tpm: Preserve TPM measurement log across kexec (ppc64)
Subsystems (10)
v7 cover tpm: Preserve TPM measurement log across kexec (ppc64)
The of-tree subsystem does not currently preserve the IBM vTPM 1.2 and vTPM 2.0 measurement logs across a kexec on PowerVM and PowerKVM. This series fixes this for the kexec_file_load() syscall using the flattened device tree (fdt) to carry the TPM measurement log's buffer across kexec.
Stefan
v7:
- Added Nageswara's Tested-by tags
- Added back original comment to inline function and removed Jarkko's R-b tag
v6:
- Add __init to get_kexec_buffer as suggested by Jonathan
- Fixed issue detected by kernel test robot
v5:
- Rebased on 1 more patch that would otherwise create merge conflicts
v4:
- Rebased on 2 patches that would otherwise create merge conflicts; posting these patches in this series with several tags removed so krobot can test the series already
- Changes to individual patches documented in patch descripitons
v3:
- Moved TPM Open Firmware related function to drivers/char/tpm/eventlog/tpm_of.c
v2:
- rearranged patches
- fixed compilation issues for x86
Jonathan McDowell (1):
x86/kexec: Carry forward IMA measurement log on kexec
Palmer Dabbelt (1):
drivers: of: kexec ima: Support 32-bit platforms
Stefan Berger (3):
tpm: of: Make of-tree specific function commonly available
of: kexec: Refactor IMA buffer related functions to make them reusable tpm/kexec: Duplicate TPM measurement log in of-tree for kexec
Vaibhav Jain (1):
of: check previous kernel's ima-kexec-buffer against memory bounds
arch/x86/Kconfig | 1 + arch/x86/include/uapi/asm/bootparam.h | 9 + arch/x86/kernel/e820.c | 6 +- arch/x86/kernel/kexec-bzimage64.c | 42 +++- arch/x86/kernel/setup.c | 63 +++++ drivers/char/tpm/eventlog/of.c | 31 +-- drivers/of/kexec.c | 342 ++++++++++++++++++++++---- include/linux/ima.h | 5 + include/linux/kexec.h | 6 + include/linux/of.h | 11 +- include/linux/tpm.h | 36 +++ kernel/kexec_file.c | 6 + security/integrity/ima/ima_kexec.c | 2 +- 13 files changed, 478 insertions(+), 82 deletions(-)
base-commit: 3d7cb6b04c3f3115719235cc6866b10326de34cd
Signature · 1 line
2.35.1
PATCH v7 1/6 of: check previous kernel's ima-kexec-buffer against memory bounds
From: Vaibhav Jain <[email protected]>
Presently ima_get_kexec_buffer() doesn't check if the previous kernel's ima-kexec-buffer lies outside the addressable memory range. This can result in a kernel panic if the new kernel is booted with 'mem=X' arg and the ima-kexec-buffer was allocated beyond that range by the previous kernel. The panic is usually of the form below:
$ sudo kexec --initrd initrd vmlinux --append='mem=16G'
<snip>
BUG: Unable to handle kernel data access on read at 0xc000c01fff7f0000 Faulting instruction address: 0xc000000000837974 Oops: Kernel access of bad area, sig: 11 [#1] <snip> NIP [c000000000837974] ima_restore_measurement_list+0x94/0x6c0 LR [c00000000083b55c] ima_load_kexec_buffer+0xac/0x160 Call Trace: [c00000000371fa80] [c00000000083b55c] ima_load_kexec_buffer+0xac/0x160 [c00000000371fb00] [c0000000020512c4] ima_init+0x80/0x108 [c00000000371fb70] [c0000000020514dc] init_ima+0x4c/0x120 [c00000000371fbf0] [c000000000012240] do_one_initcall+0x60/0x2c0 [c00000000371fcc0] [c000000002004ad0] kernel_init_freeable+0x344/0x3ec [c00000000371fda0] [c0000000000128a4] kernel_init+0x34/0x1b0 [c00000000371fe10] [c00000000000ce64] ret_from_kernel_thread+0x5c/0x64 Instruction dump: f92100b8 f92100c0 90e10090 910100a0 4182050c 282a0017 3bc00000 40810330 7c0802a6 fb610198 7c9b2378 f80101d0 <a1240000> 2c090001 40820614 e9240010 ---[ end trace 0000000000000000 ]---
Fix this issue by checking returned PFN range of previous kernel's ima-kexec-buffer with page_is_ram() to ensure correct memory bounds.
--- drivers/of/kexec.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
drivers/of/kexec.c +17 -0
@@ -126,6 +126,7 @@ int ima_get_kexec_buffer(void **addr, size_t *size) {int ret, len;unsigned long tmp_addr; +unsigned long start_pfn, end_pfn;size_t tmp_size;const void *prop;
@@ -140,6 +141,22 @@ int ima_get_kexec_buffer(void **addr, size_t *size)if (ret)return ret; +/* Do some sanity on the returned size for the ima-kexec buffer */ +if (!tmp_size) +return -ENOENT ; + +/* +* Calculate the PFNs for the buffer and ensure +* they are with in addressable memory. +*/ + start_pfn =PHYS_PFN (tmp_addr); + end_pfn =PHYS_PFN (tmp_addr + tmp_size -1 ); +if (!page_is_ram (start_pfn) || !page_is_ram (end_pfn)) { +pr_warn ("IMA buffer at 0x%lx, size = 0x%zx beyond memory\n" , + tmp_addr, tmp_size); +return -EINVAL ; + } + *addr =__va (tmp_addr); *size = tmp_size;
Signature · 7 lines
2.35.1 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
PATCH v7 2/6 drivers: of: kexec ima: Support 32-bit platforms
In reply to Stefan BergerFrom: Palmer Dabbelt <[email protected]>
RISC-V recently added kexec_file() support, which uses enables kexec IMA. We're the first 32-bit platform to support this, so we found a build bug.
--- drivers/of/kexec.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/of/kexec.c +2 -2
@@ -253,8 +253,8 @@ static int setup_ima_buffer(const struct kimage *image, void *fdt,if (ret)return -EINVAL ; -pr_debug ("IMA buffer at 0x%llx, size = 0x%zx\n" , - image->ima_buffer_addr, image->ima_buffer_size); +pr_debug ("IMA buffer at 0x%pa, size = 0x%zx\n" , + &image->ima_buffer_addr, image->ima_buffer_size);return 0 ; }
Signature · 7 lines
2.35.1 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
PATCH v7 3/6 x86/kexec: Carry forward IMA measurement log on kexec
In reply to Stefan BergerFrom: Jonathan McDowell <[email protected]>
On kexec file load, the Integrity Measurement Architecture (IMA) subsystem may verify the IMA signature of the kernel and initramfs, and measure it. The command line parameters passed to the kernel in the kexec call may also be measured by IMA.
A remote attestation service can verify a TPM quote based on the TPM event log, the IMA measurement list and the TPM PCR data. This can be achieved only if the IMA measurement log is carried over from the current kernel to the next kernel across the kexec call.
PowerPC and ARM64 both achieve this using device tree with a "linux,ima-kexec-buffer" node. x86 platforms generally don't make use of device tree, so use the setup_data mechanism to pass the IMA buffer to the new kernel.
--- arch/x86/Kconfig | 1 + arch/x86/include/uapi/asm/bootparam.h | 9 ++++ arch/x86/kernel/e820.c | 6 +-- arch/x86/kernel/kexec-bzimage64.c | 42 +++++++++++++++++- arch/x86/kernel/setup.c | 63 +++++++++++++++++++++++++++ drivers/of/kexec.c | 13 +++--- include/linux/ima.h | 5 +++ include/linux/of.h | 2 - security/integrity/ima/ima_kexec.c | 2 +- 9 files changed, 127 insertions(+), 16 deletions(-)
arch/x86/Kconfig +1 -0
@@ -2011,6 +2011,7 @@ config KEXEC_FILE bool "kexec file based system call" select KEXEC_CORE select BUILD_BIN2C + select HAVE_IMA_KEXEC if IMA depends on X86_64 depends on CRYPTO=y depends on CRYPTO_SHA256=y
arch/x86/include/uapi/asm/bootparam.h +9 -0
@@ -11,6 +11,7 @@#define SETUP_APPLE_PROPERTIES 5#define SETUP_JAILHOUSE 6#define SETUP_CC_BLOB 7 +#define SETUP_IMA 8#define SETUP_INDIRECT (1<<31)
@@ -172,6 +173,14 @@ struct jailhouse_setup_data { } __attribute__((packed)) v2; } __attribute__((packed)); +/* +* IMA buffer setup data information from the previous kernel during kexec +*/ +struct ima_setup_data { +__u64 addr; +__u64 size; +} __attribute__((packed)); +/* The so-called "zeropage" */ struct boot_params {struct screen_info screen_info;/* 0x000 */
arch/x86/kernel/e820.c +3 -3
@@ -1017,10 +1017,10 @@ void __init e820__reserve_setup_data(void)e820__range_update (pa_data,sizeof (*data)+data->len,E820_TYPE_RAM ,E820_TYPE_RESERVED_KERN );/* -* SETUP_EFI is supplied by kexec and does not need to be -* reserved. +* SETUP_EFI and SETUP_IMA are supplied by kexec and do not need +* to be reserved. */ -if (data->type !=SETUP_EFI ) +if (data->type !=SETUP_EFI && data->type !=SETUP_IMA )e820__range_update_kexec (pa_data,sizeof (*data) + data->len,E820_TYPE_RAM ,E820_TYPE_RESERVED_KERN );
arch/x86/kernel/kexec-bzimage64.c +40 -2
65 lines not on this page · raw message
arch/x86/kernel/setup.c +63 -0
87 lines not on this page · raw message
drivers/of/kexec.c +5 -8
@@ -9,6 +9,7 @@* Copyright (C) 2016 IBM Corporation */ +#include <linux/ima.h> #include <linux/kernel.h> #include <linux/kexec.h> #include <linux/memblock.h>
@@ -115,6 +116,7 @@ static int do_get_kexec_buffer(const void *prop, int len, unsigned long *addr,return 0 ; } +#ifdef CONFIG_HAVE_IMA_KEXEC /** * ima_get_kexec_buffer - get IMA buffer from the previous kernel * @addr: On successful return, set to point to the buffer contents.
@@ -122,7 +124,7 @@ static int do_get_kexec_buffer(const void *prop, int len, unsigned long *addr,* * Return: 0 on success, negative errno on error. */ -int ima_get_kexec_buffer (void **addr,size_t *size) +int __init ima_get_kexec_buffer (void **addr,size_t *size) {int ret, len;unsigned long tmp_addr;
@@ -130,9 +132,6 @@ int ima_get_kexec_buffer(void **addr, size_t *size)size_t tmp_size;const void *prop; -if (!IS_ENABLED (CONFIG_HAVE_IMA_KEXEC )) -return -ENOTSUPP ; - prop =of_get_property (of_chosen,"linux,ima-kexec-buffer" , &len);if (!prop)return -ENOENT ;
@@ -166,16 +165,13 @@ int ima_get_kexec_buffer(void **addr, size_t *size)/** * ima_free_kexec_buffer - free memory used by the IMA buffer */ -int ima_free_kexec_buffer (void ) +int __init ima_free_kexec_buffer (void ) {int ret;unsigned long addr;size_t size;struct property *prop; -if (!IS_ENABLED (CONFIG_HAVE_IMA_KEXEC )) -return -ENOTSUPP ; - prop =of_find_property (of_chosen,"linux,ima-kexec-buffer" ,NULL );if (!prop)return -ENOENT ;
@@ -190,6 +186,7 @@ int ima_free_kexec_buffer(void) return memblock_phys_free(addr, size); } +#endif /** * remove_ima_buffer - remove the IMA buffer property and reservation from @fdt
include/linux/ima.h +5 -0
@@ -140,6 +140,11 @@ static inline int ima_measure_critical_data(const char *event_label,#endif /* CONFIG_IMA */ +#ifdef CONFIG_HAVE_IMA_KEXEC +int __init ima_free_kexec_buffer (void ); +int __init ima_get_kexec_buffer (void **addr,size_t *size); +#endif +#ifdef CONFIG_IMA_SECURE_AND_OR_TRUSTED_BOOT extern bool arch_ima_get_secureboot (void );extern const char *const *arch_get_ima_policy (void );
include/linux/of.h +0 -2
@@ -441,8 +441,6 @@ void *of_kexec_alloc_and_setup_fdt(const struct kimage *image,unsigned long initrd_load_addr, unsigned long initrd_len, const char *cmdline, size_t extra_fdt_size); -int ima_get_kexec_buffer (void **addr,size_t *size); -int ima_free_kexec_buffer (void );#else /* CONFIG_OF */ static inline void of_core_init (void )
security/integrity/ima/ima_kexec.c +1 -1
@@ -137,7 +137,7 @@ void ima_add_kexec_buffer(struct kimage *image)/* * Restore the measurement list from the previous kernel. */ -void ima_load_kexec_buffer (void ) +void __init ima_load_kexec_buffer (void ) {void *kexec_buffer =NULL ;size_t kexec_buffer_size =0 ;
Signature · 7 lines
2.35.1 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
Re: [PATCH v7 3/6] x86/kexec: Carry forward IMA measurement log on kexec
Quoted 17 lines · Stefan Berger
From: Jonathan McDowell <[email protected]>
On kexec file load, the Integrity Measurement Architecture (IMA) subsystem may verify the IMA signature of the kernel and initramfs, and measure it. The command line parameters passed to the kernel in the kexec call may also be measured by IMA.
A remote attestation service can verify a TPM quote based on the TPM event log, the IMA measurement list and the TPM PCR data. This can be achieved only if the IMA measurement log is carried over from the current kernel to the next kernel across the kexec call.
PowerPC and ARM64 both achieve this using device tree with a "linux,ima-kexec-buffer" node. x86 platforms generally don't make use of device tree, so use the
setup_datamechanism to pass the IMA buffer to the new kernel.
Quoted 4 lines
Is there any particular reason to keep sending a patch which is already upstream?
Signature · 6 lines
Regards/Gruss,
Boris.
SUSE Software Solutions Germany GmbH
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)Re: [PATCH v7 3/6] x86/kexec: Carry forward IMA measurement log on kexec
Quoted 18 lines · Borislav Petkov
Quoted 17 lines · Stefan Berger
From: Jonathan McDowell <[email protected]>
On kexec file load, the Integrity Measurement Architecture (IMA) subsystem may verify the IMA signature of the kernel and initramfs, and measure it. The command line parameters passed to the kernel in the kexec call may also be measured by IMA.
A remote attestation service can verify a TPM quote based on the TPM event log, the IMA measurement list and the TPM PCR data. This can be achieved only if the IMA measurement log is carried over from the current kernel to the next kernel across the kexec call.
PowerPC and ARM64 both achieve this using device tree with a "linux,ima-kexec-buffer" node. x86 platforms generally don't make use of device tree, so use the
setup_datamechanism to pass the IMA buffer to the new kernel.
Quoted 8 lines
Quoted 4 lines
Is there any particular reason to keep sending a patch which is already upstream?
Yes, so this series can be tested by krobot. I only based this series on 5.19 so far, so if it's upstreamed since then it will go missing next time when I base it on 5.20-rc1 or so.
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Re: [PATCH v7 3/6] x86/kexec: Carry forward IMA measurement log on kexec
Quoted 1 line · Stefan Berger
Yes, so this series can be tested by krobot.
You mean Intel's 0day robot?
I believe that thing has by now enough logic to figure out which branch to base patches ontop. Or maybe there's some magic incantation to tell it which base commit to use so that you can simply do your patches ontop of latest linux-next instead of having to carry upstreamed patches.
Also, there's a little point in testing against 5.19 when you wanna test it against v6.0-rc1...
Signature · 6 lines
Regards/Gruss,
Boris.
SUSE Software Solutions Germany GmbH
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)PATCH v7 4/6 tpm: of: Make of-tree specific function commonly available
In reply to Stefan BergerSimplify tpm_read_log_of() by moving reusable parts of the code into an inline function that makes it commonly available so it can be used also for kexec support. Call the new of_tpm_get_sml_parameters() function from the TPM Open Firmware driver.
---
v7:
- Added original comment back into inlined function
v4:
- converted to inline function
--- drivers/char/tpm/eventlog/of.c | 31 +++++------------------------ include/linux/tpm.h | 36 ++++++++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+), 26 deletions(-)
drivers/char/tpm/eventlog/of.c +5 -26
@@ -12,6 +12,7 @@#include <linux/slab.h> #include <linux/of.h> +#include <linux/tpm.h> #include <linux/tpm_eventlog.h> #include "../tpm.h"
@@ -20,11 +21,10 @@int tpm_read_log_of (struct tpm_chip *chip) {struct device_node *np; -const u32 *sizep; -const u64 *basep;struct tpm_bios_log *log;u32 size;u64 base; +int ret; log = &chip->log;if (chip->dev.parent && chip->dev.parent->of_node)
@@ -35,30 +35,9 @@ int tpm_read_log_of(struct tpm_chip *chip)if (of_property_read_bool (np,"powered-while-suspended" )) chip->flags |=TPM_CHIP_FLAG_ALWAYS_POWERED ; - sizep =of_get_property (np,"linux,sml-size" ,NULL ); - basep =of_get_property (np,"linux,sml-base" ,NULL ); -if (sizep ==NULL && basep ==NULL ) -return -ENODEV ; -if (sizep ==NULL || basep ==NULL ) -return -EIO ; - -/* -* For both vtpm/tpm, firmware has log addr and log size in big -* endian format. But in case of vtpm, there is a method called -* sml-handover which is run during kernel init even before -* device tree is setup. This sml-handover function takes care -* of endianness and writes to sml-base and sml-size in little -* endian format. For this reason, vtpm doesn't need conversion -* but physical tpm needs the conversion. -*/ -if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && -of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { - size =be32_to_cpup ((__force __be32 *)sizep); - base =be64_to_cpup ((__force __be64 *)basep); - }else { - size = *sizep; - base = *basep; - } + ret =of_tpm_get_sml_parameters (np, &base, &size); +if (ret <0 ) +return ret;if (size ==0 ) {dev_warn (&chip->dev,"%s: Event log area empty\n" , __func__);
include/linux/tpm.h +36 -0
@@ -460,4 +460,40 @@ static inline struct tpm_chip *tpm_default_chip(void)return NULL ; }#endif + +#ifdef CONFIG_OF +static inline int of_tpm_get_sml_parameters (struct device_node *np, +u64 *base,u32 *size) +{ +const u32 *sizep; +const u64 *basep; + + sizep =of_get_property (np,"linux,sml-size" ,NULL ); + basep =of_get_property (np,"linux,sml-base" ,NULL ); +if (sizep ==NULL && basep ==NULL ) +return -ENODEV ; +if (sizep ==NULL || basep ==NULL ) +return -EIO ; + +/* +* For both vtpm/tpm, firmware has log addr and log size in big +* endian format. But in case of vtpm, there is a method called +* sml-handover which is run during kernel init even before +* device tree is setup. This sml-handover function takes care +* of endianness and writes to sml-base and sml-size in little +* endian format. For this reason, vtpm doesn't need conversion +* but physical tpm needs the conversion. +*/ +if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && +of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { + *size =be32_to_cpup ((__force __be32 *)sizep); + *base =be64_to_cpup ((__force __be64 *)basep); + }else { + *size = *sizep; + *base = *basep; + } +return 0 ; +} +#endif +#endif
Signature · 1 line
2.35.1
Re: [PATCH v7 4/6] tpm: of: Make of-tree specific function commonly available
Quoted 131 lines · Stefan Berger
Simplify
tpm_read_log_of()by moving reusable parts of the code into an inline function that makes it commonly available so it can be used also for kexec support. Call the newof_tpm_get_sml_parameters()function from the TPM Open Firmware driver.Signed-off-by Stefan Berger <[email protected]>Cc Jarkko Sakkinen <[email protected]>Cc Jason Gunthorpe <[email protected]>Cc Rob Herring <[email protected]>Cc Frank Rowand <[email protected]>Reviewed-by Mimi Zohar <[email protected]>Tested-by Nageswara R Sastry <[email protected]>---
v7:
- Added original comment back into inlined functionv4:
- converted to inline function--- drivers/char/tpm/eventlog/of.c | 31 +++++------------------------ include/linux/tpm.h | 36 ++++++++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+), 26 deletions(-)drivers/char/tpm/eventlog/of.c +5
-26@@ -12,6 +12,7 @@#include <linux/slab.h> #include <linux/of.h> +#include <linux/tpm.h> #include <linux/tpm_eventlog.h> #include "../tpm.h" @@ -20,11 +21,10 @@int tpm_read_log_of (struct tpm_chip *chip) {struct device_node *np; -const u32 *sizep; -const u64 *basep;struct tpm_bios_log *log;u32 size;u64 base; +int ret; log = &chip->log;if (chip->dev.parent && chip->dev.parent->of_node)@@ -35,30 +35,9 @@ int tpm_read_log_of(struct tpm_chip *chip)if (of_property_read_bool (np,"powered-while-suspended" )) chip->flags |=TPM_CHIP_FLAG_ALWAYS_POWERED ; - sizep =of_get_property (np,"linux,sml-size" ,NULL ); - basep =of_get_property (np,"linux,sml-base" ,NULL ); -if (sizep ==NULL && basep ==NULL ) -return -ENODEV ; -if (sizep ==NULL || basep ==NULL ) -return -EIO ; - -/* -* For both vtpm/tpm, firmware has log addr and log size in big -* endian format. But in case of vtpm, there is a method called -* sml-handover which is run during kernel init even before -* device tree is setup. This sml-handover function takes care -* of endianness and writes to sml-base and sml-size in little -* endian format. For this reason, vtpm doesn't need conversion -* but physical tpm needs the conversion. -*/ -if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && -of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { - size =be32_to_cpup ((__force __be32 *)sizep); - base =be64_to_cpup ((__force __be64 *)basep); - }else { - size = *sizep; - base = *basep; - } + ret =of_tpm_get_sml_parameters (np, &base, &size); +if (ret <0 ) +return ret;if (size ==0 ) {dev_warn (&chip->dev,"%s: Event log area empty\n" , __func__);include/linux/tpm.h +36
-0@@ -460,4 +460,40 @@ static inline struct tpm_chip *tpm_default_chip(void)return NULL ; }#endif + +#ifdef CONFIG_OF +static inline int of_tpm_get_sml_parameters (struct device_node *np, +u64 *base,u32 *size) +{ +const u32 *sizep; +const u64 *basep; + + sizep =of_get_property (np,"linux,sml-size" ,NULL ); + basep =of_get_property (np,"linux,sml-base" ,NULL ); +if (sizep ==NULL && basep ==NULL ) +return -ENODEV ; +if (sizep ==NULL || basep ==NULL ) +return -EIO ; + +/* +* For both vtpm/tpm, firmware has log addr and log size in big +* endian format. But in case of vtpm, there is a method called +* sml-handover which is run during kernel init even before +* device tree is setup. This sml-handover function takes care +* of endianness and writes to sml-base and sml-size in little +* endian format. For this reason, vtpm doesn't need conversion +* but physical tpm needs the conversion. +*/ +if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && +of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { + *size =be32_to_cpup ((__force __be32 *)sizep); + *base =be64_to_cpup ((__force __be64 *)basep); + }else { + *size = *sizep; + *base = *basep; + } +return 0 ; +} +#endif +#endif Signature · 1 line
2.35.1
BR, Jarkkko
Jarkko Sakkinen
maintainerRe: [PATCH v7 4/6] tpm: of: Make of-tree specific function commonly available
Quoted 135 lines · Jarkko Sakkinen
Quoted 131 lines · Stefan Berger
Simplify
tpm_read_log_of()by moving reusable parts of the code into an inline function that makes it commonly available so it can be used also for kexec support. Call the newof_tpm_get_sml_parameters()function from the TPM Open Firmware driver.Signed-off-by Stefan Berger <[email protected]>Cc Jarkko Sakkinen <[email protected]>Cc Jason Gunthorpe <[email protected]>Cc Rob Herring <[email protected]>Cc Frank Rowand <[email protected]>Reviewed-by Mimi Zohar <[email protected]>Tested-by Nageswara R Sastry <[email protected]>---
v7:
- Added original comment back into inlined functionv4:
- converted to inline function--- drivers/char/tpm/eventlog/of.c | 31 +++++------------------------ include/linux/tpm.h | 36 ++++++++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+), 26 deletions(-)drivers/char/tpm/eventlog/of.c +5
-26@@ -12,6 +12,7 @@#include <linux/slab.h> #include <linux/of.h> +#include <linux/tpm.h> #include <linux/tpm_eventlog.h> #include "../tpm.h" @@ -20,11 +21,10 @@int tpm_read_log_of (struct tpm_chip *chip) {struct device_node *np; -const u32 *sizep; -const u64 *basep;struct tpm_bios_log *log;u32 size;u64 base; +int ret; log = &chip->log;if (chip->dev.parent && chip->dev.parent->of_node)@@ -35,30 +35,9 @@ int tpm_read_log_of(struct tpm_chip *chip)if (of_property_read_bool (np,"powered-while-suspended" )) chip->flags |=TPM_CHIP_FLAG_ALWAYS_POWERED ; - sizep =of_get_property (np,"linux,sml-size" ,NULL ); - basep =of_get_property (np,"linux,sml-base" ,NULL ); -if (sizep ==NULL && basep ==NULL ) -return -ENODEV ; -if (sizep ==NULL || basep ==NULL ) -return -EIO ; - -/* -* For both vtpm/tpm, firmware has log addr and log size in big -* endian format. But in case of vtpm, there is a method called -* sml-handover which is run during kernel init even before -* device tree is setup. This sml-handover function takes care -* of endianness and writes to sml-base and sml-size in little -* endian format. For this reason, vtpm doesn't need conversion -* but physical tpm needs the conversion. -*/ -if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && -of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { - size =be32_to_cpup ((__force __be32 *)sizep); - base =be64_to_cpup ((__force __be64 *)basep); - }else { - size = *sizep; - base = *basep; - } + ret =of_tpm_get_sml_parameters (np, &base, &size); +if (ret <0 ) +return ret;if (size ==0 ) {dev_warn (&chip->dev,"%s: Event log area empty\n" , __func__);include/linux/tpm.h +36
-0@@ -460,4 +460,40 @@ static inline struct tpm_chip *tpm_default_chip(void)return NULL ; }#endif + +#ifdef CONFIG_OF +static inline int of_tpm_get_sml_parameters (struct device_node *np, +u64 *base,u32 *size) +{ +const u32 *sizep; +const u64 *basep; + + sizep =of_get_property (np,"linux,sml-size" ,NULL ); + basep =of_get_property (np,"linux,sml-base" ,NULL ); +if (sizep ==NULL && basep ==NULL ) +return -ENODEV ; +if (sizep ==NULL || basep ==NULL ) +return -EIO ; + +/* +* For both vtpm/tpm, firmware has log addr and log size in big +* endian format. But in case of vtpm, there is a method called +* sml-handover which is run during kernel init even before +* device tree is setup. This sml-handover function takes care +* of endianness and writes to sml-base and sml-size in little +* endian format. For this reason, vtpm doesn't need conversion +* but physical tpm needs the conversion. +*/ +if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && +of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { + *size =be32_to_cpup ((__force __be32 *)sizep); + *base =be64_to_cpup ((__force __be64 *)basep); + }else { + *size = *sizep; + *base = *basep; + } +return 0 ; +} +#endif +#endif Signature · 1 line
2.35.1Reviewed-by Jarkko Sakkinen <[email protected]>
Quoted 1 line
BR, Jarkkko
Should I pick this or will the full patch set be picked by someone?
BR, Jarkko
PATCH v7 5/6 of: kexec: Refactor IMA buffer related functions to make them reusable
In reply to Stefan BergerRefactor IMA buffer related functions to make them reusable for carrying TPM logs across kexec.
---
v6:
- Add __init to get_kexec_buffer as suggested by Jonathan
v5:
- Rebased on Jonathan McDowell's commit "b69a2afd5afc x86/kexec: Carry forward IMA measurement log on kexec"
v4:
- Move debug output into setup_buffer()
--- drivers/of/kexec.c | 126 ++++++++++++++++++++++++++------------------- 1 file changed, 74 insertions(+), 52 deletions(-)
drivers/of/kexec.c +74 -52
@@ -117,45 +117,57 @@ static int do_get_kexec_buffer(const void *prop, int len, unsigned long *addr, }#ifdef CONFIG_HAVE_IMA_KEXEC -/** -* ima_get_kexec_buffer - get IMA buffer from the previous kernel -* @addr: On successful return, set to point to the buffer contents. -* @size: On successful return, set to the buffer size. -* -* Return: 0 on success, negative errno on error. -*/ -int __init ima_get_kexec_buffer (void **addr,size_t *size) +static int __init get_kexec_buffer (const char *name,unsigned long *addr, +size_t *size) {int ret, len; -unsigned long tmp_addr;unsigned long start_pfn, end_pfn; -size_t tmp_size;const void *prop; - prop =of_get_property (of_chosen,"linux,ima-kexec-buffer" , &len); + prop =of_get_property (of_chosen, name, &len);if (!prop)return -ENOENT ; - ret =do_get_kexec_buffer (prop, len, &tmp_addr, &tmp_size); + ret =do_get_kexec_buffer (prop, len, addr, size);if (ret)return ret; -/* Do some sanity on the returned size for the ima-kexec buffer */ -if (!tmp_size) +/* Do some sanity on the returned size for the kexec buffer */ +if (!*size)return -ENOENT ;/* * Calculate the PFNs for the buffer and ensure * they are with in addressable memory. */ - start_pfn =PHYS_PFN (tmp_addr); - end_pfn =PHYS_PFN (tmp_addr + tmp_size -1 ); + start_pfn =PHYS_PFN (*addr); + end_pfn =PHYS_PFN (*addr + *size -1 );if (!page_is_ram (start_pfn) || !page_is_ram (end_pfn)) { -pr_warn ("IMA buffer at 0x%lx, size = 0x%zx beyond memory\n" , - tmp_addr, tmp_size); +pr_warn ("%s buffer at 0x%lx, size = 0x%zx beyond memory\n" , + name, *addr, *size);return -EINVAL ; } +return 0 ; +} + +/** +* ima_get_kexec_buffer - get IMA buffer from the previous kernel +* @addr: On successful return, set to point to the buffer contents. +* @size: On successful return, set to the buffer size. +* +* Return: 0 on success, negative errno on error. +*/ +int __init ima_get_kexec_buffer (void **addr,size_t *size) +{ +int ret; +unsigned long tmp_addr; +size_t tmp_size; + + ret =get_kexec_buffer ("linux,ima-kexec-buffer" , &tmp_addr, &tmp_size); +if (ret) +return ret; + *addr =__va (tmp_addr); *size = tmp_size;
@@ -188,72 +200,82 @@ int __init ima_free_kexec_buffer(void) }#endif -/** -* remove_ima_buffer - remove the IMA buffer property and reservation from @fdt -* -* @fdt: Flattened Device Tree to update -* @chosen_node: Offset to the chosen node in the device tree -* -* The IMA measurement buffer is of no use to a subsequent kernel, so we always -* remove it from the device tree. -*/ -static void remove_ima_buffer (void *fdt,int chosen_node) +static int remove_buffer (void *fdt,int chosen_node,const char *name) {int ret, len;unsigned long addr;size_t size;const void *prop; -if (!IS_ENABLED (CONFIG_HAVE_IMA_KEXEC )) -return ; - - prop =fdt_getprop (fdt, chosen_node,"linux,ima-kexec-buffer" , &len); + prop =fdt_getprop (fdt, chosen_node, name, &len);if (!prop) -return ; +return -ENOENT ; ret =do_get_kexec_buffer (prop, len, &addr, &size); -fdt_delprop (fdt, chosen_node,"linux,ima-kexec-buffer" ); +fdt_delprop (fdt, chosen_node, name);if (ret) -return ; +return ret; ret =fdt_find_and_del_mem_rsv (fdt, addr, size);if (!ret) -pr_debug ("Removed old IMA buffer reservation.\n" ); +pr_debug ("Remove old %s buffer reserveration" , name); +return ret; } -#ifdef CONFIG_IMA_KEXEC /** -* setup_ima_buffer - add IMA buffer information to the fdt -* @image: kexec image being loaded. -* @fdt: Flattened device tree for the next kernel. -* @chosen_node: Offset to the chosen node. +* remove_ima_buffer - remove the IMA buffer property and reservation from @fdt * -* Return: 0 on success, or negative errno on error. +* @fdt: Flattened Device Tree to update +* @chosen_node: Offset to the chosen node in the device tree +* +* The IMA measurement buffer is of no use to a subsequent kernel, so we always +* remove it from the device tree. */ -static int setup_ima_buffer (const struct kimage *image,void *fdt, -int chosen_node) +static void remove_ima_buffer (void *fdt,int chosen_node) +{ +if (!IS_ENABLED (CONFIG_HAVE_IMA_KEXEC )) +return ; + +remove_buffer (fdt, chosen_node,"linux,ima-kexec-buffer" ); +} + +#ifdef CONFIG_IMA_KEXEC +static int setup_buffer (void *fdt,int chosen_node,const char *name, +phys_addr_t addr,size_t size) {int ret; -if (!image->ima_buffer_size) +if (!size)return 0 ; ret =fdt_appendprop_addrrange (fdt,0 , chosen_node, -"linux,ima-kexec-buffer" , - image->ima_buffer_addr, - image->ima_buffer_size); + name, addr, size);if (ret <0 )return -EINVAL ; - ret =fdt_add_mem_rsv (fdt, image->ima_buffer_addr, - image->ima_buffer_size); + ret =fdt_add_mem_rsv (fdt, addr, size);if (ret)return -EINVAL ; -pr_debug ("IMA buffer at 0x%pa, size = 0x%zx\n" , - &image->ima_buffer_addr, image->ima_buffer_size); +pr_debug ("%s at 0x%pa, size = 0x%zx\n" , name, &addr, size);return 0 ; + +} + +/** +* setup_ima_buffer - add IMA buffer information to the fdt +* @image: kexec image being loaded. +* @fdt: Flattened device tree for the next kernel. +* @chosen_node: Offset to the chosen node. +* +* Return: 0 on success, or negative errno on error. +*/ +static int setup_ima_buffer (const struct kimage *image,void *fdt, +int chosen_node) +{ +return setup_buffer (fdt, chosen_node,"linux,ima-kexec-buffer" , + image->ima_buffer_addr, image->ima_buffer_size); }#else /* CONFIG_IMA_KEXEC */ static inline int setup_ima_buffer(const struct kimage *image, void *fdt,
Signature · 7 lines
2.35.1 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
PATCH v7 6/6 tpm/kexec: Duplicate TPM measurement log in of-tree for kexec
In reply to Stefan BergerThe memory area of the TPM measurement log is currently not properly duplicated for carrying it across kexec when an Open Firmware Devicetree is used. Therefore, the contents of the log get corrupted. Fix this for the kexec_file_load() syscall by allocating a buffer and copying the contents of the existing log into it. The new buffer is preserved across the kexec and a pointer to it is available when the new kernel is started. To achieve this, store the allocated buffer's address in the flattened device tree (fdt) under the name linux,tpm-kexec-buffer and search for this entry early in the kernel startup before the TPM subsystem starts up. Adjust the pointer in the of-tree stored under linux,sml-base to point to this buffer holding the preserved log. The TPM driver can then read the base address from this entry when making the log available. Invalidate the log by removing 'linux,sml-base' from the devicetree if anything goes wrong with updating the buffer.
Use subsys_initcall() to call the function to restore the buffer even if the TPM subsystem or driver are not used. This allows the buffer to be carried across the next kexec without involvement of the TPM subsystem and ensures a valid buffer pointed to by the of-tree.
Use the subsys_initcall(), rather than an ealier initcall, since page_is_ram() in get_kexec_buffer() only starts working at this stage.
---
v6:
- Define prototype for tpm_add_kexec_buffer under same config options as drivers/of/kexec.c is compiled, provide inline function otherwise. (kernel test robot)
v4:
- Added #include <linux/vmalloc.h> due to parisc
- Use phys_addr_t for physical address rather than void *
- Remove linux,sml-base if the buffer cannot be updated after a kexec
- Added __init to functions where possible
--- drivers/of/kexec.c | 216 +++++++++++++++++++++++++++++++++++++++++- include/linux/kexec.h | 6 ++ include/linux/of.h | 9 +- kernel/kexec_file.c | 6 ++ 4 files changed, 234 insertions(+), 3 deletions(-)
drivers/of/kexec.c +214 -2
@@ -19,6 +19,8 @@#include <linux/random.h> #include <linux/slab.h> #include <linux/types.h> +#include <linux/tpm.h> +#include <linux/vmalloc.h> #define RNG_SEED_SIZE 128
@@ -116,7 +118,6 @@ static int do_get_kexec_buffer(const void *prop, int len, unsigned long *addr,return 0 ; } -#ifdef CONFIG_HAVE_IMA_KEXEC static int __init get_kexec_buffer (const char *name,unsigned long *addr,size_t *size) {
@@ -151,6 +152,7 @@ static int __init get_kexec_buffer(const char *name, unsigned long *addr,return 0 ; } +#ifdef CONFIG_HAVE_IMA_KEXEC /** * ima_get_kexec_buffer - get IMA buffer from the previous kernel * @addr: On successful return, set to point to the buffer contents.
@@ -239,7 +241,6 @@ static void remove_ima_buffer(void *fdt, int chosen_node)remove_buffer (fdt, chosen_node,"linux,ima-kexec-buffer" ); } -#ifdef CONFIG_IMA_KEXEC static int setup_buffer (void *fdt,int chosen_node,const char *name,phys_addr_t addr,size_t size) {
@@ -263,6 +264,7 @@ static int setup_buffer(void *fdt, int chosen_node, const char *name, } +#ifdef CONFIG_IMA_KEXEC /** * setup_ima_buffer - add IMA buffer information to the fdt * @image: kexec image being loaded.
@@ -285,6 +287,213 @@ static inline int setup_ima_buffer(const struct kimage *image, void *fdt, }#endif /* CONFIG_IMA_KEXEC */ +/** +* tpm_get_kexec_buffer - get TPM log buffer from the previous kernel +* @phyaddr: On successful return, set to physical address of buffer +* @size: On successful return, set to the buffer size. +* +* Return: 0 on success, negative errno on error. +*/ +static int __init tpm_get_kexec_buffer (phys_addr_t *phyaddr,size_t *size) +{ +unsigned long tmp_addr; +size_t tmp_size; +int ret; + + ret =get_kexec_buffer ("linux,tpm-kexec-buffer" , &tmp_addr, &tmp_size); +if (ret) +return ret; + + *phyaddr = (phys_addr_t )tmp_addr; + *size = tmp_size; + +return 0 ; +} + +/** +* tpm_of_remove_kexec_buffer - remove the linux,tpm-kexec-buffer node +*/ +static int __init tpm_of_remove_kexec_buffer (void ) +{ +struct property *prop; + + prop =of_find_property (of_chosen,"linux,tpm-kexec-buffer" ,NULL ); +if (!prop) +return -ENOENT ; + +return of_remove_property (of_chosen, prop); +} + +/** +* remove_tpm_buffer - remove the TPM log buffer property and reservation from @fdt +* +* @fdt: Flattened Device Tree to update +* @chosen_node: Offset to the chosen node in the device tree +* +* The TPM log measurement buffer is of no use to a subsequent kernel, so we always +* remove it from the device tree. +*/ +static void remove_tpm_buffer (void *fdt,int chosen_node) +{ +if (!IS_ENABLED (CONFIG_PPC64 )) +return ; + +remove_buffer (fdt, chosen_node,"linux,tpm-kexec-buffer" ); +} + +/** +* setup_tpm_buffer - add TPM measurement log buffer information to the fdt +* @image: kexec image being loaded. +* @fdt: Flattened device tree for the next kernel. +* @chosen_node: Offset to the chosen node. +* +* Return: 0 on success, or negative errno on error. +*/ +static int setup_tpm_buffer (const struct kimage *image,void *fdt, +int chosen_node) +{ +if (!IS_ENABLED (CONFIG_PPC64 )) +return 0 ; + +return setup_buffer (fdt, chosen_node,"linux,tpm-kexec-buffer" , + image->tpm_buffer_addr, image->tpm_buffer_size); +} + +void tpm_add_kexec_buffer (struct kimage *image) +{ +struct kexec_buf kbuf = { .image = image, .buf_align =1 , + .buf_min =0 , .buf_max =ULONG_MAX , + .top_down = true }; +struct device_node *np; +void *buffer; +u32 size; +u64 base; +int ret; + +if (!IS_ENABLED (CONFIG_PPC64 )) +return ; + + np =of_find_node_by_name (NULL ,"vtpm" ); +if (!np) +return ; + +if (of_tpm_get_sml_parameters (np, &base, &size) <0 ) +return ; + + buffer =vmalloc (size); +if (!buffer) +return ; +memcpy (buffer,__va (base), size); + + kbuf.buffer = buffer; + kbuf.bufsz = size; + kbuf.memsz = size; + ret =kexec_add_buffer (&kbuf); +if (ret) { +pr_err ("Error passing over kexec TPM measurement log buffer: %d\n" , + ret); +return ; + } + + image->tpm_buffer = buffer; + image->tpm_buffer_addr = kbuf.mem; + image->tpm_buffer_size = size; +} + +/** +* tpm_post_kexec - Make stored TPM log buffer available in of-tree +*/ +static int __init tpm_post_kexec (void ) +{ +struct property *newprop, *p; +struct device_node *np; +phys_addr_t phyaddr; +u32 oflogsize; +size_t size; +u64 unused; +int ret; + +if (!IS_ENABLED (CONFIG_PPC64 )) +return 0 ; + + np =of_find_node_by_name (NULL ,"vtpm" ); +if (!np) +return 0 ; + +if (!of_get_property (of_chosen,"linux,tpm-kexec-buffer" ,NULL )) { +/* +* linux,tpm-kexec-buffer may be missing on initial boot +* or if previous kernel didn't pass a buffer. +*/ +if (of_get_property (of_chosen,"linux,booted-from-kexec" ,NULL )) { +/* no buffer but kexec'd: remove 'linux,sml-base' */ + ret = -EINVAL ; + goto err_remove_sml_base; + } +return 0 ; + } + +/* +* If any one of the following steps fails we remove linux,sml-base +* to invalidate the TPM log. +*/ + ret =tpm_get_kexec_buffer (&phyaddr, &size); +if (ret) + goto err_remove_kexec_buffer; + +/* logsize must not have changed */ + ret =of_tpm_get_sml_parameters (np, &unused, &oflogsize); +if (ret <0 ) + goto err_free_memblock; + ret = -EINVAL ; +if (oflogsize != size) + goto err_free_memblock; + +/* replace linux,sml-base with new physical address of buffer */ + ret = -ENOMEM ; + newprop =kzalloc (sizeof (*newprop),GFP_KERNEL ); +if (!newprop) + goto err_free_memblock; + + newprop->name =kstrdup ("linux,sml-base" ,GFP_KERNEL ); + newprop->length =sizeof (phyaddr); + newprop->value =kmalloc (sizeof (phyaddr),GFP_KERNEL ); +if (!newprop->name || !newprop->value) + goto err_free_newprop_struct; + +if (of_property_match_string (np,"compatible" ,"IBM,vtpm" ) <0 && +of_property_match_string (np,"compatible" ,"IBM,vtpm20" ) <0 ) { + ret = -ENODEV ; + goto err_free_newprop_struct; + }else { + *(phys_addr_t *)newprop->value = phyaddr; + } + + ret =of_update_property (np, newprop); +if (ret) { +pr_err ("Could not update linux,sml-base with new address" ); + goto err_free_newprop_struct; + } + +return 0 ; + +err_free_newprop_struct: +kfree (newprop->value); +kfree (newprop->name); +kfree (newprop); +err_free_memblock: +memblock_phys_free ((phys_addr_t )phyaddr, size); +err_remove_kexec_buffer: +tpm_of_remove_kexec_buffer (); +err_remove_sml_base: + p =of_find_property (np,"linux,sml-base" ,NULL ); +if (p) +of_remove_property (np, p); + +return ret; +} +subsys_initcall (tpm_post_kexec); +/* * of_kexec_alloc_and_setup_fdt - Alloc and setup a new Flattened Device Tree *
@@ -483,6 +692,9 @@ void *of_kexec_alloc_and_setup_fdt(const struct kimage *image,remove_ima_buffer (fdt, chosen_node); ret =setup_ima_buffer (image, fdt,fdt_path_offset (fdt,"/chosen" )); +remove_tpm_buffer (fdt, chosen_node); + ret =setup_tpm_buffer (image, fdt,fdt_path_offset (fdt,"/chosen" )); + out:if (ret) {kvfree (fdt);
include/linux/kexec.h +6 -0
@@ -349,6 +349,12 @@ struct kimage {void *elf_headers;unsigned long elf_headers_sz;unsigned long elf_load_addr; + +/* Virtual address of TPM log buffer for kexec syscall */ +void *tpm_buffer; + +phys_addr_t tpm_buffer_addr; +size_t tpm_buffer_size; };/* kexec interface functions */
include/linux/of.h +8 -1
@@ -100,6 +100,8 @@ struct of_reconfig_data {struct property *old_prop; }; +struct kimage ; +/* initialize a node */ extern struct kobj_type of_node_ktype;extern const struct fwnode_operations of_fwnode_ops;
@@ -436,7 +438,6 @@ int of_map_id(struct device_node *np, u32 id,phys_addr_t of_dma_get_max_cpu_address (struct device_node *np); -struct kimage ; void *of_kexec_alloc_and_setup_fdt(const struct kimage *image, unsigned long initrd_load_addr, unsigned long initrd_len,
@@ -1606,4 +1607,10 @@ static inline int of_overlay_notifier_unregister(struct notifier_block *nb)#endif +#if defined(CONFIG_KEXEC_FILE ) && defined(CONFIG_OF_FLATTREE ) +void tpm_add_kexec_buffer (struct kimage *image); +#else +static inline void tpm_add_kexec_buffer (struct kimage *image) { } +#endif +#endif /* _LINUX_OF_H */
kernel/kexec_file.c +6 -0
@@ -27,6 +27,7 @@#include <linux/kernel_read_file.h> #include <linux/syscalls.h> #include <linux/vmalloc.h> +#include <linux/of.h> #include "kexec_internal.h" #ifdef CONFIG_KEXEC_SIG
@@ -146,6 +147,9 @@ void kimage_file_post_load_cleanup(struct kimage *image) image->ima_buffer =NULL ;#endif /* CONFIG_IMA_KEXEC */ +vfree (image->tpm_buffer); + image->tpm_buffer =NULL ; +/* See if architecture has anything to cleanup post load */ arch_kimage_file_post_load_cleanup (image);
@@ -252,6 +256,8 @@ kimage_file_prepare_segments(struct kimage *image, int kernel_fd, int initrd_fd,/* IMA needs to pass the measurement list to the next kernel. */ ima_add_kexec_buffer (image); +/* Pass the TPM measurement log to next kernel */ +tpm_add_kexec_buffer (image);/* Call arch image load handlers */ ldata =arch_kexec_kernel_image_load (image);
Signature · 7 lines
2.35.1 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
Re: [PATCH v7 6/6] tpm/kexec: Duplicate TPM measurement log in of-tree for kexec
Quoted 40 lines · Stefan Berger
The memory area of the TPM measurement log is currently not properly duplicated for carrying it across kexec when an Open Firmware Devicetree is used. Therefore, the contents of the log get corrupted. Fix this for the
kexec_file_load()syscall by allocating a buffer and copying the contents of the existing log into it. The new buffer is preserved across the kexec and a pointer to it is available when the new kernel is started. To achieve this, store the allocated buffer's address in the flattened device tree (fdt) under the name linux,tpm-kexec-buffer and search for this entry early in the kernel startup before the TPM subsystem starts up. Adjust the pointer in the of-tree stored under linux,sml-base to point to this buffer holding the preserved log. The TPM driver can then read the base address from this entry when making the log available. Invalidate the log by removing 'linux,sml-base' from the devicetree if anything goes wrong with updating the buffer.Use
subsys_initcall()to call the function to restore the buffer even if the TPM subsystem or driver are not used. This allows the buffer to be carried across the next kexec without involvement of the TPM subsystem and ensures a valid buffer pointed to by the of-tree.Use the
subsys_initcall(), rather than an ealier initcall, sincepage_is_ram()inget_kexec_buffer()only starts working at this stage.Signed-off-by Stefan Berger <[email protected]>Cc Rob Herring <[email protected]>Cc Frank Rowand <[email protected]>Cc Eric Biederman <[email protected]>Tested-by Nageswara R Sastry <[email protected]>---
v6:
- Define prototype fortpm_add_kexec_bufferunder same config options asdrivers/of/kexec.cis compiled, provide inline function otherwise. (kernel test robot)v4:
- Added #include <linux/vmalloc.h> due to parisc
- Usephys_addr_tfor physical address rather than void *
- Remove linux,sml-base if the buffer cannot be updated after a kexec
- Added__initto functions where possible
Quoted 7 lines
--- drivers/of/kexec.c | 216 +++++++++++++++++++++++++++++++++++++++++- include/linux/kexec.h | 6 ++ include/linux/of.h | 9 +- kernel/kexec_file.c | 6 ++ 4 files changed, 234 insertions(+), 3 deletions(-)
I can confirm this patch set fixes an issue that guest kdump kernel crashes on POWER9 host by applying it to 5.19.1 (there is a conflict when applying this patch set to latest kernel i.e. 6.0.0-rc1).
Previously, the issue can be reproduced as follows,
1. revert commit 7c5ed82b800d ("powerpc: Set crashkernel offset to mid
of RMA region") which masks the issue
2. build & boot the kernel with crashkernel=512M 3. load kdump kernel and trigger kernel crash,
[ 4.209792] Oops: Kernel access of bad area, sig: 11 [#1] [ 4.210373] LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries [ 4.210985] Modules linked in: [ 4.211570] CPU: 12 PID: 1 Comm: swapper/12 Not tainted 6.0.0-rc1+ #3 [ 4.212204] NIP: c0000000080a6540 LR: c00000000840c630 CTR: 0000000000000800 [ 4.212871] REGS: c00000000e66b130 TRAP: 0300 Not tainted (6.0.0-rc1+) [ 4.213529] MSR: 800000000280b033 <SF,VEC,VSX,EE,FP,ME,IR,DR,RI,LE> CR: 24000280 XER: 0000005b [ 4.214242] CFAR: c0000000080a6528 DAR: c00000002ffb0000 DSISR: 40000000 IRQMASK: 0 [ 4.214242] GPR00: 000000000e000000 c00000000e66b3d0 c00000000aa69a00 c0000000117c0000 [ 4.214242] GPR04: c00000002ffb0000 0000000000040000 0000000000000800 03fffffffee84000 [ 4.214242] GPR08: 0000000080000000 0000000000000010 0000000000000020 0000000000000030 [ 4.214242] GPR12: 0000000000000040 c00000000ad61000 0000000000000050 0000000000000060 [ 4.214242] GPR16: 0000000000000070 0000000000000000 0000000000000000 0000000000000000 [ 4.214242] GPR20: 0000000000000000 0000000000000000 c000000008f7ebd8 c00000000e418420 [ 4.214242] GPR24: c00000000ad3c389 c000000008f7ed60 c00000000ecc9000 c000000011271848 [ 4.214242] GPR28: c000000027fcee5c c00000002ffb0000 0000000000040000 c0000000117c0000 [ 4.220176] NIP [c0000000080a6540] memcpy_power7+0x400/0x7d0 [ 4.220808] LR [c00000000840c630] kmemdup+0x50/0x80 [ 4.221428] Call Trace: [ 4.222006] [c00000000e66b3d0] [c0000000080a63b4] memcpy_power7+0x274/0x7d0 (unreliable) [ 4.222662] [c00000000e66b4d0] [c00000000840c630] kmemdup+0x50/0x80 [ 4.223293] [c00000000e66b510] [c000000008973ee0] tpm_read_log_of+0x110/0x1f0 [ 4.223929] [c00000000e66b5a0] [c000000008973020] tpm_bios_log_setup+0x80/0x250 [ 4.224552] [c00000000e66b630] [c00000000896b938] tpm_chip_register.part.0+0x38/0x2a0 [ 4.225170] [c00000000e66b6b0] [c000000008979a80] tpm_ibmvtpm_probe+0x530/0x7d0 [ 4.225777] [c00000000e66b790] [c0000000081055e4] vio_bus_probe+0x94/0x150 [ 4.226366] [c00000000e66b7e0] [c0000000089936d4] really_probe+0x104/0x550 [ 4.226947] [c00000000e66b860] [c000000008993bd4] __driver_probe_device+0xb4/0x240 [ 4.227560] [c00000000e66b8e0] [c000000008993db4] driver_probe_device+0x54/0x130 [ 4.228159] [c00000000e66b920] [c000000008994a88] __driver_attach+0x128/0x2d0 [ 4.228748] [c00000000e66b9a0] [c00000000898fcf8] bus_for_each_dev+0xa8/0x130 [ 4.229325] [c00000000e66ba00] [c000000008992bd4] driver_attach+0x34/0x50 [ 4.229887] [c00000000e66ba20] [c0000000089922d8] bus_add_driver+0x218/0x300 [ 4.230453] [c00000000e66bab0] [c000000008995f94] driver_register+0xb4/0x1c0 [ 4.231012] [c00000000e66bb20] [c000000008103fb0] __vio_register_driver+0x80/0xf0 [ 4.231569] [c00000000e66bba0] [c00000000a06807c] ibmvtpm_module_init+0x34/0x48 [ 4.232119] [c00000000e66bbc0] [c000000008011c14] do_one_initcall+0x64/0x300 [ 4.232664] [c00000000e66bca0] [c00000000a0053f0] do_initcalls+0x13c/0x190 [ 4.233183] [c00000000e66bd50] [c00000000a00570c] kernel_init_freeable+0x22c/0x2a0 [ 4.233706] [c00000000e66bdb0] [c000000008012240] kernel_init+0x30/0x1a0 [ 4.234204] [c00000000e66be10] [c00000000800ca54] ret_from_kernel_thread+0x5c/0x64 [ 4.234721] Instruction dump: [ 4.235191] fa010080 39800040 39c00050 39e00060 3a000070 7cc903a6 48000018 60000000 [ 4.235738] 60000000 60000000 60000000 60000000 <7ce020ce> 7cc448ce 7ca450ce 7c8458ce [ 4.236302] ---[ end trace 0000000000000000 ]---
Quoted 72 lines · Stefan Berger
The of-tree subsystem does not currently preserve the IBM vTPM 1.2 and vTPM 2.0 measurement logs across a kexec on PowerVM and PowerKVM. This series fixes this for the
kexec_file_load()syscall using the flattened device tree (fdt) to carry the TPM measurement log's buffer across kexec.Stefan
v7:
- Added Nageswara's Tested-by tags
- Added back original comment to inline function and removed Jarkko's R-b tagv6:
- Add__inittoget_kexec_bufferas suggested by Jonathan
- Fixed issue detected by kernel test robotv5:
- Rebased on 1 more patch that would otherwise create merge conflictsv4:
- Rebased on 2 patches that would otherwise create merge conflicts; posting these patches in this series with several tags removed so krobot can test the series already
- Changes to individual patches documented in patch descripitonsv3:
- Moved TPM Open Firmware related function todrivers/char/tpm/eventlog/tpm_of.cv2:
- rearranged patches
- fixed compilation issues for x86Jonathan McDowell (1):
x86/kexec: Carry forward IMA measurement log on kexecPalmer Dabbelt (1):
drivers: of: kexec ima: Support 32-bit platformsStefan Berger (3):
tpm: of: Make of-tree specific function commonly available
of: kexec: Refactor IMA buffer related functions to make them reusable tpm/kexec: Duplicate TPM measurement log in of-tree for kexecVaibhav Jain (1):
of: check previous kernel's ima-kexec-buffer against memory boundsarch/x86/Kconfig | 1 + arch/x86/include/uapi/asm/bootparam.h | 9 + arch/x86/kernel/e820.c | 6 +- arch/x86/kernel/kexec-bzimage64.c | 42 +++- arch/x86/kernel/setup.c | 63 +++++ drivers/char/tpm/eventlog/of.c | 31 +-- drivers/of/kexec.c | 342 ++++++++++++++++++++++---- include/linux/ima.h | 5 + include/linux/kexec.h | 6 + include/linux/of.h | 11 +- include/linux/tpm.h | 36 +++ kernel/kexec_file.c | 6 + security/integrity/ima/ima_kexec.c | 2 +- 13 files changed, 478 insertions(+), 82 deletions(-)base-commit:
3d7cb6b04c3f3115719235cc6866b10326de34cdSignature · 7 lines
2.35.1 _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
Signature · 2 lines
Best regards, Coiby
Quoted 3 lines · Coiby Xu
I can confirm this patch set fixes an issue that guest kdump kernel crashes on POWER9 host by applying it to 5.19.1 (there is a conflict when applying this patch set to latest kernel i.e. 6.0.0-rc1).
FYI, here's the error of applying it to 6.0.0-rc1,
[root@localhost linux]# git am ../v7_tpm_log.mbox Applying: of: check previous kernel's ima-kexec-buffer against memory bounds
error: patch failed: drivers/of/kexec.c:126
error: drivers/of/kexec.c: patch does not apply
Patch failed at 0001 of: check previous kernel's ima-kexec-buffer against memory bounds
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Signature · 8 lines
Best regards, Coiby _______________________________________________ kexec mailing list [email protected] http://lists.infradead.org/mailman/listinfo/kexec
Quoted 3 lines · Coiby Xu
I can confirm this patch set fixes an issue that guest kdump kernel crashes on POWER9 host by applying it to 5.19.1 (there is a conflict when applying this patch set to latest kernel i.e. 6.0.0-rc1)
I rebased it. 2 of the borrowed patches disappeared now since they are upstream already and the rest applied without conflict...
Quoted 2 lines
Tested-by Coiby Xu <[email protected]>
Thanks.