Support kdump with LUKS encryption by reusing LUKS volume keys
Subsystems (6)
v4 cover Support kdump with LUKS encryption by reusing LUKS volume keys
LUKS is the standard for Linux disk encryption. Many users choose LUKS and in some use cases like Confidential VM it's mandated. With kdump enabled, when the 1st kernel crashes, the system could boot into the kdump/crash kernel and dump the memory image i.e. /proc/vmcore to a specified target. Currently, when dumping vmcore to a LUKS encrypted device, there are two problems,
- Kdump kernel may not be able to decrypt the LUKS partition. For some machines, a system administrator may not have a chance to enter the password to decrypt the device in kdump initramfs after the 1st kernel crashes; For cloud confidential VMs, depending on the policy the kdump kernel may not be able to unseal the keys with TPM and the console virtual keyboard is untrusted.
- LUKS2 by default use the memory-hard Argon2 key derivation function which is quite memory-consuming compared to the limited memory reserved for kdump. Take Fedora example, by default, only 256M is reserved for systems having memory between 4G-64G. With LUKS enabled, ~1300M needs to be reserved for kdump. Note if the memory reserved for kdump can't be used by 1st kernel i.e. an user sees ~1300M memory missing in the 1st kernel.
Besides users (at least for Fedora) usually expect kdump to work out of the box i.e. no manual password input is needed. And it doesn't make sense to derivate the keys again in kdump kernel which seems to be redundant work.
This patch set addresses the above issues by make the LUKS volume keys persistent for kdump kernel with the help of cryptsetup's new APIs (--link-vk-to-keyring/--volume-key-keyring). Here is the life cycle of this kdump copy of LUKS volume keys,
1. After the 1st kernel loads the initramfs during boot, systemd use an user-input passphrase or TPM-sealed key to de-crypt the LUKS volume keys and then save the volume keys to specified keyring (using the --link-vk-to-keyring API) and the key will expire within specified time.
2. A user space tool (kdump initramfs builder) writes a key description to /sys/kernel/crash_dm_crypt_keys to inform the 1st kernel to record the key while building the kdump initramfs
3. The kexec_file_load syscall read the volume keys by recorded key descriptions and then save them key to kdump reserved memory and wipe the copy.
4. When the 1st kernel crashes and the kdump initramfs is booted, the kdump initramfs asks the kdump kernel to create a user key using the key stored in kdump reserved memory by writing to to /sys/kernel/crash_dm_crypt_keys. Then the LUKS encrypted devide is unlocked with libcryptsetup's --volume-key-keyring API.
5. The system gets rebooted to the 1st kernel after dumping vmcore to the LUKS encrypted device is finished
After libcryptsetup saving the LUKS volume keys to specified keyring, whoever takes this should be responsible for the safety of these copies of keys. The keys will be saved in the memory area exclusively reserved for kdump where even the 1st kernel has no direct access. And further more, two additional protections are added,
- save the copy randomly in kdump reserved memory as suggested by Jan
- clear the _PAGE_PRESENT flag of the page that stores the copy as suggested by Pingfan
This patch set only supports x86. There will be patches to support other architectures once this patch set gets merged.
v5
- Baoquan
- limit the feature of placing kexec_buf randomly to kdump (CONFIG_CRASH_DUMP)
- add documentation for added sysfs API
- allow to re-send init command to support the case of user switching to a different LUKS-encrypted target
- make CONFIG_CRASH_DM_CRYPT depends on CONFIG_DM_CRYPT
- check if the number of keys exceed KEY_NUM_MAX
- rename (struct keys_header).key_count as (struct keys_header).total_keys to improve code readiblity
- improve commit message
- fix the failure of calling crash_exclude_mem_range (there is a split of mem_range)
- use ret instead of r as return code
- Greg
- add documentation for added sysfs API
- avoid spamming kernel logs
- fix a buffer overflow issue
- keep the state enums synced up with the string values
- use sysfs_emit other than sprintf
- explain KEY_NUM_MAX and KEY_SIZE_MAX
- s/EXPORT_SYMBOL_GPL/EXPORT_SYMBOL/g
- improve code readability
- Rebase onto latest Linus tree
v4
- rebase onto latest Linus tree so Baoquan can apply the patches for code review
- fix kernel test robot warnings
v3
- Support CPU/memory hot-plugging [Baoquan]
- Don't save the keys temporarily to simplify the implementation [Baoquan]
- Support multiple LUKS encrypted volumes
- Read logon key instead of user key to improve security [Ondrej]
- A kernel config option CRASH_DM_CRYPT for this feature (disabled by default)
- Fix warnings found by kernel test robot
- Rebase the code onto 6.9.0-rc5+
v2
- work together with libscryptsetup's --link-vk-to-keyring/--volume-key-keyring APIs [Milan and Ondrej]
- add the case where console virtual keyboard is untrusted for confidential VM
- use dm_crypt_key instead of LUKS volume key [Milan and Eric]
- fix some code format issues
- don't move "struct kexec_segment" declaration
- Rebase the code onto latest Linus tree (6.7.0)
v1
- "Put the luks key handling related to crash_dump out into a separate file kernel/crash_dump_luks.c" [Baoquan]
- Put the generic luks handling code before the x86 specific code to make it easier for other arches to follow suit [Baoquan]
- Use phys_to_virt instead of "pfn -> page -> vaddr" [Dave Hansen]
- Drop the RFC prefix [Dave Young]
- Rebase the code onto latest Linus tree (6.4.0-rc4)
RFC v2
- libcryptsetup interacts with the kernel via sysfs instead of "hacking" dm-crypt
- to save a kdump copy of the LUKS volume key in 1st kernel
- to add a logon key using the copy for libcryptsetup in kdump kernel [Milan]
- to avoid the incorrect usage of LUKS master key in dm-crypt [Milan]
- save the kdump copy of LUKS volume key randomly [Jan]
- mark the kdump copy inaccessible [Pingfan]
- Miscellaneous
- explain when operations related to the LUKS volume key happen [Jan]
- s/master key/volume key/g
- use crash_ instead of kexec_ as function prefix
- fix commit subject prefixes e.g. "x86, kdump" to x86/crash
Coiby Xu (7):
kexec_file: allow to place kexec_buf randomly
crash_dump: make dm crypt keys persist for the kdump kernel
crash_dump: store dm crypt keys in kdump reserved memory
crash_dump: reuse saved dm crypt keys for CPU/memory hot-plugging
crash_dump: retrieve dm crypt keys in kdump kernel x86/crash: pass dm crypt keys to kdump kernel
x86/crash: make the page that stores the dm crypt keys inaccessible
Documentation/ABI/testing/crash_dm_crypt_keys | 35 ++ arch/x86/kernel/crash.c | 20 +- arch/x86/kernel/kexec-bzimage64.c | 7 + arch/x86/kernel/machine_kexec_64.c | 22 ++ include/linux/crash_core.h | 9 +- include/linux/crash_dump.h | 2 + include/linux/kexec.h | 8 + kernel/Kconfig.kexec | 9 + kernel/Makefile | 1 + kernel/crash_dump_dm_crypt.c | 338 ++++++++++++++++++ kernel/kexec_file.c | 21 ++ kernel/ksysfs.c | 24 ++ 12 files changed, 493 insertions(+), 3 deletions(-) create mode 100644 Documentation/ABI/testing/crash_dm_crypt_keys create mode 100644 kernel/crash_dump_dm_crypt.c
base-commit: 8a92980606e3585d72d510a03b59906e96755b8a
Signature · 1 line
2.45.1
PATCH v5 1/7 kexec_file: allow to place kexec_buf randomly
Currently, kexec_buf is placed in order which means for the same machine, the info in the kexec_buf is always located at the same position each time the machine is booted. This may cause a risk for sensitive information like LUKS volume key. Now struct kexec_buf has a new field random which indicates it's supposed to be placed in a random position.
--- include/linux/kexec.h | 4 ++++ kernel/kexec_file.c | 21 +++++++++++++++++++++ 2 files changed, 25 insertions(+)
include/linux/kexec.h +4 -0
@@ -171,6 +171,7 @@ int kexec_image_post_load_cleanup_default(struct kimage *image);* @buf_min: The buffer can't be placed below this address. * @buf_max: The buffer can't be placed above this address. * @top_down: Allocate from top of memory. +* @random: Place the buffer at a random position. */ struct kexec_buf {struct kimage *image;
@@ -182,6 +183,9 @@ struct kexec_buf {unsigned long buf_min;unsigned long buf_max;bool top_down; +#ifdef CONFIG_CRASH_DUMP +bool random; +#endif };int kexec_load_purgatory (struct kimage *image,struct kexec_buf *kbuf);
kernel/kexec_file.c +21 -0
@@ -25,6 +25,9 @@#include <linux/elfcore.h> #include <linux/kernel.h> #include <linux/kernel_read_file.h> +#ifdef CONFIG_CRASH_DUMP +#include <linux/prandom.h> +#endif #include <linux/syscalls.h> #include <linux/vmalloc.h> #include "kexec_internal.h"
@@ -437,6 +440,18 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,return ret; } +#ifdef CONFIG_CRASH_DUMP +static unsigned long kexec_random_start (unsigned long start,unsigned long end) +{ +unsigned long temp_start; +unsigned short i; + +get_random_bytes (&i,sizeof (unsigned short )); + temp_start = start + (end - start) /USHRT_MAX * i; +return temp_start; +} +#endif +static int locate_mem_hole_top_down (unsigned long start,unsigned long end,struct kexec_buf *kbuf) {
@@ -445,6 +460,10 @@ static int locate_mem_hole_top_down(unsigned long start, unsigned long end, temp_end =min (end, kbuf->buf_max); temp_start = temp_end - kbuf->memsz +1 ; +#ifdef CONFIG_CRASH_DUMP +if (kbuf->random) + temp_start =kexec_random_start (temp_start, temp_end); +#endif do { /* align down start */
@@ -482,6 +501,8 @@ static int locate_mem_hole_bottom_up(unsigned long start, unsigned long end,unsigned long temp_start, temp_end; temp_start =max (start, kbuf->buf_min); +if (kbuf->random) + temp_start =kexec_random_start (temp_start, end); do { temp_start = ALIGN(temp_start, kbuf->buf_align);
Signature · 1 line
2.45.1
kernel test robot
build botRe: [PATCH v5 1/7] kexec_file: allow to place kexec_buf randomly
kernel test robot noticed the following build errors:
Hi Coiby,
kernel test robot noticed the following build errors:
[auto build test ERROR on 8a92980606e3585d72d510a03b59906e96755b8a]
url: https://github.com/intel-lab-lkp/linux/commits/Coiby-Xu/kexec_file-allow-to-place-kexec_buf-randomly/20240607-203154 base: 8a92980606e3585d72d510a03b59906e96755b8a patch link: https://lore.kernel.org/r/20240607122622.167228-2-coxu%40redhat.com patch subject: [PATCH v5 1/7] kexec_file: allow to place kexec_buf randomly config: parisc-allyesconfig (https://download.01.org/0day-ci/archive/20240608/[email protected]/config) compiler: hppa-linux-gcc (GCC) 13.2.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240608/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All errors (new ones prefixed by >>):
kernel/kexec_file.c: In function 'locate_mem_hole_bottom_up':
>> kernel/kexec_file.c:504:17: error: 'struct kexec_buf' has no member named 'random'
504 | if (kbuf->random)
| ^~
>> kernel/kexec_file.c:505:30: error: implicit declaration of function 'kexec_random_start' [-Werror=implicit-function-declaration]
505 | temp_start = kexec_random_start(temp_start, end);
| ^~~~~~~~~~~~~~~~~~cc1: some warnings being treated as errors
vim +504 kernel/kexec_file.c
496
497 static int locate_mem_hole_bottom_up(unsigned long start, unsigned long end,
498 struct kexec_buf *kbuf)
499 {
500 struct kimage *image = kbuf->image;
501 unsigned long temp_start, temp_end;
502
503 temp_start = max(start, kbuf->buf_min);
> 504 if (kbuf->random) > 505 temp_start = kexec_random_start(temp_start, end);
506
507 do {
508 temp_start = ALIGN(temp_start, kbuf->buf_align);
509 temp_end = temp_start + kbuf->memsz - 1;
510
511 if (temp_end > end || temp_end > kbuf->buf_max)
512 return 0;
513 /*
514 * Make sure this does not conflict with any of existing
515 * segments
516 */
517 if (kimage_is_destination_range(image, temp_start, temp_end)) {
518 temp_start = temp_start + PAGE_SIZE;
519 continue;
520 }
521
522 /* We found a suitable memory range */
523 break;
524 } while (1);
525
526 /* If we are here, we found a suitable memory range */
527 kbuf->mem = temp_start;
528
529 /* Success, stop navigating through remaining System RAM ranges */
530 return 1;
531 }
532
Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
kernel test robot
build botRe: [PATCH v5 1/7] kexec_file: allow to place kexec_buf randomly
In reply to Coiby Xukernel test robot noticed the following build errors:
Hi Coiby,
kernel test robot noticed the following build errors:
[auto build test ERROR on 8a92980606e3585d72d510a03b59906e96755b8a]
url: https://github.com/intel-lab-lkp/linux/commits/Coiby-Xu/kexec_file-allow-to-place-kexec_buf-randomly/20240607-203154 base: 8a92980606e3585d72d510a03b59906e96755b8a patch link: https://lore.kernel.org/r/20240607122622.167228-2-coxu%40redhat.com patch subject: [PATCH v5 1/7] kexec_file: allow to place kexec_buf randomly config: x86_64-kexec (https://download.01.org/0day-ci/archive/20240608/[email protected]/config) compiler: clang version 18.1.5 (https://github.com/llvm/llvm-project 617a15a9eac96088ae5e9134248d8236e34b91b1) reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240608/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All errors (new ones prefixed by >>):
>> kernel/kexec_file.c:504:12: error: no member named 'random' in 'struct kexec_buf'
504 | if (kbuf->random)
| ~~~~ ^
>> kernel/kexec_file.c:505:16: error: call to undeclared function 'kexec_random_start'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
505 | temp_start = kexec_random_start(temp_start, end);
| ^2 errors generated.
vim +504 kernel/kexec_file.c
496
497 static int locate_mem_hole_bottom_up(unsigned long start, unsigned long end,
498 struct kexec_buf *kbuf)
499 {
500 struct kimage *image = kbuf->image;
501 unsigned long temp_start, temp_end;
502
503 temp_start = max(start, kbuf->buf_min);
> 504 if (kbuf->random) > 505 temp_start = kexec_random_start(temp_start, end);
506
507 do {
508 temp_start = ALIGN(temp_start, kbuf->buf_align);
509 temp_end = temp_start + kbuf->memsz - 1;
510
511 if (temp_end > end || temp_end > kbuf->buf_max)
512 return 0;
513 /*
514 * Make sure this does not conflict with any of existing
515 * segments
516 */
517 if (kimage_is_destination_range(image, temp_start, temp_end)) {
518 temp_start = temp_start + PAGE_SIZE;
519 continue;
520 }
521
522 /* We found a suitable memory range */
523 break;
524 } while (1);
525
526 /* If we are here, we found a suitable memory range */
527 kbuf->mem = temp_start;
528
529 /* Success, stop navigating through remaining System RAM ranges */
530 return 1;
531 }
532
Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
Greg KH
maintainerRe: [PATCH v5 1/7] kexec_file: allow to place kexec_buf randomly
In reply to Coiby XuQuoted 27 lines · Coiby Xu
Currently,
kexec_bufis placed in order which means for the same machine, the info in thekexec_bufis always located at the same position each time the machine is booted. This may cause a risk for sensitive information like LUKS volume key. Now structkexec_bufhas a new field random which indicates it's supposed to be placed in a random position.Suggested-by Jan Pazdziora <[email protected]>Signed-off-by Coiby Xu <[email protected]>--- include/linux/kexec.h | 4 ++++ kernel/kexec_file.c | 21 +++++++++++++++++++++ 2 files changed, 25 insertions(+)include/linux/kexec.h +1
-0@@ -171,6 +171,7 @@ int kexec_image_post_load_cleanup_default(struct kimage *image);* @buf_min: The buffer can't be placed below this address. * @buf_max: The buffer can't be placed above this address. * @top_down: Allocate from top of memory. +* @random: Place the buffer at a random position. */ struct kexec_buf {struct kimage *image;
include/linux/kexec.h struct kexec_buf { · line 188
Quoted 6 lines
include/linux/kexec.h +3
-0@@ -182,6 +183,9 @@ struct kexec_buf {unsigned long buf_min;unsigned long buf_max;bool top_down; +#ifdef CONFIG_CRASH_DUMP +bool random; +#endif
Why is the ifdef needed?
Quoted 8 lines
include/linux/kexec.h +0
-0};int kexec_load_purgatory (struct kimage *image,struct kexec_buf *kbuf);kernel/kexec_file.c +0
-0
kernel/kexec_file.c line 30
Quoted 6 lines
kernel/kexec_file.c +3
-0@@ -25,6 +25,9 @@#include <linux/elfcore.h> #include <linux/kernel.h> #include <linux/kernel_read_file.h> +#ifdef CONFIG_CRASH_DUMP +#include <linux/prandom.h> +#endif
No ifdef in .c files please. This should not be an issue.
Quoted 8 lines
kernel/kexec_file.c +1
-0#include <linux/syscalls.h> #include <linux/vmalloc.h> #include "kexec_internal.h" @@ -437,6 +440,18 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,return ret; } +#ifdef CONFIG_CRASH_DUMP
kernel/kexec_file.c SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd, · line 453
Quoted 10 lines
kernel/kexec_file.c +10
-0+static unsigned long kexec_random_start (unsigned long start,unsigned long end) +{ +unsigned long temp_start; +unsigned short i; + +get_random_bytes (&i,sizeof (unsigned short )); + temp_start = start + (end - start) /USHRT_MAX * i; +return temp_start; +} +#endif
This #ifdef should be handled properly in the .h file.
kernel/kexec_file.c static int locate_mem_hole_top_down(unsigned long start, unsigned long end, · line 466
Quoted 12 lines
kernel/kexec_file.c +5
-0+static int locate_mem_hole_top_down (unsigned long start,unsigned long end,struct kexec_buf *kbuf) {@@ -445,6 +460,10 @@ static int locate_mem_hole_top_down(unsigned long start, unsigned long end, temp_end =min (end, kbuf->buf_max); temp_start = temp_end - kbuf->memsz +1 ; +#ifdef CONFIG_CRASH_DUMP +if (kbuf->random) + temp_start =kexec_random_start (temp_start, temp_end); +#endif
Same with this.
And why do you need the boolean at all, why not just have kexec_random_start() handle this properly for you?
thanks,
greg k-h
PATCH v5 2/7 crash_dump: make dm crypt keys persist for the kdump kernel
In reply to Coiby XuA sysfs /sys/kernel/crash_dm_crypt_keys is provided for user space to make the dm crypt keys persist for the kdump kernel. Take the case of dumping to a LUKS-encrypted target as an example, here is the life cycle of this kdump copy of LUKS volume keys,
1. After the 1st kernel loads the initramfs during boot, systemd use an user-input passphrase or TPM-sealed key to de-crypt the LUKS volume keys and then save the volume keys to specified keyring (using the --link-vk-to-keyring API) and the key will expire within specified time.
2. A user space tool (kdump initramfs builder) writes a key description to /sys/kernel/crash_dm_crypt_keys to inform the 1st kernel to record the key while building the kdump initramfs
3. The kexec_file_load syscall read the volume keys by recorded key descriptions and then save them key to kdump reserved memory and wipe the copy.
4. When the 1st kernel crashes and the kdump initramfs is booted, the kdump initramfs asks the kdump kernel to create a user key using the key stored in kdump reserved memory by writing to to /sys/kernel/crash_dm_crypt_keys. Then the LUKS encrypted devide is unlocked with libcryptsetup's --volume-key-keyring API.
5. The system gets rebooted to the 1st kernel after dumping vmcore to the LUKS encrypted device is finished
I assume 1) there are 128 LUKS devices at maximum to be unlocked thus MAX_KEY_NUM=128 2) and a key won't exceed 256 bytes thus MAX_KEY_SIZE=256 according to "cryptsetup benchmark".
For details on usage of the API, please check the new doc file Documentation/ABI/testing/crash_dm_crypt_keys.
--- Documentation/ABI/testing/crash_dm_crypt_keys | 30 ++++ include/linux/crash_core.h | 5 +- kernel/Kconfig.kexec | 9 ++ kernel/Makefile | 1 + kernel/crash_dump_dm_crypt.c | 130 ++++++++++++++++++ kernel/ksysfs.c | 24 ++++ 6 files changed, 198 insertions(+), 1 deletion(-) create mode 100644 Documentation/ABI/testing/crash_dm_crypt_keys create mode 100644 kernel/crash_dump_dm_crypt.c
Documentation/ABI/testing/crash_dm_crypt_keys new file +30 -0
@@ -0,0 +1,30 @@ +What: /sys/kernel/crash_dm_crypt_keys +Date: Jun 2024 +KernelVersion: 6.11 +Contact: [email protected] +Description: read/write + Make dm crypt keys persistent for the kdump kernel. + + Assume the key size won't exceed 256 bytes and the maximum number of keys is 128. + + You can write the following commands before kexec'ing the kdump kernel, + - "init KEY_NUM" + Let the kernel know the number of dm crypt keys so it will initialize + needed structures. KEY_NUM=128 dm crypt keys at maximum. + - "record KEY_DESC" + Record a key description. For security reason, the key must be a logon + key whose payload can't be read by user space. For details, please refer + to security/keys/core.rst. + + And you can also read this API to know the command eructation status, + - fresh + Waiting for a command + - initialized + The "init KEY_NUM" command has been executed successfully + - recorded + Specified number of keys have been recorded + - loaded + the kdump kernel has been loaded with the dm crypt keys stored to kdump + reserved memory + +User: Kdump service
include/linux/crash_core.h +4 -1
@@ -34,7 +34,10 @@ static inline void arch_kexec_protect_crashkres(void) { }static inline void arch_kexec_unprotect_crashkres (void ) { }#endif - +#ifdef CONFIG_CRASH_DM_CRYPT +int crash_sysfs_dm_crypt_keys_read (char *buf); +int crash_sysfs_dm_crypt_keys_write (const char *buf,size_t count); +#endif #ifndef arch_crash_handle_hotplug_eventstatic inline void arch_crash_handle_hotplug_event (struct kimage *image,void *arg) { }
kernel/Kconfig.kexec +9 -0
@@ -116,6 +116,15 @@ config CRASH_DUMP For s390, this option also enables zfcpdump. See also <file:Documentation/arch/s390/zfcpdump.rst> +config CRASH_DM_CRYPT + bool "Support saving crash dump to dm-crypt encrypted volume" + depends on CRASH_DUMP + depends on DM_CRYPT + help + With this option enabled, user space can intereact with + /sys/kernel/crash_dm_crypt_keys to make the dm crypt keys + persistent for the crash dump kernel. + config CRASH_HOTPLUG bool "Update the crash elfcorehdr on system configuration changes" default y
kernel/Makefile +1 -0
@@ -72,6 +72,7 @@ obj-$(CONFIG_VMCORE_INFO) += vmcore_info.o elfcorehdr.o obj-$(CONFIG_CRASH_RESERVE) += crash_reserve.o obj-$(CONFIG_KEXEC_CORE) += kexec_core.o obj-$(CONFIG_CRASH_DUMP) += crash_core.o +obj-$(CONFIG_CRASH_DM_CRYPT) += crash_dump_dm_crypt.o obj-$(CONFIG_KEXEC) += kexec.o obj-$(CONFIG_KEXEC_FILE) += kexec_file.o obj-$(CONFIG_KEXEC_ELF) += kexec_elf.o
kernel/crash_dump_dm_crypt.c new file +130 -0
130 lines not on this page · raw message
kernel/ksysfs.c +24 -0
@@ -167,6 +167,27 @@ static ssize_t vmcoreinfo_show(struct kobject *kobj, }KERNEL_ATTR_RO (vmcoreinfo); +#ifdef CONFIG_CRASH_DM_CRYPT +static ssize_t crash_dm_crypt_keys_show (struct kobject *kobj, +struct kobj_attribute *attr,char *buf) +{ +return crash_sysfs_dm_crypt_keys_read (buf); +} + +static ssize_t crash_dm_crypt_keys_store (struct kobject *kobj, +struct kobj_attribute *attr, +const char *buf,size_t count) +{ +int ret; + + ret =crash_sysfs_dm_crypt_keys_write (buf, count); +if (ret <0 ) +return ret; +return count; +} +KERNEL_ATTR_RW (crash_dm_crypt_keys); +#endif /* CONFIG_CRASH_DM_CRYPT */ +#ifdef CONFIG_CRASH_HOTPLUG static ssize_t crash_elfcorehdr_size_show (struct kobject *kobj,struct kobj_attribute *attr,char *buf)
@@ -271,6 +292,9 @@ static struct attribute * kernel_attrs[] = {#endif #ifdef CONFIG_VMCORE_INFO &vmcoreinfo_attr.attr, +#ifdef CONFIG_CRASH_DM_CRYPT + &crash_dm_crypt_keys_attr.attr, +#endif #ifdef CONFIG_CRASH_HOTPLUG &crash_elfcorehdr_size_attr.attr, #endif
Signature · 1 line
2.45.1
Greg KH
maintainerRe: [PATCH v5 2/7] crash_dump: make dm crypt keys persist for the kdump kernel
Quoted 54 lines · Coiby Xu
A sysfs
/sys/kernel/crash_dm_crypt_keysis provided for user space to make the dm crypt keys persist for the kdump kernel. Take the case of dumping to a LUKS-encrypted target as an example, here is the life cycle of this kdump copy of LUKS volume keys,1. After the 1st kernel loads the initramfs during boot, systemd use an user-input passphrase or TPM-sealed key to de-crypt the LUKS volume keys and then save the volume keys to specified keyring (using the --link-vk-to-keyring API) and the key will expire within specified time.
2. A user space tool (kdump initramfs builder) writes a key description to
/sys/kernel/crash_dm_crypt_keysto inform the 1st kernel to record the key while building the kdump initramfs3. The
kexec_file_loadsyscall read the volume keys by recorded key descriptions and then save them key to kdump reserved memory and wipe the copy.4. When the 1st kernel crashes and the kdump initramfs is booted, the kdump initramfs asks the kdump kernel to create a user key using the key stored in kdump reserved memory by writing to to
/sys/kernel/crash_dm_crypt_keys. Then the LUKS encrypted devide is unlocked with libcryptsetup's --volume-key-keyring API.5. The system gets rebooted to the 1st kernel after dumping vmcore to the LUKS encrypted device is finished
I assume 1) there are 128 LUKS devices at maximum to be unlocked thus MAX_KEY_NUM=128 2) and a key won't exceed 256 bytes thus MAX_KEY_SIZE=256 according to "cryptsetup benchmark".
For details on usage of the API, please check the new doc file
Documentation/ABI/testing/crash_dm_crypt_keys.Cc Greg KH <[email protected]>Signed-off-by Coiby Xu <[email protected]>--- Documentation/ABI/testing/crash_dm_crypt_keys | 30 ++++ include/linux/crash_core.h | 5 +- kernel/Kconfig.kexec | 9 ++ kernel/Makefile | 1 + kernel/crash_dump_dm_crypt.c | 130 ++++++++++++++++++ kernel/ksysfs.c | 24 ++++ 6 files changed, 198 insertions(+), 1 deletion(-) create mode 100644 Documentation/ABI/testing/crash_dm_crypt_keys create mode 100644 kernel/crash_dump_dm_crypt.cDocumentation/ABI/testing/crash_dm_crypt_keys new file +0
-0
Documentation/ABI/testing/crash_dm_crypt_keys line 6
Quoted 6 lines
Documentation/ABI/testing/crash_dm_crypt_keys new file +6
-0@@ -0,0 +1,30 @@ +What: /sys/kernel/crash_dm_crypt_keys +Date: Jun 2024 +KernelVersion: 6.11 +Contact: [email protected] +Description: read/write + Make dm crypt keys persistent for the kdump kernel.
Tabs please.
Documentation/ABI/testing/crash_dm_crypt_keys line 8
Quoted 2 lines
Documentation/ABI/testing/crash_dm_crypt_keys new file +2
-0+ + Assume the key size won't exceed 256 bytes and the maximum number of keys is 128.
Please wrap the documentation lines.
And who makes this assumption? The kernel? Userspace? Something else? Who enforces it? What happens if it is not true?
Quoted 10 lines
Documentation/ABI/testing/crash_dm_crypt_keys new file +10
-0+ + You can write the following commands before kexec'ing the kdump kernel, + - "init KEY_NUM" + Let the kernel know the number of dm crypt keys so it will initialize + needed structures. KEY_NUM=128 dm crypt keys at maximum. + - "record KEY_DESC" + Record a key description. For security reason, the key must be a logon + key whose payload can't be read by user space. For details, please refer + to security/keys/core.rst. +
Documentation/ABI/testing/crash_dm_crypt_keys line 28
Quoted 10 lines
Documentation/ABI/testing/crash_dm_crypt_keys new file +10
-0+ And you can also read this API to know the command eructation status, + - fresh + Waiting for a command + - initialized + The "init KEY_NUM" command has been executed successfully + - recorded + Specified number of keys have been recorded + - loaded + the kdump kernel has been loaded with the dm crypt keys stored to kdump + reserved memory
I do not understand, this is a command that you write to this file? That's not how sysfs files work at all.
And this isn't really very well documented, I do not understand if this is read/write or just read or just write or a mix? And what are the valid write values? Valid read? Why "init" vs. "initialized"?
Quoted 7 lines
Documentation/ABI/testing/crash_dm_crypt_keys new file +2
-0+ +User: Kdump serviceinclude/linux/crash_core.h +0
-0
include/linux/crash_core.h static inline void arch_kexec_protect_crashkres(void) { } · line 40
Quoted 8 lines
include/linux/crash_core.h +4
-1@@ -34,7 +34,10 @@ static inline void arch_kexec_protect_crashkres(void) { }static inline void arch_kexec_unprotect_crashkres (void ) { }#endif - +#ifdef CONFIG_CRASH_DM_CRYPT +int crash_sysfs_dm_crypt_keys_read (char *buf); +int crash_sysfs_dm_crypt_keys_write (const char *buf,size_t count); +#endif
No versions for when this is not enabled?
Quoted 41 lines
include/linux/crash_core.h +0
-0#ifndef arch_crash_handle_hotplug_eventstatic inline void arch_crash_handle_hotplug_event (struct kimage *image,void *arg) { }kernel/Kconfig.kexec +9
-0@@ -116,6 +116,15 @@ config CRASH_DUMP For s390, this option also enables zfcpdump. See also <file:Documentation/arch/s390/zfcpdump.rst> +config CRASH_DM_CRYPT + bool "Support saving crash dump to dm-crypt encrypted volume" + depends on CRASH_DUMP + depends on DM_CRYPT + help + With this option enabled, user space can intereact with + /sys/kernel/crash_dm_crypt_keys to make the dm crypt keys + persistent for the crash dump kernel. + config CRASH_HOTPLUG bool "Update the crash elfcorehdr on system configuration changes" default ykernel/Makefile +1
-0@@ -72,6 +72,7 @@ obj-$(CONFIG_VMCORE_INFO) += vmcore_info.o elfcorehdr.o obj-$(CONFIG_CRASH_RESERVE) += crash_reserve.o obj-$(CONFIG_KEXEC_CORE) += kexec_core.o obj-$(CONFIG_CRASH_DUMP) += crash_core.o +obj-$(CONFIG_CRASH_DM_CRYPT) += crash_dump_dm_crypt.o obj-$(CONFIG_KEXEC) += kexec.o obj-$(CONFIG_KEXEC_FILE) += kexec_file.o obj-$(CONFIG_KEXEC_ELF) += kexec_elf.okernel/crash_dump_dm_crypt.c new file +0
-0
kernel/crash_dump_dm_crypt.c line 6
Quoted 6 lines
kernel/crash_dump_dm_crypt.c new file +6
-0@@ -0,0 +1,130 @@ +// SPDX-License-Identifier: GPL-2.0-only +#include <keys/user-type.h> +#include <linux/crash_dump.h> + +#define KEY_NUM_MAX 128/* maximum dm crypt keys */ +#define KEY_SIZE_MAX 256/* maximum dm crypt key size */
Tabs please.
kernel/crash_dump_dm_crypt.c line 9
Quoted 3 lines
kernel/crash_dump_dm_crypt.c new file +3
-0+ +// The key scription has the format: cryptsetup:UUID 11+36+1(NULL)=48 +#define KEY_DESC_LEN 48
"scription"? And is this a string length? Why not use the in-kernel uuid code if you are dealing with uuid values?
Quoted 7 lines
kernel/crash_dump_dm_crypt.c new file +7
-0+ +static enum STATE_ENUM { +FRESH =0 , +INITIALIZED , +RECORDED , +LOADED , +} state;
kernel/crash_dump_dm_crypt.c line 26
Quoted 10 lines
kernel/crash_dump_dm_crypt.c new file +10
-0+ +static const char *const STATE_STR [] = { + [FRESH ] ="fresh" , + [INITIALIZED ] ="initialized" , + [RECORDED ] ="recorded" , + [LOADED ] ="loaded" +}; + +static unsigned int key_count; +static size_t keys_header_size;
So all of the keys have a single header size?
Quoted 77 lines
kernel/crash_dump_dm_crypt.c new file +77
-0+ +struct dm_crypt_key { +unsigned int key_size; +char key_desc[KEY_DESC_LEN ]; +u8 data[KEY_SIZE_MAX ]; +}; + +static struct keys_header { +unsigned int total_keys; + struct dm_crypt_key keys[] __counted_by(total_keys); +} *keys_header; + +static size_t get_keys_header_size (struct keys_header *keys_header, +size_t total_keys) +{ +return struct_size (keys_header, keys, total_keys); +} + +/* +* Let the kernel know the number of dm crypt keys and allocate memory to +* initialize related structures. +*/ +static int init (const char *buf) +{ +unsigned int total_keys; + +if (sscanf (buf,"init %u" , &total_keys) !=1 ) +return -EINVAL ; + +if (total_keys >KEY_NUM_MAX ) { +kexec_dprintk ( +"Exceed the maximum number of keys (KEY_NUM_MAX=%u)\n" , +KEY_NUM_MAX ); +return -EINVAL ; + } + + keys_header_size =get_keys_header_size (keys_header, total_keys); + key_count =0 ; + +if (keys_header !=NULL ) +kvfree (keys_header); + + keys_header =kzalloc (keys_header_size,GFP_KERNEL ); +if (!keys_header) +return -ENOMEM ; + + keys_header->total_keys = total_keys; + state =INITIALIZED ; +return 0 ; +} + +/* +* Record the key description of a dm crypt key. +*/ +static int record_key_desc (const char *buf,struct dm_crypt_key *dm_key) +{ +char key_desc[KEY_DESC_LEN ]; + +if (state !=INITIALIZED ) { +kexec_dprintk ("Please send the cmd 'init <KEY_NUM>' first\n" ); +return -EINVAL ; + } + +if (sscanf (buf,"record %s" , key_desc) !=1 ) +return -EINVAL ; + +if (key_count >= keys_header->total_keys) { +kexec_dprintk ("Already have %u keys" , key_count); +return -EINVAL ; + } + +strscpy (dm_key->key_desc, key_desc,KEY_DESC_LEN ); +kexec_dprintk ("Key%d (%s) recorded\n" , key_count, dm_key->key_desc); + key_count++; + +if (key_count == keys_header->total_keys) + state =RECORDED ;
kernel/crash_dump_dm_crypt.c line 113
Quoted 10 lines
kernel/crash_dump_dm_crypt.c new file +10
-0+ +return 0 ; +} + +static int process_cmd (const char *buf,size_t count) +{ +if (strncmp (buf,"init " ,5 ) ==0 ) +return init (buf); +else if (strncmp (buf,"record " ,7 ) ==0 && count ==KEY_DESC_LEN +6 ) +return record_key_desc (buf, &keys_header->keys[key_count]);
So you are parsing the "init" and "record" strings twice? That feels wrong, and might get out of sync easily.
But again, having a sysfs file that does "actions" like this is odd too, as you have multiple things in the same string. sysfs is "one value per file", why aren't you using configfs instead? That's meant for configuring stuff, not sysfs.
Quoted 32 lines
kernel/crash_dump_dm_crypt.c new file +17
-0+ +return -EINVAL ; +} + +int crash_sysfs_dm_crypt_keys_write (const char *buf,size_t count) +{ +if (!is_kdump_kernel ()) +return process_cmd (buf, count); +return -EINVAL ; +} +EXPORT_SYMBOL_GPL (crash_sysfs_dm_crypt_keys_write); + +int crash_sysfs_dm_crypt_keys_read (char *buf) +{ +return sysfs_emit (buf,"%s\n" ,STATE_STR [state]); +} +EXPORT_SYMBOL_GPL (crash_sysfs_dm_crypt_keys_read);kernel/ksysfs.c +7
-0@@ -167,6 +167,27 @@ static ssize_t vmcoreinfo_show(struct kobject *kobj, }KERNEL_ATTR_RO (vmcoreinfo); +#ifdef CONFIG_CRASH_DM_CRYPT +static ssize_t crash_dm_crypt_keys_show (struct kobject *kobj, +struct kobj_attribute *attr,char *buf) +{ +return crash_sysfs_dm_crypt_keys_read (buf); +} +
kernel/ksysfs.c static ssize_t vmcoreinfo_show(struct kobject *kobj, · line 186
Quoted 10 lines
kernel/ksysfs.c +10
-0+static ssize_t crash_dm_crypt_keys_store (struct kobject *kobj, +struct kobj_attribute *attr, +const char *buf,size_t count) +{ +int ret; + + ret =crash_sysfs_dm_crypt_keys_write (buf, count); +if (ret <0 ) +return ret; +return count;
Why doesn't crash_sysfs_dm_crypt_keys_write just return the count value processed? No need to check the return value again.
thanks
greg k-h
PATCH v5 3/7 crash_dump: store dm crypt keys in kdump reserved memory
In reply to Coiby XuWhen the kdump kernel image and initrd are loaded, the dm crypts keys will be read from keyring and then stored in kdump reserved memory.
--- include/linux/crash_core.h | 3 ++ include/linux/crash_dump.h | 2 + include/linux/kexec.h | 4 ++ kernel/crash_dump_dm_crypt.c | 87 ++++++++++++++++++++++++++++++++++++ 4 files changed, 96 insertions(+)
include/linux/crash_core.h +3 -0
@@ -37,6 +37,9 @@ static inline void arch_kexec_unprotect_crashkres(void) { }#ifdef CONFIG_CRASH_DM_CRYPT int crash_sysfs_dm_crypt_keys_read (char *buf);int crash_sysfs_dm_crypt_keys_write (const char *buf,size_t count); +int crash_load_dm_crypt_keys (struct kimage *image); +#else +static inline int crash_load_dm_crypt_keys (struct kimage *image) {return 0 ; }#endif #ifndef arch_crash_handle_hotplug_event
include/linux/crash_dump.h +2 -0
@@ -15,6 +15,8 @@extern unsigned long long elfcorehdr_addr;extern unsigned long long elfcorehdr_size; +extern unsigned long long dm_crypt_keys_addr; +#ifdef CONFIG_CRASH_DUMP extern int elfcorehdr_alloc (unsigned long long *addr,unsigned long long *size);extern void elfcorehdr_free (unsigned long long addr);
include/linux/kexec.h +4 -0
@@ -372,6 +372,10 @@ struct kimage {void *elf_headers;unsigned long elf_headers_sz;unsigned long elf_load_addr; + +/* dm crypt keys buffer */ +unsigned long dm_crypt_keys_addr; +unsigned long dm_crypt_keys_sz; };/* kexec interface functions */
kernel/crash_dump_dm_crypt.c +87 -0
94 lines not on this page · raw message
Signature · 1 line
2.45.1
PATCH v5 4/7 crash_dump: reuse saved dm crypt keys for CPU/memory hot-plugging
In reply to Coiby XuWhen there are CPU and memory hot un/plugs, the crash elfcorehdr which describes CPUs and memory in the system needs to be updated for the kdump kernel.
Currently, there are two solutions to support this case. One is to utilizes udev to instruct user space to reload the kdump kernel image and initrd, elfcorehdr and etc. again. The other is to only update the elfcorehdr segment. For the 1st solution, the dm crypt keys need to be reloaded again. The user space can write the "reuse" command to /sys/kernel/crash_dm_crypt_key so the stored keys can be re-saved again.
Note only x86 (commit ea53ad9cf73b ("x86/crash: add x86 crash hotplug support")) and ppc (WIP) supports the new infrastructure (commit 247262756121 ("crash: add generic infrastructure for crash hotplug support")). If the new infrastructure get extended to all arches, this patch can be dropped.
--- Documentation/ABI/testing/crash_dm_crypt_keys | 2 ++ kernel/crash_dump_dm_crypt.c | 32 ++++++++++++++++--- 2 files changed, 30 insertions(+), 4 deletions(-)
Documentation/ABI/testing/crash_dm_crypt_keys +2 -0
@@ -15,6 +15,8 @@ Description: read/write Record a key description. For security reason, the key must be a logon key whose payload can't be read by user space. For details, please refer to security/keys/core.rst. + - "reuse" + Reuse the dm crypt keys stored in kdump reserved memory. And you can also read this API to know the command eructation status, - fresh
kernel/crash_dump_dm_crypt.c +28 -4
@@ -15,13 +15,15 @@ static enum STATE_ENUM {INITIALIZED ,RECORDED ,LOADED , +REUSE , } state;static const char *const STATE_STR [] = { [FRESH ] ="fresh" , [INITIALIZED ] ="initialized" , [RECORDED ] ="recorded" , - [LOADED ] ="loaded" + [LOADED ] ="loaded" , + [REUSE ] ="reuse" };static unsigned int key_count;
@@ -107,12 +109,32 @@ static int record_key_desc(const char *buf, struct dm_crypt_key *dm_key)return 0 ; } +static void get_keys_from_kdump_reserved_memory (void ) +{ +struct keys_header *keys_header_loaded; + +arch_kexec_unprotect_crashkres (); + + keys_header_loaded =kmap_local_page (pfn_to_page ( + kexec_crash_image->dm_crypt_keys_addr >>PAGE_SHIFT )); + +memcpy (keys_header, keys_header_loaded, keys_header_size); +kunmap_local (keys_header_loaded); + state =RECORDED ; +arch_kexec_protect_crashkres (); +} +static int process_cmd (const char *buf,size_t count) {if (strncmp (buf,"init " ,5 ) ==0 )return init (buf);else if (strncmp (buf,"record " ,7 ) ==0 && count ==KEY_DESC_LEN +6 )return record_key_desc (buf, &keys_header->keys[key_count]); +else if (!strcmp (buf,"reuse" )) { + state =REUSE ; +get_keys_from_kdump_reserved_memory (); +return 0 ; + }return -EINVAL ; }
@@ -192,9 +214,11 @@ int crash_load_dm_crypt_keys(struct kimage *image) } image->dm_crypt_keys_addr =0 ; - r =build_keys_header (); -if (r) -return r; +if (state !=REUSE ) { + r =build_keys_header (); +if (r) +return r; + } kbuf.buffer = keys_header; kbuf.bufsz = keys_header_size;
Signature · 1 line
2.45.1
PATCH v5 5/7 crash_dump: retrieve dm crypt keys in kdump kernel
In reply to Coiby XuCrash kernel will retrieve the dm crypt keys based on the dmcryptkeys command line parameter. When user space writes the key description to /sys/kernel/crash_dm_crypt_key, the crash kernel will save the encryption keys to the user keyring. Then user space e.g. cryptsetup's --volume-key-keyring API can use it to unlock the encrypted device.
--- Documentation/ABI/testing/crash_dm_crypt_keys | 3 + include/linux/crash_core.h | 1 + kernel/crash_dump_dm_crypt.c | 101 +++++++++++++++++- 3 files changed, 103 insertions(+), 2 deletions(-)
Documentation/ABI/testing/crash_dm_crypt_keys +3 -0
@@ -29,4 +29,7 @@ Description: read/write the kdump kernel has been loaded with the dm crypt keys stored to kdump reserved memory + After the kdump kernel gets booted, user space can write anything to this API + so the dm crypt keys can be restored to the keyring. + User: Kdump service
include/linux/crash_core.h +1 -0
@@ -38,6 +38,7 @@ static inline void arch_kexec_unprotect_crashkres(void) { }int crash_sysfs_dm_crypt_keys_read (char *buf);int crash_sysfs_dm_crypt_keys_write (const char *buf,size_t count);int crash_load_dm_crypt_keys (struct kimage *image); +ssize_t dm_crypt_keys_read (char *buf,size_t count,u64 *ppos);#else static inline int crash_load_dm_crypt_keys (struct kimage *image) {return 0 ; }#endif
kernel/crash_dump_dm_crypt.c +99 -2
129 lines not on this page · raw message
Signature · 1 line
2.45.1
PATCH v5 6/7 x86/crash: pass dm crypt keys to kdump kernel
In reply to Coiby Xu1st kernel will build up the kernel command parameter dmcryptkeys as similar to elfcorehdr to pass the memory address of the stored info of dm crypt key to kdump kernel.
--- arch/x86/kernel/crash.c | 20 ++++++++++++++++++-- arch/x86/kernel/kexec-bzimage64.c | 7 +++++++ 2 files changed, 25 insertions(+), 2 deletions(-)
arch/x86/kernel/crash.c +18 -2
@@ -266,6 +266,7 @@ static int memmap_exclude_ranges(struct kimage *image, struct crash_mem *cmem, unsigned long long mend) {unsigned long start, end; +int ret; cmem->ranges[0 ].start = mstart; cmem->ranges[0 ].end = mend;
@@ -274,22 +275,37 @@ static int memmap_exclude_ranges(struct kimage *image, struct crash_mem *cmem,/* Exclude elf header region */ start = image->elf_load_addr; end = start + image->elf_headers_sz -1 ; -return crash_exclude_mem_range (cmem, start, end); + ret =crash_exclude_mem_range (cmem, start, end); + +if (ret) +return ret; + +/* Exclude dm crypt keys region */ +if (image->dm_crypt_keys_addr) { + start = image->dm_crypt_keys_addr; + end = start + image->dm_crypt_keys_sz -1 ; +return crash_exclude_mem_range (cmem, start, end); + } + +return ret; }/* Prepare memory map for crash dump kernel */ int crash_setup_memmap_entries (struct kimage *image,struct boot_params *params) { +unsigned int max_nr_ranges =3 ;int i, ret =0 ;unsigned long flags;struct e820_entry ei;struct crash_memmap_data cmd;struct crash_mem *cmem; - cmem =vzalloc (struct_size (cmem, ranges,1 )); + cmem =vzalloc (struct_size (cmem, ranges, max_nr_ranges));if (!cmem)return -ENOMEM ; + cmem->max_nr_ranges = max_nr_ranges; +memset (&cmd,0 ,sizeof (struct crash_memmap_data )); cmd.params = params;
arch/x86/kernel/kexec-bzimage64.c +7 -0
@@ -76,6 +76,10 @@ static int setup_cmdline(struct kimage *image, struct boot_params *params,if (image->type ==KEXEC_TYPE_CRASH ) { len =sprintf (cmdline_ptr,"elfcorehdr=0x%lx " , image->elf_load_addr); + +if (image->dm_crypt_keys_addr !=0 ) + len +=sprintf (cmdline_ptr + len, +"dmcryptkeys=0x%lx " , image->dm_crypt_keys_addr); }memcpy (cmdline_ptr + len, cmdline, cmdline_len); cmdline_len += len;
@@ -441,6 +445,9 @@ static void *bzImage64_load(struct kimage *image, char *kernel, ret =crash_load_segments (image);if (ret)return ERR_PTR (ret); + ret =crash_load_dm_crypt_keys (image); +if (ret) + pr_debug("Either no dm crypt key or error to retrieve the dm crypt key\n"); }#endif
Signature · 1 line
2.45.1
PATCH v5 7/7 x86/crash: make the page that stores the dm crypt keys inaccessible
In reply to Coiby XuThis adds an addition layer of protection for the saved copy of dm crypt key. Trying to access the saved copy will cause page fault.
--- arch/x86/kernel/machine_kexec_64.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+)
arch/x86/kernel/machine_kexec_64.c +22 -0
@@ -545,13 +545,35 @@ static void kexec_mark_crashkres(bool protect) kexec_mark_range(control, crashk_res.end, protect); } +/* make the memory storing dm crypt keys in/accessible */ +static void kexec_mark_dm_crypt_keys (bool protect) +{ +unsigned long start_paddr, end_paddr; +unsigned int nr_pages; + +if (kexec_crash_image->dm_crypt_keys_addr) { + start_paddr = kexec_crash_image->dm_crypt_keys_addr; + end_paddr = start_paddr + kexec_crash_image->dm_crypt_keys_sz -1 ; + nr_pages = (PAGE_ALIGN (end_paddr) -PAGE_ALIGN_DOWN (start_paddr))/PAGE_SIZE ; +if (protect) +set_memory_np ((unsigned long )phys_to_virt (start_paddr), nr_pages); +else +__set_memory_prot ( + (unsigned long )phys_to_virt (start_paddr), + nr_pages, +__pgprot (_PAGE_PRESENT | _PAGE_NX | _PAGE_RW)); + } +} +void arch_kexec_protect_crashkres (void ) {kexec_mark_crashkres (true); +kexec_mark_dm_crypt_keys (true); }void arch_kexec_unprotect_crashkres (void ) { +kexec_mark_dm_crypt_keys (false);kexec_mark_crashkres (false); }#endif
Signature · 1 line
2.45.1
The subject prefix should be "[PATCH v5 0/7]". I'm sorry if it causes any confusion.
Quoted 170 lines · Coiby Xu
LUKS is the standard for Linux disk encryption. Many users choose LUKS and in some use cases like Confidential VM it's mandated. With kdump enabled, when the 1st kernel crashes, the system could boot into the kdump/crash kernel and dump the memory image i.e.
/proc/vmcoreto a specified target. Currently, when dumping vmcore to a LUKS encrypted device, there are two problems,- Kdump kernel may not be able to decrypt the LUKS partition. For some machines, a system administrator may not have a chance to enter the password to decrypt the device in kdump initramfs after the 1st kernel crashes; For cloud confidential VMs, depending on the policy the kdump kernel may not be able to unseal the keys with TPM and the console virtual keyboard is untrusted.
- LUKS2 by default use the memory-hard Argon2 key derivation function which is quite memory-consuming compared to the limited memory reserved for kdump. Take Fedora example, by default, only 256M is reserved for systems having memory between 4G-64G. With LUKS enabled, ~1300M needs to be reserved for kdump. Note if the memory reserved for kdump can't be used by 1st kernel i.e. an user sees ~1300M memory missing in the 1st kernel.
Besides users (at least for Fedora) usually expect kdump to work out of the box i.e. no manual password input is needed. And it doesn't make sense to derivate the keys again in kdump kernel which seems to be redundant work.
This patch set addresses the above issues by make the LUKS volume keys persistent for kdump kernel with the help of cryptsetup's new APIs (--link-vk-to-keyring/--volume-key-keyring). Here is the life cycle of this kdump copy of LUKS volume keys,
1. After the 1st kernel loads the initramfs during boot, systemd use an user-input passphrase or TPM-sealed key to de-crypt the LUKS volume keys and then save the volume keys to specified keyring (using the --link-vk-to-keyring API) and the key will expire within specified time.
2. A user space tool (kdump initramfs builder) writes a key description to
/sys/kernel/crash_dm_crypt_keysto inform the 1st kernel to record the key while building the kdump initramfs3. The
kexec_file_loadsyscall read the volume keys by recorded key descriptions and then save them key to kdump reserved memory and wipe the copy.4. When the 1st kernel crashes and the kdump initramfs is booted, the kdump initramfs asks the kdump kernel to create a user key using the key stored in kdump reserved memory by writing to to
/sys/kernel/crash_dm_crypt_keys. Then the LUKS encrypted devide is unlocked with libcryptsetup's --volume-key-keyring API.5. The system gets rebooted to the 1st kernel after dumping vmcore to the LUKS encrypted device is finished
After libcryptsetup saving the LUKS volume keys to specified keyring, whoever takes this should be responsible for the safety of these copies of keys. The keys will be saved in the memory area exclusively reserved for kdump where even the 1st kernel has no direct access. And further more, two additional protections are added,
- save the copy randomly in kdump reserved memory as suggested by Jan
- clear the_PAGE_PRESENTflag of the page that stores the copy as suggested by PingfanThis patch set only supports x86. There will be patches to support other architectures once this patch set gets merged.
v5
- Baoquan
- limit the feature of placingkexec_bufrandomly to kdump (CONFIG_CRASH_DUMP)
- add documentation for added sysfs API
- allow to re-send init command to support the case of user switching to a different LUKS-encrypted target
- makeCONFIG_CRASH_DM_CRYPTdepends onCONFIG_DM_CRYPT
- check if the number of keys exceedKEY_NUM_MAX
- rename (struct keys_header).key_count as (struct keys_header).total_keys to improve code readiblity
- improve commit message
- fix the failure of callingcrash_exclude_mem_range(there is a split ofmem_range)
- use ret instead of r as return code- Greg
- add documentation for added sysfs API
- avoid spamming kernel logs
- fix a buffer overflow issue
- keep the state enums synced up with the string values
- usesysfs_emitother than sprintf
- explainKEY_NUM_MAXandKEY_SIZE_MAX
- s/EXPORT_SYMBOL_GPL/EXPORT_SYMBOL/g
- improve code readability- Rebase onto latest Linus tree
v4
- rebase onto latest Linus tree so Baoquan can apply the patches for code review
- fix kernel test robot warningsv3
- Support CPU/memory hot-plugging [Baoquan]
- Don't save the keys temporarily to simplify the implementation [Baoquan]
- Support multiple LUKS encrypted volumes
- Read logon key instead of user key to improve security [Ondrej]
- A kernel config optionCRASH_DM_CRYPTfor this feature (disabled by default)
- Fix warnings found by kernel test robot
- Rebase the code onto 6.9.0-rc5+v2
- work together with libscryptsetup's --link-vk-to-keyring/--volume-key-keyring APIs [Milan and Ondrej]
- add the case where console virtual keyboard is untrusted for confidential VM
- usedm_crypt_keyinstead of LUKS volume key [Milan and Eric]
- fix some code format issues
- don't move "structkexec_segment" declaration
- Rebase the code onto latest Linus tree (6.7.0)v1
- "Put the luks key handling related tocrash_dumpout into a separate filekernel/crash_dump_luks.c" [Baoquan]
- Put the generic luks handling code before the x86 specific code to make it easier for other arches to follow suit [Baoquan]
- Usephys_to_virtinstead of "pfn -> page -> vaddr" [Dave Hansen]
- Drop the RFC prefix [Dave Young]
- Rebase the code onto latest Linus tree (6.4.0-rc4)RFC v2
- libcryptsetup interacts with the kernel via sysfs instead of "hacking" dm-crypt
- to save a kdump copy of the LUKS volume key in 1st kernel
- to add a logon key using the copy for libcryptsetup in kdump kernel [Milan]
- to avoid the incorrect usage of LUKS master key in dm-crypt [Milan]
- save the kdump copy of LUKS volume key randomly [Jan]
- mark the kdump copy inaccessible [Pingfan]
- Miscellaneous
- explain when operations related to the LUKS volume key happen [Jan]
- s/master key/volume key/g
- usecrash_instead ofkexec_as function prefix
- fix commit subject prefixes e.g. "x86, kdump" to x86/crashCoiby Xu (7):
kexec_file: allow to placekexec_bufrandomly
crash_dump: make dm crypt keys persist for the kdump kernel
crash_dump: store dm crypt keys in kdump reserved memory
crash_dump: reuse saved dm crypt keys for CPU/memory hot-plugging
crash_dump: retrieve dm crypt keys in kdump kernel x86/crash: pass dm crypt keys to kdump kernel
x86/crash: make the page that stores the dm crypt keys inaccessibleDocumentation/ABI/testing/crash_dm_crypt_keys | 35 ++ arch/x86/kernel/crash.c | 20 +- arch/x86/kernel/kexec-bzimage64.c | 7 + arch/x86/kernel/machine_kexec_64.c | 22 ++ include/linux/crash_core.h | 9 +- include/linux/crash_dump.h | 2 + include/linux/kexec.h | 8 + kernel/Kconfig.kexec | 9 + kernel/Makefile | 1 + kernel/crash_dump_dm_crypt.c | 338 ++++++++++++++++++ kernel/kexec_file.c | 21 ++ kernel/ksysfs.c | 24 ++ 12 files changed, 493 insertions(+), 3 deletions(-)create mode 100644
Documentation/ABI/testing/crash_dm_crypt_keyscreate mode 100644kernel/crash_dump_dm_crypt.cbase-commit:
8a92980606e3585d72d510a03b59906e96755b8aSignature · 1 line
2.45.1
Signature · 2 lines
Best regards, Coiby
Greg KH
maintainerQuoted 2 lines · Coiby Xu
The subject prefix should be "[PATCH v5 0/7]". I'm sorry if it causes any confusion.
It will probably break our tools (b4) that want to look this up, so you might want to fix that for the next version you send out, given that this one has some 0-day bot issues.
thanks,
greg k-h