONFIG_ASPEED_ENABLE_SUPERIO
>> leaves the ast2x00 built-in Super I/O device enabled, as it is
>> required for some systems, and CONFIG_ASPEED_ENABLE_DEBUG_UART leaves
>> the hardware debug UART enabled, since it provides a relatively high
>> ratio of utility to security risk during development.
>>
>> This patch is based on a patch by Andrew Jeffery for an older u-boot
>> branch in the OpenBMC tree for the df-isolate-bmc distro feature flag.
>>
>> Signed-off-by: Zev Weiss <zev at bewilderbeest.net>
>> ---
>>
>> Tested on ast2500 and (hostless, BMC-only) ast2400.
>>
>> Ryan, are you OK with having an option (off by default) to leave the
>> debug UART enabled as in this version of the patch?
>>
>> Ian, if you could test this out with CONFIG_ASPEED_ENABLE_SUPERIO=y on
>> one of your systems and confirm that that setting works as intended
>> that would be great.
>
>Ian, did you get around to testing this?
>
>I've given it a careful review.
>
Thanks Joel!
>>
>> Changes since v2 [1]:
>> - made most of the changes unconditional/unconfigurable, but added
>> Kconfig options to leave Super I/O and debug UART enabled
>>
>> Changes since v1 [0]:
>> - extended to cover ast2400
>> - inverted sense of Kconfig option, default (n) is now secure mode
>> - renamed some register/bit macros more appropriately
>>
>> [0] https://lore.kernel.org/openbmc/20220414040448.27100-1-zev@bewilderbeest.net/
>> [1] https://lore.kernel.org/openbmc/20220414224004.29703-1-zev@bewilderbeest.net/
>>
>> arch/arm/include/asm/arch-aspeed/platform.h | 7 ++
>> .../arm/include/asm/arch-aspeed/scu_ast2400.h | 7 ++
>> .../arm/include/asm/arch-aspeed/scu_ast2500.h | 8 ++
>> arch/arm/mach-aspeed/Kconfig | 22 ++++++
>> arch/arm/mach-aspeed/ast2400/board_common.c | 66 +++++++++++++++++
>> arch/arm/mach-aspeed/ast2500/board_common.c | 73 +++++++++++++++++++
>> 6 files changed, 183 insertions(+)
>>
>> diff --git a/arch/arm/include/asm/arch-aspeed/platform.h b/arch/arm/include/asm/arch-aspeed/platform.h
>> index f016bdaba3e7..f05747642f38 100644
>> --- a/arch/arm/include/asm/arch-aspeed/platform.h
>> +++ b/arch/arm/include/asm/arch-aspeed/platform.h
>> @@ -15,24 +15,31 @@
>> /*********************************************************************************/
>> #if defined(CONFIG_ASPEED_AST2400)
>> #define ASPEED_MAC_COUNT 2
>> +#define ASPEED_SDRAM_CTRL 0x1e6e0000
>> #define ASPEED_HW_STRAP1 0x1e6e2070
>> #define ASPEED_REVISION_ID 0x1e6e207C
>> #define ASPEED_SYS_RESET_CTRL 0x1e6e203C
>> #define ASPEED_VGA_HANDSHAKE0 0x1e6e2040 /* VGA fuction handshake register */
>
>sp: function
>
I'm guessing we don't want to bundle unrelated spelling fixes in
existing code into the same patch, but I can post a separate one to
address this?
>> +#define ASPEED_PCIE_CONFIG_SET 0x1e6e2180
>> #define ASPEED_DRAM_BASE 0x40000000
>> #define ASPEED_SRAM_BASE 0x1E720000
>> +#define ASPEED_LPC_CTRL 0x1e789000
>> #define ASPEED_SRAM_SIZE 0x8000
>> #define ASPEED_FMC_CS0_BASE 0x20000000
>> #elif defined(CONFIG_ASPEED_AST2500)
>> #define ASPEED_MAC_COUNT 2
>> +#define ASPEED_SDRAM_CTRL 0x1e6e0000
>> +#define ASPEED_MISC1_CTRL 0x1e6e202C
>> #define ASPEED_HW_STRAP1 0x1e6e2070
>> #define ASPEED_HW_STRAP2 0x1e6e20D0
>> #define ASPEED_REVISION_ID 0x1e6e207C
>> #define ASPEED_SYS_RESET_CTRL 0x1e6e203C
>> #define ASPEED_VGA_HANDSHAKE0 0x1e6e2040 /* VGA fuction handshake register */
>
>sp: function
>
(As above.)
>> +#define ASPEED_PCIE_CONFIG_SET 0x1e6e2180
>> #define ASPEED_MAC_COUNT 2
>> #define ASPEED_DRAM_BASE 0x80000000
>> #define ASPEED_SRAM_BASE 0x1E720000
>> +#define ASPEED_LPC_CTRL 0x1e789000
>> #define ASPEED_SRAM_SIZE 0x9000
>> #define ASPEED_FMC_CS0_BASE 0x20000000
>> #elif defined(CONFIG_ASPEED_AST2600)
>> diff --git a/arch/arm/include/asm/arch-aspeed/scu_ast2400.h b/arch/arm/include/asm/arch-aspeed/scu_ast2400.h
>> index 9c5d96ae84b9..55875fd8312f 100644
>> --- a/arch/arm/include/asm/arch-aspeed/scu_ast2400.h
>> +++ b/arch/arm/include/asm/arch-aspeed/scu_ast2400.h
>> @@ -8,6 +8,7 @@
>> #define SCU_HWSTRAP_VGAMEM_MASK (3 << SCU_HWSTRAP_VGAMEM_SHIFT)
>> #define SCU_HWSTRAP_MAC1_RGMII (1 << 6)
>> #define SCU_HWSTRAP_MAC2_RGMII (1 << 7)
>> +#define SCU_HWSTRAP_LPC_SIO_DEC_DIS (1 << 20)
>> #define SCU_HWSTRAP_DDR4 (1 << 24)
>> #define SCU_HWSTRAP_CLKIN_25MHZ (1 << 23)
>>
>> @@ -104,6 +105,12 @@
>> #define SCU_CLKDUTY_RGMII2TXCK_SHIFT 16
>> #define SCU_CLKDUTY_RGMII2TXCK_MASK (0x7f << SCU_CLKDUTY_RGMII2TXCK_SHIFT)
>>
>> +#define SCU_PCIE_CONFIG_SET_VGA_MMIO (1 << 1)
>> +#define SCU_PCIE_CONFIG_SET_BMC_EN (1 << 8)
>> +#define SCU_PCIE_CONFIG_SET_BMC_MMIO (1 << 9)
>> +#define SCU_PCIE_CONFIG_SET_BMC_DMA (1 << 14)
>> +
>> +
>> struct ast2400_clk_priv {
>> struct ast2400_scu *scu;
>> };
>> diff --git a/arch/arm/include/asm/arch-aspeed/scu_ast2500.h b/arch/arm/include/asm/arch-aspeed/scu_ast2500.h
>> index 8fe4028e4ff0..06dc998afaa8 100644
>> --- a/arch/arm/include/asm/arch-aspeed/scu_ast2500.h
>> +++ b/arch/arm/include/asm/arch-aspeed/scu_ast2500.h
>> @@ -11,6 +11,7 @@
>> #define SCU_HWSTRAP_VGAMEM_MASK (3 << SCU_HWSTRAP_VGAMEM_SHIFT)
>> #define SCU_HWSTRAP_MAC1_RGMII (1 << 6)
>> #define SCU_HWSTRAP_MAC2_RGMII (1 << 7)
>> +#define SCU_HWSTRAP_LPC_SIO_DEC_DIS (1 << 20)
>> #define SCU_HWSTRAP_DDR4 (1 << 24)
>> #define SCU_HWSTRAP_CLKIN_25MHZ (1 << 23)
>>
>> @@ -107,6 +108,13 @@
>> #define SCU_CLKDUTY_RGMII2TXCK_SHIFT 16
>> #define SCU_CLKDUTY_RGMII2TXCK_MASK (0x7f << SCU_CLKDUTY_RGMII2TXCK_SHIFT)
>>
>> +#define SCU_PCIE_CONFIG_SET_VGA_MMIO (1 << 1)
>> +#define SCU_PCIE_CONFIG_SET_BMC_EN (1 << 8)
>> +#define SCU_PCIE_CONFIG_SET_BMC_MMIO (1 << 9)
>> +#define SCU_PCIE_CONFIG_SET_BMC_DMA (1 << 14)
>> +
>> +#define SCU_MISC_DEBUG_UART_DISABLE (1 << 10)
>> +
>> struct ast2500_clk_priv {
>> struct ast2500_scu *scu;
>> };
>> diff --git a/arch/arm/mach-aspeed/Kconfig b/arch/arm/mach-aspeed/Kconfig
>> index 579a547df61e..fc565e0da830 100644
>> --- a/arch/arm/mach-aspeed/Kconfig
>> +++ b/arch/arm/mach-aspeed/Kconfig
>> @@ -45,6 +45,28 @@ config ASPEED_AST2600
>> which is enabled by support of LPC and eSPI peripherals.
>> endchoice
>>
>> +config ASPEED_ENABLE_SUPERIO
>> + bool "Enable built-in AST2x00 Super I/O hardware"
>> + depends on ASPEED_AST2400 || ASPEED_AST2500
>> + help
>> + The Aspeed AST2400 and AST2500 include a built-in Super I/O
>> + device that is normally disabled; say Y here to enable it.
>
>Can you add a line break here.
>
Sure.
>> + Note that this has security implications: it grants the host
>> + read access to the BMC's entire address space. This should
>> + thus be left disabled unless required by a specific system.
>
>Change "note" to WARNING or DANGER or an all caps label of your choosing.
>
>Do you think we should put these behind a config
>ASPEED_SHOW_DANGEROUS_BACKDOOR_FEATURE so users have to acknowledge
>that option before enabling these?
>
My first thought was that this seemed like it might be overkill given
the emphasis in the description, but then remembered that depending on
the particular Kconfig UI it'd be fairly easy to only see the one-line
summary, so yeah, that seems like a reasonable extra seatbelt to add.
>
>> +
>> +config ASPEED_ENABLE_DEBUG_UART
>> + bool "Enable AST2500 hardware debug UART"
>> + depends on ASPEED_AST2500
>> + help
>> + The Aspeed AST2500 include a hardware-supported, UART-based
>> + debug interface that is normally disabled; say Y here to
>> + enable it.
>
>line break
>
Ack.
>> Note that this has security implications: the
>> + debug UART provide read/write access to the BMC's entire
>
>provides
>
Ack.
>> + address space. This should thus be left disabled on
>> + production systems, but may be useful to enable for
>> + debugging during development.
>> +
>
>I did a build as follows:
>
>$ cat build-ast2500.sh
>#!/bin/bash
>
>set -e
>
>OBJ=ast2500-obj
>CONFIG=evb-ast2500_defconfig
>IMG="$OBJ/test.img"
>
>make -j8 O="$OBJ" -s clean
>make -j8 O="$OBJ" -j8 -s $CONFIG
>CROSS_COMPILE=arm-linux-gnueabi- ARCH=arm make -j8 O="$OBJ" -j8
>DEVICE_TREE=ast2500-evb -s
>size "$OBJ/u-boot"
>
>cp "$OBJ/u-boot.bin" "$OBJ/test.img"
>truncate -s 32M "$OBJ/test.img"
>
>Testing on a qemu instance, before your patch:
>
>$ qemu-system-arm -M ast2500-evb -drive
>file=ast2500-obj/test.img,if=mtd -nographic -nic user,tftp=/srv/tftp/
>
># ./culvert probe
>[*] Probing AHB interfaces
>[*] Performing interface discovery via devmem
>SuperIO: Enabled
>iLPC2AHB Bridge: Read-write
>VGA PCIe device: Enabled
>MMIO on VGA device: Enabled
>P2A write filter state:
>0x00000000-0x0fffffff (Firmware): Read-write
>0x10000000-0x1fffffff (SoC IO): Read-write
>0x20000000-0x2fffffff (BMC Flash): Read-write
>0x30000000-0x3fffffff (Host Flash): Read-write
>0x40000000-0x5fffffff (Reserved): Read-write
>0x60000000-0x7fffffff (LPC Host): Read-write
>0x80000000-0xffffffff (DRAM): Read-write
>X-DMA on VGA device: Enabled
>BMC PCIe device: Disabled
>X-DMA is unconstrained: Yes
>Debug UART: Enabled
>Debug UART enabled on: UART5
>
>After applying your patch:
>
># ./culvert probe
>[*] Probing AHB interfaces
>[*] Performing interface discovery via devmem
>SuperIO: Disabled
>VGA PCIe device: Enabled
>MMIO on VGA device: Disabled
>X-DMA on VGA device: Enabled
>BMC PCIe device: Disabled
>X-DMA is unconstrained: No
>Debug UART: Disabled
>
>
>Similarly with the ast2400:
>
>qemu-system-arm -M palmetto-bmc -drive
>file=ast2400-obj/test.img,if=mtd -nographic -nic user,tftp=/srv/tftp/
>
># ./culvert probe
>[*] Probing AHB interfaces
>[*] Performing interface discovery via devmem
>SuperIO: Enabled
>iLPC2AHB Bridge: Read-write
>VGA PCIe device: Enabled
>MMIO on VGA device: Enabled
>P2A write filter state:
>0x00000000-0x17ffffff (Firmware): Read-write
>0x18000000-0x1fffffff (SoC IO): Read-write
>0x20000000-0x2fffffff (BMC Flash): Read-write
>0x30000000-0x3fffffff (Host Flash): Read-write
>0x40000000-0x5fffffff (DRAM): Read-write
>0x60000000-0x7fffffff (LPC Host): Read-write
>0x80000000-0xffffffff (Reserved): Read-write
>X-DMA on VGA device: Enabled
>BMC PCIe device: Disabled
>X-DMA is unconstrained: Yes
>Debug UART: Absent
>
># ./culvert probe
>[*] Probing AHB interfaces
>[*] Performing interface discovery via devmem
>SuperIO: Disabled
>VGA PCIe device: Enabled
>MMIO on VGA device: Disabled
>X-DMA on VGA device: Enabled
>BMC PCIe device: Disabled
>X-DMA is unconstrained: No
>Debug UART: Absent
>
>These "after" settings look good. I didn't check that qemu models the
>default state of these bits in detail, but the fact that they change
>from "insecure" to "secure" indicates your patch is setting the
>correct things.
>
>Tested-by: Joel Stanley <joel at jms.id.au>
>Reviewed-by: Joel Stanley <joel at jms.id.au>
>
Great, thanks -- I'll add those tags and post v4 with the above changes
soon.
Zev
More information about the openbmc
mailing list