Skip to contentSkip to conversation

Split crash out from kexec and clean up related config items

PATCH In review Baoquan He 7 participants 37 messages Jan 24, 2024 – Aug 23, 2024

Subsystems (25)
Arm (32-bit) x86 VFS PowerPC +21 more subsystems
lore
Share & tools
lore

Baoquan He

author
#0 lore raw

v3 cover Split crash out from kexec and clean up related config items

Motivation:
=============
Previously, LKP reported a building error. When investigating, it can't be resolved reasonablly with the present messy kdump config items.

https://lore.kernel.org/oe-kbuild-all/[email protected]/

The kdump (crash dumping) related config items could causes confusions:

Firstly,
---
CRASH_CORE enables codes including
- crashkernel reservation;
- elfcorehdr updating;
- vmcoreinfo exporting;
- crash hotplug handling;

Now fadump of powerpc, kcore dynamic debugging and kdump all selects CRASH_CORE, while fadump
- fadump needs crashkernel parsing, vmcoreinfo exporting, and accessing global variable 'elfcorehdr_addr';
- kcore only needs vmcoreinfo exporting;
- kdump needs all of the current kernel/crash_core.c.

So only enabling PROC_CORE or FA_DUMP will enable CRASH_CORE, this mislead people that we enable crash dumping, actual it's not.

Secondly,
---
It's not reasonable to allow KEXEC_CORE select CRASH_CORE.

Because KEXEC_CORE enables codes which allocate control pages, copy kexec/kdump segments, and prepare for switching. These codes are shared by both kexec reboot and kdump. We could want kexec reboot, but disable kdump. In that case, CRASH_CORE should not be selected.

--------------------
CONFIG_CRASH_CORE=y
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC=y
CONFIG_KEXEC_FILE=y
---------------------

Thirdly,
---
It's not reasonable to allow CRASH_DUMP select KEXEC_CORE.

That could make KEXEC_CORE, CRASH_DUMP are enabled independently from KEXEC or KEXEC_FILE. However, w/o KEXEC or KEXEC_FILE, the KEXEC_CORE code built in doesn't make any sense because no kernel loading or switching will happen to utilize the KEXEC_CORE code.
---------------------
CONFIG_CRASH_CORE=y
CONFIG_KEXEC_CORE=y
CONFIG_CRASH_DUMP=y
---------------------

In this case, what is worse, on arch sh and arm, KEXEC relies on MMU, while CRASH_DUMP can still be enabled when !MMU, then compiling error is seen as the lkp test robot reported in above link.

------arch/sh/Kconfig------
config ARCH_SUPPORTS_KEXEC
def_bool MMU

config ARCH_SUPPORTS_CRASH_DUMP
def_bool BROKEN_ON_SMP
---------------------------

Changes:
===========
1, split out crash_reserve.c from crash_core.c;
2, split out vmcore_infoc. from crash_core.c;
3, move crash related codes in kexec_core.c into crash_core.c; 4, remove dependency of FA_DUMP on CRASH_DUMP;
5, clean up kdump related config items;
6, wrap up crash codes in crash related ifdefs on all 8 arch-es
which support crash dumping, except of ppc;

Achievement:
===========
With above changes, I can rearrange the config item logic as below (the right item depends on or is selected by the left item):

    PROC_KCORE -----------> VMCORE_INFO

               |----------> VMCORE_INFO
    FA_DUMP----|
               |----------> CRASH_RESERVE

                                                    ---->VMCORE_INFO
                                                   /
                                                   |---->CRASH_RESERVE
    KEXEC      --|                                /|
                 |--> KEXEC_CORE--> CRASH_DUMP-->/-|---->PROC_VMCORE
    KEXEC_FILE --|                               \ |
                                                   \---->CRASH_HOTPLUG


    KEXEC      --|
                 |--> KEXEC_CORE (for kexec reboot only)
    KEXEC_FILE --|

Test
========
On all 8 architectures, including x86_64, arm64, s390x, sh, arm, mips, riscv, loongarch, I did below three cases of config item setting and building all passed. Take configs on x86_64 as exampmle here:

(1) Both CONFIG_KEXEC and KEXEC_FILE is unset, then all kexec/kdump items are unset automatically:
# Kexec and crash features
# CONFIG_KEXEC is not set
# CONFIG_KEXEC_FILE is not set

# end of Kexec and crash features

(2) set CONFIG_KEXEC_FILE and 'make olddefconfig':
---------------
# Kexec and crash features CONFIG_CRASH_RESERVE=y CONFIG_VMCORE_INFO=y
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC_FILE=y
CONFIG_CRASH_DUMP=y
CONFIG_CRASH_HOTPLUG=y CONFIG_CRASH_MAX_MEMORY_RANGES=8192

# end of Kexec and crash features
---------------

(3) unset CONFIG_CRASH_DUMP in case 2 and execute 'make olddefconfig':
------------------------
# Kexec and crash features
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC_FILE=y

# end of Kexec and crash features
------------------------

Note:
For ppc, it needs investigation to make clear how to split out crash code in arch folder. Hope Hari and Pingfan can help have a look, see if it's doable. Now, I make it either have both kexec and crash enabled, or disable both of them altogether.

Changelog
==========
v2->v3:
- In patch 2, there's conflict when rebasing to linux-next in kernel/crash_core.c because of below commits from Uladzislau:
- commit 699d9351822e ("mm: vmalloc: Fix a warning in the crash_save_vmcoreinfo_init()")
- commit 5f4c0c1e2a51 (:mm/vmalloc: remove vmap_area_list")
- In patch 13, fix the lkp reported issue by using CONFIG_CRASH_RESERVE
ifdef, giving up the earlier IS_ENABLED(CONFIG_CRASH_RESERVE) checking in v2.
- In patch 14, update code change after below commit merged into
mainline:
- commit 78de91b45860 ("LoongArch: Use generic interface to support crashkernel=X,[high,low]")

Baoquan He (14):
kexec: split crashkernel reservation code out from crash_core.c
crash: split vmcoreinfo exporting code out from crash_core.c
crash: remove dependency of FA_DUMP on CRASH_DUMP
crash: split crash dumping code out from kexec_core.c
crash: clean up kdump related config items
x86, crash: wrap crash dumping code into crash related ifdefs arm64, crash: wrap crash dumping code into crash related ifdefs ppc, crash: enforce KEXEC and KEXEC_FILE to select CRASH_DUMP s390, crash: wrap crash dumping code into crash related ifdefs sh, crash: wrap crash dumping code into crash related ifdefs mips, crash: wrap crash dumping code into crash related ifdefs riscv, crash: wrap crash dumping code into crash related ifdefs arm, crash: wrap crash dumping code into crash related ifdefs loongarch, crash: wrap crash dumping code into crash related ifdefs

 arch/arm/kernel/setup.c                       |   4 +-
 arch/arm64/Kconfig                            |   2 +-
 .../asm/{crash_core.h => crash_reserve.h}     |   4 +-
 arch/arm64/include/asm/kexec.h                |   2 +-
 arch/arm64/kernel/Makefile                    |   2 +-
 arch/arm64/kernel/machine_kexec.c             |   2 +-
 arch/arm64/kernel/machine_kexec_file.c        |  10 +-
 .../kernel/{crash_core.c => vmcore_info.c}    |   2 +-
 arch/arm64/mm/init.c                          |   2 +-
 arch/loongarch/kernel/setup.c                 |   2 +-
 arch/mips/kernel/setup.c                      |  17 +-
 arch/powerpc/Kconfig                          |   9 +-
 arch/powerpc/kernel/setup-common.c            |   2 +-
 arch/powerpc/mm/nohash/kaslr_booke.c          |   4 +-
 arch/powerpc/platforms/powernv/opal-core.c    |   2 +-
 arch/riscv/Kconfig                            |   2 +-
 .../asm/{crash_core.h => crash_reserve.h}     |   4 +-
 arch/riscv/kernel/Makefile                    |   2 +-
 arch/riscv/kernel/elf_kexec.c                 |   9 +-
 .../kernel/{crash_core.c => vmcore_info.c}    |   2 +-
 arch/riscv/mm/init.c                          |   2 +-
 arch/s390/kernel/kexec_elf.c                  |   2 +
 arch/s390/kernel/kexec_image.c                |   2 +
 arch/s390/kernel/machine_kexec_file.c         |  10 +
 arch/sh/kernel/machine_kexec.c                |   3 +
 arch/sh/kernel/setup.c                        |   2 +-
 arch/x86/Kconfig                              |   2 +-
 .../asm/{crash_core.h => crash_reserve.h}     |   6 +-
 arch/x86/kernel/Makefile                      |   6 +-
 arch/x86/kernel/cpu/mshyperv.c                |   4 +
 arch/x86/kernel/kexec-bzimage64.c             |   4 +
 arch/x86/kernel/kvm.c                         |   4 +-
 arch/x86/kernel/machine_kexec_64.c            |   3 +
 arch/x86/kernel/reboot.c                      |   2 +-
 arch/x86/kernel/setup.c                       |   2 +-
 arch/x86/kernel/smp.c                         |   2 +-
 .../{crash_core_32.c => vmcore_info_32.c}     |   2 +-
 .../{crash_core_64.c => vmcore_info_64.c}     |   2 +-
 arch/x86/xen/enlighten_hvm.c                  |   4 +
 drivers/base/cpu.c                            |   6 +-
 drivers/firmware/qemu_fw_cfg.c                |  14 +-
 fs/proc/Kconfig                               |   2 +-
 fs/proc/kcore.c                               |   2 +-
 include/linux/buildid.h                       |   2 +-
 include/linux/crash_core.h                    | 152 ++--
 include/linux/crash_reserve.h                 |  48 ++
 include/linux/kexec.h                         |  47 +-
 include/linux/vmcore_info.h                   |  81 ++
 init/initramfs.c                              |   2 +-
 kernel/Kconfig.kexec                          |  12 +-
 kernel/Makefile                               |   5 +-
 kernel/crash_core.c                           | 762 +++++-------------
 kernel/crash_reserve.c                        | 464 +++++++++++
 kernel/{crash_dump.c => elfcorehdr.c}         |   0
 kernel/kexec.c                                |  11 +-
 kernel/kexec_core.c                           | 250 +-----
 kernel/kexec_file.c                           |  13 +-
 kernel/kexec_internal.h                       |   2 +
 kernel/ksysfs.c                               |  10 +-
 kernel/printk/printk.c                        |   4 +-
 kernel/vmcore_info.c                          | 231 ++++++
 lib/buildid.c                                 |   2 +-
 62 files changed, 1228 insertions(+), 1043 deletions(-)
 rename arch/arm64/include/asm/{crash_core.h => crash_reserve.h} (81%)
 rename arch/arm64/kernel/{crash_core.c => vmcore_info.c} (97%)
 rename arch/riscv/include/asm/{crash_core.h => crash_reserve.h} (78%)
 rename arch/riscv/kernel/{crash_core.c => vmcore_info.c} (96%)
 rename arch/x86/include/asm/{crash_core.h => crash_reserve.h} (92%)
 rename arch/x86/kernel/{crash_core_32.c => vmcore_info_32.c} (90%)
 rename arch/x86/kernel/{crash_core_64.c => vmcore_info_64.c} (94%)
 create mode 100644 include/linux/crash_reserve.h
 create mode 100644 include/linux/vmcore_info.h
 create mode 100644 kernel/crash_reserve.c
 rename kernel/{crash_dump.c => elfcorehdr.c} (100%)
 create mode 100644 kernel/vmcore_info.c
Signature · 1 line
2.41.0
author
#1 lore raw

PATCH v3 1/14 kexec: split crashkernel reservation code out from crash_core.c

Both kdump and fa_dump of ppc rely on crashkernel reservation. Move the relevant codes into separate files:
crash_reserve.c, include/linux/crash_reserve.h.

And also add config item CRASH_RESERVE to control its enabling of the codes. And update config items which has relationship with crashkernel reservation.

And also change ifdeffery from CONFIG_CRASH_CORE to CONFIG_CRASH_RESERVE when those scopes are only crashkernel reservation related.

And also rename arch/XXX/include/asm/{crash_core.h => crash_reserve.h} on arm64, x86 and risc-v because those architectures' crash_core.h is only related to crashkernel reservation.

---
 arch/arm64/Kconfig                            |   2 +-
 .../asm/{crash_core.h => crash_reserve.h}     |   4 +-
 arch/powerpc/Kconfig                          |   1 +
 arch/powerpc/mm/nohash/kaslr_booke.c          |   4 +-
 arch/riscv/Kconfig                            |   2 +-
 .../asm/{crash_core.h => crash_reserve.h}     |   4 +-
 arch/x86/Kconfig                              |   2 +-
 .../asm/{crash_core.h => crash_reserve.h}     |   6 +-
 include/linux/crash_core.h                    |  40 --
 include/linux/crash_reserve.h                 |  48 ++
 include/linux/kexec.h                         |   1 +
 kernel/Kconfig.kexec                          |   5 +-
 kernel/Makefile                               |   1 +
 kernel/crash_core.c                           | 438 -----------------
 kernel/crash_reserve.c                        | 464 ++++++++++++++++++
 15 files changed, 531 insertions(+), 491 deletions(-)
 rename arch/arm64/include/asm/{crash_core.h => crash_reserve.h} (81%)
 rename arch/riscv/include/asm/{crash_core.h => crash_reserve.h} (78%)
 rename arch/x86/include/asm/{crash_core.h => crash_reserve.h} (92%)
 create mode 100644 include/linux/crash_reserve.h
 create mode 100644 kernel/crash_reserve.c
arch/arm64/Kconfig +1 -1
index 3d2300aca238..96fb363d2f52 100644
@@ -1520,7 +1520,7 @@ config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-	def_bool CRASH_CORE
+	def_bool CRASH_RESERVE
 
 config TRANS_TABLE
 	def_bool y
arch/arm64/include/asm/crash_reserve.h renamed from arch/arm64/include/asm/crash_core.h +2 -2
similarity index 81%
rename from arch/arm64/include/asm/crash_core.h
rename to arch/arm64/include/asm/crash_reserve.h
index 9f5c8d339f44..4afe027a4e7b 100644
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
-#ifndef _ARM64_CRASH_CORE_H
-#define _ARM64_CRASH_CORE_H
+#ifndef _ARM64_CRASH_RESERVE_H
+#define _ARM64_CRASH_RESERVE_H
 
 /* Current arm64 boot protocol requires 2MB alignment */
 #define CRASH_ALIGN                     SZ_2M
arch/powerpc/Kconfig +1 -0
index b9fc064d38d2..7f704ae5c5ef 100644
@@ -691,6 +691,7 @@ config FA_DUMP
 	bool "Firmware-assisted dump"
 	depends on PPC64 && (PPC_RTAS || PPC_POWERNV)
 	select CRASH_CORE
+	select CRASH_RESERVE
 	select CRASH_DUMP
 	help
 	  A robust mechanism to get reliable kernel crash dump with
arch/powerpc/mm/nohash/kaslr_booke.c +2 -2
index b4f2786a7d2b..cdff129abb14 100644
@@ -13,7 +13,7 @@
 #include <linux/delay.h>
 #include <linux/memblock.h>
 #include <linux/libfdt.h>
-#include <linux/crash_core.h>
+#include <linux/crash_reserve.h>
 #include <linux/of.h>
 #include <linux/of_fdt.h>
 #include <asm/cacheflush.h>
@@ -173,7 +173,7 @@ static __init bool overlaps_region(const void *fdt, u32 start,
 
 static void __init get_crash_kernel(void *fdt, unsigned long size)
 {
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_CRASH_RESERVE
 	unsigned long long crash_size, crash_base;
 	int ret;
 
arch/riscv/Kconfig +1 -1
index 69d24f513922..7c6869e43a4a 100644
@@ -767,7 +767,7 @@ config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-	def_bool CRASH_CORE
+	def_bool CRASH_RESERVE
 
 config COMPAT
 	bool "Kernel support for 32-bit U-mode"
arch/riscv/include/asm/crash_reserve.h renamed from arch/riscv/include/asm/crash_core.h +2 -2
similarity index 78%
rename from arch/riscv/include/asm/crash_core.h
rename to arch/riscv/include/asm/crash_reserve.h
index e1874b23feaf..013962e63587 100644
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
-#ifndef _RISCV_CRASH_CORE_H
-#define _RISCV_CRASH_CORE_H
+#ifndef _RISCV_CRASH_RESERVE_H
+#define _RISCV_CRASH_RESERVE_H
 
 #define CRASH_ALIGN			PMD_SIZE
 
arch/x86/Kconfig +1 -1
index 502986237cb6..a9243e0948a3 100644
@@ -2106,7 +2106,7 @@ config ARCH_SUPPORTS_CRASH_HOTPLUG
 	def_bool y
 
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-	def_bool CRASH_CORE
+	def_bool CRASH_RESEERVE
 
 config PHYSICAL_START
 	hex "Physical address where the kernel is loaded" if (EXPERT || CRASH_DUMP)
arch/x86/include/asm/crash_reserve.h renamed from arch/x86/include/asm/crash_core.h +3 -3
similarity index 92%
rename from arch/x86/include/asm/crash_core.h
rename to arch/x86/include/asm/crash_reserve.h
index 76af98f4e801..152239f95541 100644
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
-#ifndef _X86_CRASH_CORE_H
-#define _X86_CRASH_CORE_H
+#ifndef _X86_CRASH_RESERVE_H
+#define _X86_CRASH_RESERVE_H
 
 /* 16M alignment for crash kernel regions */
 #define CRASH_ALIGN             SZ_16M
@@ -39,4 +39,4 @@ static inline unsigned long crash_low_size_default(void)
 #endif
 }
 
-#endif /* _X86_CRASH_CORE_H */
+#endif /* _X86_CRASH_RESERVE_H */
include/linux/crash_core.h +0 -40
index 9eaeaafe0cad..1fde49246fa6 100644
@@ -5,14 +5,6 @@
 #include <linux/linkage.h>
 #include <linux/elfcore.h>
 #include <linux/elf.h>
-#ifdef CONFIG_ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-#include <asm/crash_core.h>
-#endif
-
-/* Location of a reserved region to hold the crash kernel.
- */
-extern struct resource crashk_res;
-extern struct resource crashk_low_res;
 
 #define CRASH_CORE_NOTE_NAME	   "CORE"
 #define CRASH_CORE_NOTE_HEAD_BYTES ALIGN(sizeof(struct elf_note), 4)
@@ -87,38 +79,6 @@ Elf_Word *append_elf_note(Elf_Word *buf, char *name, unsigned int type,
 			  void *data, size_t data_len);
 void final_note(Elf_Word *buf);
 
-int __init parse_crashkernel(char *cmdline, unsigned long long system_ram,
-		unsigned long long *crash_size, unsigned long long *crash_base,
-		unsigned long long *low_size, bool *high);
-
-#ifdef CONFIG_ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-#ifndef DEFAULT_CRASH_KERNEL_LOW_SIZE
-#define DEFAULT_CRASH_KERNEL_LOW_SIZE	(128UL << 20)
-#endif
-#ifndef CRASH_ALIGN
-#define CRASH_ALIGN			SZ_2M
-#endif
-#ifndef CRASH_ADDR_LOW_MAX
-#define CRASH_ADDR_LOW_MAX		SZ_4G
-#endif
-#ifndef CRASH_ADDR_HIGH_MAX
-#define CRASH_ADDR_HIGH_MAX		memblock_end_of_DRAM()
-#endif
-
-void __init reserve_crashkernel_generic(char *cmdline,
-		unsigned long long crash_size,
-		unsigned long long crash_base,
-		unsigned long long crash_low_size,
-		bool high);
-#else
-static inline void __init reserve_crashkernel_generic(char *cmdline,
-		unsigned long long crash_size,
-		unsigned long long crash_base,
-		unsigned long long crash_low_size,
-		bool high)
-{}
-#endif
-
 /* Alignment required for elf header segment */
 #define ELF_CORE_HEADER_ALIGN   4096
 
include/linux/crash_reserve.h new file +48 -0

48 lines not on this page · raw message

include/linux/kexec.h +1 -0
index 400cb6c02176..6d79bfb52e5b 100644
@@ -16,6 +16,7 @@
 #if !defined(__ASSEMBLY__)
 
 #include <linux/crash_core.h>
+#include <linux/crash_reserve.h>
 #include <asm/io.h>
 #include <linux/range.h>
 
kernel/Kconfig.kexec +4 -1
index 946dffa048b7..8b7be71edd85 100644
@@ -2,11 +2,15 @@
 
 menu "Kexec and crash features"
 
+config CRASH_RESERVE
+	bool
+
 config CRASH_CORE
 	bool
 
 config KEXEC_CORE
 	select CRASH_CORE
+	select CRASH_RESERVE
 	bool
 
 config KEXEC_ELF
@@ -96,7 +100,6 @@ config KEXEC_JUMP
 config CRASH_DUMP
 	bool "kernel crash dumps"
 	depends on ARCH_SUPPORTS_CRASH_DUMP
-	select CRASH_CORE
 	select KEXEC_CORE
 	help
 	  Generate crash dump after being started by kexec.
kernel/Makefile +1 -0
index ce105a5558fc..05fa88b3ab74 100644
@@ -69,6 +69,7 @@ obj-$(CONFIG_KALLSYMS) += kallsyms.o
 obj-$(CONFIG_KALLSYMS_SELFTEST) += kallsyms_selftest.o
 obj-$(CONFIG_BSD_PROCESS_ACCT) += acct.o
 obj-$(CONFIG_CRASH_CORE) += crash_core.o
+obj-$(CONFIG_CRASH_RESERVE) += crash_reserve.o
 obj-$(CONFIG_KEXEC_CORE) += kexec_core.o
 obj-$(CONFIG_KEXEC) += kexec.o
 obj-$(CONFIG_KEXEC_FILE) += kexec_file.o
kernel/crash_core.c +0 -438

444 lines not on this page · raw message

kernel/crash_reserve.c new file +464 -0

464 lines not on this page · raw message

Signature · 1 line
2.41.0
#2 lore raw

Re: [PATCH linux-next v3 01/14] kexec: split crashkernel reservation code out from crash_core.c

Hi,

Quoted 12 lines · Baoquan He

And also add config item CRASH_RESERVE to control its enabling of the codes. And update config items which has relationship with crashkernel reservation.

And also change ifdeffery from CONFIG_CRASH_CORE to CONFIG_CRASH_RESERVE when those scopes are only crashkernel reservation related.

arch/x86/Kconfig +0 -0
index 502986237cb6..a9243e0948a3 100644

arch/x86/Kconfig config ARCH_SUPPORTS_CRASH_HOTPLUG · line 2112

Quoted 8 lines
arch/x86/Kconfig +1 -1
@@ -2106,7 +2106,7 @@ config ARCH_SUPPORTS_CRASH_HOTPLUG
  	def_bool y
 
  config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-	def_bool CRASH_CORE
+	def_bool CRASH_RESEERVE
 
  config PHYSICAL_START
  	hex "Physical address where the kernel is loaded" if (EXPERT || CRASH_DUMP)

CRASH_RESEERVE is probably a typo and should be CRASH_RESERVE (with the former ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION isn't defined in my .config and crashkernel=... parameter has no effect).

Kind regards,
Klara Modin

author
#3 lore raw

Re: [PATCH linux-next v3 01/14] kexec: split crashkernel reservation code out from crash_core.c

Quoted 10 lines · Klara Modin

Hi,

Quoted 6 lines · Baoquan He

And also add config item CRASH_RESERVE to control its enabling of the codes. And update config items which has relationship with crashkernel reservation.

And also change ifdeffery from CONFIG_CRASH_CORE to CONFIG_CRASH_RESERVE when those scopes are only crashkernel reservation related.

Quoted 15 lines

arch/x86/Kconfig config ARCH_SUPPORTS_CRASH_HOTPLUG · line 2110

Quoted 11 lines
arch/x86/Kconfig +1 -1
index 502986237cb6..a9243e0948a3 100644
@@ -2106,7 +2106,7 @@ config ARCH_SUPPORTS_CRASH_HOTPLUG
  	def_bool y
  config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
-	def_bool CRASH_CORE
+	def_bool CRASH_RESEERVE
  config PHYSICAL_START
  	hex "Physical address where the kernel is loaded" if (EXPERT || CRASH_DUMP)

CRASH_RESEERVE is probably a typo and should be CRASH_RESERVE (with the former ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION isn't defined in my .config and crashkernel=... parameter has no effect).

You are right, thanks a lot. Will fix that.

author
#4 lore raw

PATCH v3 2/14 crash: split vmcoreinfo exporting code out from crash_core.c

In reply to Baoquan He

Now move the relevant codes into separate files:
kernel/crash_reserve.c, include/linux/crash_reserve.h.

And add config item CRASH_RESERVE to control its enabling.

And also update the old ifdeffery of CONFIG_CRASH_CORE, including of <linux/crash_core.h> and config item dependency on CRASH_CORE accordingly.

And also do renaming as follows:
- arch/xxx/kernel/{crash_core.c => vmcore_info.c}
because they are only related to vmcoreinfo exporting on x86, arm64, riscv.

And also Remove config item CRASH_CORE, and rely on CONFIG_KEXEC_CORE to decide if build in crash_core.c.

---
v2->v3:
- There's conflict when rebasing to linux-next in kernel/crash_core.c because of below commits from Uladzislau:
commit 699d9351822e ("mm: vmalloc: Fix a warning in the crash_save_vmcoreinfo_init()") commit 5f4c0c1e2a51 (:mm/vmalloc: remove vmap_area_list")

 arch/arm64/kernel/Makefile                    |   2 +-
 .../kernel/{crash_core.c => vmcore_info.c}    |   2 +-
 arch/powerpc/Kconfig                          |   2 +-
 arch/powerpc/kernel/setup-common.c            |   2 +-
 arch/powerpc/platforms/powernv/opal-core.c    |   2 +-
 arch/riscv/kernel/Makefile                    |   2 +-
 .../kernel/{crash_core.c => vmcore_info.c}    |   2 +-
 arch/x86/kernel/Makefile                      |   2 +-
 .../{crash_core_32.c => vmcore_info_32.c}     |   2 +-
 .../{crash_core_64.c => vmcore_info_64.c}     |   2 +-
 drivers/firmware/qemu_fw_cfg.c                |  14 +-
 fs/proc/Kconfig                               |   2 +-
 fs/proc/kcore.c                               |   2 +-
 include/linux/buildid.h                       |   2 +-
 include/linux/crash_core.h                    |  73 ------
 include/linux/kexec.h                         |   1 +
 include/linux/vmcore_info.h                   |  81 ++++++
 kernel/Kconfig.kexec                          |   4 +-
 kernel/Makefile                               |   4 +-
 kernel/crash_core.c                           | 206 ----------------
 kernel/ksysfs.c                               |   6 +-
 kernel/printk/printk.c                        |   4 +-
 kernel/vmcore_info.c                          | 231 ++++++++++++++++++
 lib/buildid.c                                 |   2 +-
 24 files changed, 343 insertions(+), 309 deletions(-)
 rename arch/arm64/kernel/{crash_core.c => vmcore_info.c} (97%)
 rename arch/riscv/kernel/{crash_core.c => vmcore_info.c} (96%)
 rename arch/x86/kernel/{crash_core_32.c => vmcore_info_32.c} (90%)
 rename arch/x86/kernel/{crash_core_64.c => vmcore_info_64.c} (94%)
 create mode 100644 include/linux/vmcore_info.h
 create mode 100644 kernel/vmcore_info.c
arch/arm64/kernel/Makefile +1 -1
index e5d03a7039b4..9fd638b91f38 100644
@@ -66,7 +66,7 @@ obj-$(CONFIG_KEXEC_FILE)		+= machine_kexec_file.o kexec_image.o
 obj-$(CONFIG_ARM64_RELOC_TEST)		+= arm64-reloc-test.o
 arm64-reloc-test-y := reloc_test_core.o reloc_test_syms.o
 obj-$(CONFIG_CRASH_DUMP)		+= crash_dump.o
-obj-$(CONFIG_CRASH_CORE)		+= crash_core.o
+obj-$(CONFIG_VMCORE_INFO)		+= vmcore_info.o
 obj-$(CONFIG_ARM_SDE_INTERFACE)		+= sdei.o
 obj-$(CONFIG_ARM64_PTR_AUTH)		+= pointer_auth.o
 obj-$(CONFIG_ARM64_MTE)			+= mte.o
arch/arm64/kernel/vmcore_info.c renamed from arch/arm64/kernel/crash_core.c +1 -1
similarity index 97%
rename from arch/arm64/kernel/crash_core.c
rename to arch/arm64/kernel/vmcore_info.c
index 2a24199a9b81..b19d5d6cb8b3 100644
@@ -4,7 +4,7 @@
  * Copyright (C) Huawei Futurewei Technologies.
  */
 
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <asm/cpufeature.h>
 #include <asm/memory.h>
 #include <asm/pgtable-hwdef.h>
arch/powerpc/Kconfig +1 -1
index 7f704ae5c5ef..495d197c9b27 100644
@@ -690,7 +690,7 @@ config ARCH_SELECTS_CRASH_DUMP
 config FA_DUMP
 	bool "Firmware-assisted dump"
 	depends on PPC64 && (PPC_RTAS || PPC_POWERNV)
-	select CRASH_CORE
+	select VMCORE_INFO
 	select CRASH_RESERVE
 	select CRASH_DUMP
 	help
arch/powerpc/kernel/setup-common.c +1 -1
index 9b142b9d5187..733f210ffda1 100644
@@ -109,7 +109,7 @@ int ppc_do_canonicalize_irqs;
 EXPORT_SYMBOL(ppc_do_canonicalize_irqs);
 #endif
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 /* This keeps a track of which one is the crashing cpu. */
 int crashing_cpu = -1;
 #endif
arch/powerpc/platforms/powernv/opal-core.c +1 -1
index bb7657115f1d..c9a9b759cc92 100644
@@ -16,7 +16,7 @@
 #include <linux/kobject.h>
 #include <linux/sysfs.h>
 #include <linux/slab.h>
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/of.h>
 
 #include <asm/page.h>
arch/riscv/kernel/Makefile +1 -1
index f71910718053..d6fd8dcfceb5 100644
@@ -92,7 +92,7 @@ obj-$(CONFIG_KGDB)		+= kgdb.o
 obj-$(CONFIG_KEXEC_CORE)	+= kexec_relocate.o crash_save_regs.o machine_kexec.o
 obj-$(CONFIG_KEXEC_FILE)	+= elf_kexec.o machine_kexec_file.o
 obj-$(CONFIG_CRASH_DUMP)	+= crash_dump.o
-obj-$(CONFIG_CRASH_CORE)	+= crash_core.o
+obj-$(CONFIG_VMCORE_INFO)	+= vmcore_info.o
 
 obj-$(CONFIG_JUMP_LABEL)	+= jump_label.o
 
arch/riscv/kernel/vmcore_info.c renamed from arch/riscv/kernel/crash_core.c +1 -1
similarity index 96%
rename from arch/riscv/kernel/crash_core.c
rename to arch/riscv/kernel/vmcore_info.c
index d18d529fd9b9..6d7a22522d63 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/pagemap.h>
 
 void arch_crash_save_vmcoreinfo(void)
arch/x86/kernel/Makefile +1 -1
index 0000325ab98f..913d4022131e 100644
@@ -98,7 +98,7 @@ obj-$(CONFIG_FTRACE_SYSCALLS)	+= ftrace.o
 obj-$(CONFIG_X86_TSC)		+= trace_clock.o
 obj-$(CONFIG_TRACING)		+= trace.o
 obj-$(CONFIG_RETHOOK)		+= rethook.o
-obj-$(CONFIG_CRASH_CORE)	+= crash_core_$(BITS).o
+obj-$(CONFIG_VMCORE_INFO)	+= vmcore_info_$(BITS).o
 obj-$(CONFIG_KEXEC_CORE)	+= machine_kexec_$(BITS).o
 obj-$(CONFIG_KEXEC_CORE)	+= relocate_kernel_$(BITS).o crash.o
 obj-$(CONFIG_KEXEC_FILE)	+= kexec-bzimage64.o
arch/x86/kernel/vmcore_info_32.c renamed from arch/x86/kernel/crash_core_32.c +1 -1
similarity index 90%
rename from arch/x86/kernel/crash_core_32.c
rename to arch/x86/kernel/vmcore_info_32.c
index 8a89c109e20a..5995a749288a 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/pgtable.h>
 
 #include <asm/setup.h>
arch/x86/kernel/vmcore_info_64.c renamed from arch/x86/kernel/crash_core_64.c +1 -1
similarity index 94%
rename from arch/x86/kernel/crash_core_64.c
rename to arch/x86/kernel/vmcore_info_64.c
index 7d255f882afe..0dec7d868754 100644
@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/pgtable.h>
 
 #include <asm/setup.h>
drivers/firmware/qemu_fw_cfg.c +7 -7
index 03da9a4354f8..5f43dfa22f79 100644
@@ -37,7 +37,7 @@
 #include <uapi/linux/qemu_fw_cfg.h>
 #include <linux/delay.h>
 #include <linux/crash_dump.h>
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 
 MODULE_AUTHOR("Gabriel L. Somlo <[email protected]>");
 MODULE_DESCRIPTION("QEMU fw_cfg sysfs support");
@@ -67,7 +67,7 @@ static void fw_cfg_sel_endianness(u16 key)
 		iowrite16(key, fw_cfg_reg_ctrl);
 }
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 static inline bool fw_cfg_dma_enabled(void)
 {
 	return (fw_cfg_rev & FW_CFG_VERSION_DMA) && fw_cfg_reg_dma;
@@ -156,7 +156,7 @@ static ssize_t fw_cfg_read_blob(u16 key,
 	return count;
 }
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 /* write chunk of given fw_cfg blob (caller responsible for sanity-check) */
 static ssize_t fw_cfg_write_blob(u16 key,
 				 void *buf, loff_t pos, size_t count)
@@ -195,7 +195,7 @@ static ssize_t fw_cfg_write_blob(u16 key,
 
 	return ret;
 }
-#endif /* CONFIG_CRASH_CORE */
+#endif /* CONFIG_VMCORE_INFO */
 
 /* clean up fw_cfg device i/o */
 static void fw_cfg_io_cleanup(void)
@@ -319,7 +319,7 @@ struct fw_cfg_sysfs_entry {
 	struct list_head list;
 };
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 static ssize_t fw_cfg_write_vmcoreinfo(const struct fw_cfg_file *f)
 {
 	static struct fw_cfg_vmcoreinfo *data;
@@ -343,7 +343,7 @@ static ssize_t fw_cfg_write_vmcoreinfo(const struct fw_cfg_file *f)
 	kfree(data);
 	return ret;
 }
-#endif /* CONFIG_CRASH_CORE */
+#endif /* CONFIG_VMCORE_INFO */
 
 /* get fw_cfg_sysfs_entry from kobject member */
 static inline struct fw_cfg_sysfs_entry *to_entry(struct kobject *kobj)
@@ -583,7 +583,7 @@ static int fw_cfg_register_file(const struct fw_cfg_file *f)
 	int err;
 	struct fw_cfg_sysfs_entry *entry;
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 	if (fw_cfg_dma_enabled() &&
 		strcmp(f->name, FW_CFG_VMCOREINFO_FILENAME) == 0 &&
 		!is_kdump_kernel()) {
fs/proc/Kconfig +1 -1
index 32b1116ae137..d80a1431ef7b 100644
@@ -32,7 +32,7 @@ config PROC_FS
 config PROC_KCORE
 	bool "/proc/kcore support" if !ARM
 	depends on PROC_FS && MMU
-	select CRASH_CORE
+	select VMCORE_INFO
 	help
 	  Provides a virtual ELF core file of the live kernel.  This can
 	  be read with gdb and other ELF tools.  No modifications can be
fs/proc/kcore.c +1 -1
index 6422e569b080..8e08a9a1b7ed 100644
@@ -10,7 +10,7 @@
  *	Safe accesses to vmalloc/direct-mapped discontiguous areas, Kanoj Sarcar <[email protected]>
  */
 
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/mm.h>
 #include <linux/proc_fs.h>
 #include <linux/kcore.h>
include/linux/buildid.h +1 -1
index 8a582d242f06..20aa3c2d89f7 100644
@@ -11,7 +11,7 @@ int build_id_parse(struct vm_area_struct *vma, unsigned char *build_id,
 		   __u32 *size);
 int build_id_parse_buf(const void *buf, unsigned char *build_id, u32 buf_size);
 
-#if IS_ENABLED(CONFIG_STACKTRACE_BUILD_ID) || IS_ENABLED(CONFIG_CRASH_CORE)
+#if IS_ENABLED(CONFIG_STACKTRACE_BUILD_ID) || IS_ENABLED(CONFIG_VMCORE_INFO)
 extern unsigned char vmlinux_build_id[BUILD_ID_SIZE_MAX];
 void init_vmlinux_build_id(void);
 #else
include/linux/crash_core.h +0 -73

79 lines not on this page · raw message

include/linux/kexec.h +1 -0
index 6d79bfb52e5b..9c7bb8b56ed6 100644
@@ -16,6 +16,7 @@
 #if !defined(__ASSEMBLY__)
 
 #include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/crash_reserve.h>
 #include <asm/io.h>
 #include <linux/range.h>
include/linux/vmcore_info.h new file +81 -0

81 lines not on this page · raw message

kernel/Kconfig.kexec +2 -2
index 8b7be71edd85..8faf27043432 100644
@@ -5,11 +5,11 @@ menu "Kexec and crash features"
 config CRASH_RESERVE
 	bool
 
-config CRASH_CORE
+config VMCORE_INFO
 	bool
 
 config KEXEC_CORE
-	select CRASH_CORE
+	select VMCORE_INFO
 	select CRASH_RESERVE
 	bool
 
kernel/Makefile +2 -2
index 05fa88b3ab74..649272a1d6b9 100644
@@ -68,9 +68,9 @@ obj-$(CONFIG_MODULE_SIG_FORMAT) += module_signature.o
 obj-$(CONFIG_KALLSYMS) += kallsyms.o
 obj-$(CONFIG_KALLSYMS_SELFTEST) += kallsyms_selftest.o
 obj-$(CONFIG_BSD_PROCESS_ACCT) += acct.o
-obj-$(CONFIG_CRASH_CORE) += crash_core.o
+obj-$(CONFIG_VMCORE_INFO) += vmcore_info.o
 obj-$(CONFIG_CRASH_RESERVE) += crash_reserve.o
-obj-$(CONFIG_KEXEC_CORE) += kexec_core.o
+obj-$(CONFIG_KEXEC_CORE) += kexec_core.o crash_core.o
 obj-$(CONFIG_KEXEC) += kexec.o
 obj-$(CONFIG_KEXEC_FILE) += kexec_file.o
 obj-$(CONFIG_KEXEC_ELF) += kexec_elf.o
kernel/crash_core.c +0 -206

218 lines not on this page · raw message

kernel/ksysfs.c +3 -3
index 1d4bc493b2f4..11526fc42bc2 100644
@@ -154,7 +154,7 @@ KERNEL_ATTR_RW(kexec_crash_size);
 
 #endif /* CONFIG_KEXEC_CORE */
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 
 static ssize_t vmcoreinfo_show(struct kobject *kobj,
 			       struct kobj_attribute *attr, char *buf)
@@ -177,7 +177,7 @@ KERNEL_ATTR_RO(crash_elfcorehdr_size);
 
 #endif
 
-#endif /* CONFIG_CRASH_CORE */
+#endif /* CONFIG_VMCORE_INFO */
 
 /* whether file capabilities are enabled */
 static ssize_t fscaps_show(struct kobject *kobj,
@@ -265,7 +265,7 @@ static struct attribute * kernel_attrs[] = {
 	&kexec_crash_loaded_attr.attr,
 	&kexec_crash_size_attr.attr,
 #endif
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 	&vmcoreinfo_attr.attr,
 #ifdef CONFIG_CRASH_HOTPLUG
 	&crash_elfcorehdr_size_attr.attr,
kernel/printk/printk.c +2 -2
index f2444b581e16..7d74b000b43a 100644
@@ -34,7 +34,7 @@
 #include <linux/security.h>
 #include <linux/memblock.h>
 #include <linux/syscalls.h>
-#include <linux/crash_core.h>
+#include <linux/vmcore_info.h>
 #include <linux/ratelimit.h>
 #include <linux/kmsg_dump.h>
 #include <linux/syslog.h>
@@ -951,7 +951,7 @@ const struct file_operations kmsg_fops = {
 	.release = devkmsg_release,
 };
 
-#ifdef CONFIG_CRASH_CORE
+#ifdef CONFIG_VMCORE_INFO
 /*
  * This appends the listed symbols to /proc/vmcore
  *
kernel/vmcore_info.c new file +231 -0

231 lines not on this page · raw message

lib/buildid.c +1 -1
index e3a7acdeef0e..3e6868c86b45 100644
@@ -174,7 +174,7 @@ int build_id_parse_buf(const void *buf, unsigned char *build_id, u32 buf_size)
 	return parse_build_id_buf(build_id, NULL, buf, buf_size);
 }
 
-#if IS_ENABLED(CONFIG_STACKTRACE_BUILD_ID) || IS_ENABLED(CONFIG_CRASH_CORE)
+#if IS_ENABLED(CONFIG_STACKTRACE_BUILD_ID) || IS_ENABLED(CONFIG_VMCORE_INFO)
 unsigned char vmlinux_build_id[BUILD_ID_SIZE_MAX] __ro_after_init;
 
 /**
Signature · 1 line
2.41.0
#5 lore raw

Re: [PATCH linux-next v3 02/14] crash: split vmcoreinfo exporting code out from crash_core.c

Hi Baoquan,

Quoted 17 lines · Baoquan He

Now move the relevant codes into separate files:
kernel/crash_reserve.c, include/linux/crash_reserve.h.

And add config item CRASH_RESERVE to control its enabling.

And also update the old ifdeffery of CONFIG_CRASH_CORE, including of <linux/crash_core.h> and config item dependency on CRASH_CORE accordingly.

And also do renaming as follows:
- arch/xxx/kernel/{crash_core.c => vmcore_info.c}
because they are only related to vmcoreinfo exporting on x86, arm64, riscv.

And also Remove config item CRASH_CORE, and rely on CONFIG_KEXEC_CORE to decide if build in crash_core.c.

Quoted 7 lines

---
v2->v3:
- There's conflict when rebasing to linux-next in kernel/crash_core.c because of below commits from Uladzislau:
commit 699d9351822e ("mm: vmalloc: Fix a warning in the crash_save_vmcoreinfo_init()") commit 5f4c0c1e2a51 (:mm/vmalloc: remove vmap_area_list")

Thanks for your patch, which is now commit 443cbaf9e2fdbef7 ("crash: split vmcoreinfo exporting code out from crash_core.c") in v6.9-rc1.

After this, there are still two references left to CRASH_CORE:

1. Documentation/admin-guide/kdump/kdump.rst:

Subsequently, CRASH_CORE is selected by KEXEC_CORE::

2. arch/loongarch/Kconfig

config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
def_bool CRASH_CORE

Gr{oetje,eeting}s,

Geert

Signature · 5 lines
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
author
#6 lore raw

Re: [PATCH linux-next v3 02/14] crash: split vmcoreinfo exporting code out from crash_core.c

Quoted 34 lines · Geert Uytterhoeven

Hi Baoquan,

Quoted 17 lines · Baoquan He

Now move the relevant codes into separate files:
kernel/crash_reserve.c, include/linux/crash_reserve.h.

And add config item CRASH_RESERVE to control its enabling.

And also update the old ifdeffery of CONFIG_CRASH_CORE, including of <linux/crash_core.h> and config item dependency on CRASH_CORE accordingly.

And also do renaming as follows:
- arch/xxx/kernel/{crash_core.c => vmcore_info.c}
because they are only related to vmcoreinfo exporting on x86, arm64, riscv.

And also Remove config item CRASH_CORE, and rely on CONFIG_KEXEC_CORE to decide if build in crash_core.c.

Quoted 7 lines

---
v2->v3:
- There's conflict when rebasing to linux-next in kernel/crash_core.c because of below commits from Uladzislau:
commit 699d9351822e ("mm: vmalloc: Fix a warning in the crash_save_vmcoreinfo_init()") commit 5f4c0c1e2a51 (:mm/vmalloc: remove vmap_area_list")

Thanks for your patch, which is now commit 443cbaf9e2fdbef7 ("crash: split vmcoreinfo exporting code out from crash_core.c") in v6.9-rc1.

After this, there are still two references left to CRASH_CORE:

Quoted 8 lines

1. Documentation/admin-guide/kdump/kdump.rst:

Subsequently, CRASH_CORE is selected by KEXEC_CORE::

2. arch/loongarch/Kconfig

config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
def_bool CRASH_CORE

Sure, will post patch to clean them up. Thanks a lot.

author
#7 lore raw

PATCH v3 3/14 crash: remove dependency of FA_DUMP on CRASH_DUMP

In reply to Baoquan He

In kdump kernel, /proc/vmcore is an elf file mapping the crashed kernel's old memory content. Its elf header is constructed in 1st kernel and passed to kdump kernel via elfcorehdr_addr. Config CRASH_DUMP enables the code of 1st kernel's old memory accessing in different architectures.

Currently, config FA_DUMP has dependency on CRASH_DUMP because fadump needs access global variable 'elfcorehdr_addr' to judge if it's in kdump kernel within function is_kdump_kernel(). In the current kernel/crash_dump.c, variable 'elfcorehdr_addr' is defined, and function setup_elfcorehdr() used to parse kernel parameter to fetch the passed value of elfcorehdr_addr. Only for accessing elfcorehdr_addr, FA_DUMP really doesn't have to depends on CRASH_DUMP.

To remove the dependency of FA_DUMP on CRASH_DUMP to avoid confusion, rename kernel/crash_dump.c to kernel/elfcorehdr.c, and build it when CONFIG_VMCORE_INFO is ebabled. With this, FA_DUMP doesn't need to depend on CRASH_DUMP.

---
 arch/powerpc/Kconfig                  | 1 -
 kernel/Makefile                       | 3 +--
 kernel/{crash_dump.c => elfcorehdr.c} | 0
 kernel/kexec_internal.h               | 2 ++
 4 files changed, 3 insertions(+), 3 deletions(-)
 rename kernel/{crash_dump.c => elfcorehdr.c} (100%)
arch/powerpc/Kconfig +0 -1
index 495d197c9b27..e66fd9923250 100644
@@ -692,7 +692,6 @@ config FA_DUMP
 	depends on PPC64 && (PPC_RTAS || PPC_POWERNV)
 	select VMCORE_INFO
 	select CRASH_RESERVE
-	select CRASH_DUMP
 	help
 	  A robust mechanism to get reliable kernel crash dump with
 	  assistance from firmware. This approach does not use kexec,
kernel/Makefile +1 -2
index 649272a1d6b9..35abc65e1f1a 100644
@@ -68,7 +68,7 @@ obj-$(CONFIG_MODULE_SIG_FORMAT) += module_signature.o
 obj-$(CONFIG_KALLSYMS) += kallsyms.o
 obj-$(CONFIG_KALLSYMS_SELFTEST) += kallsyms_selftest.o
 obj-$(CONFIG_BSD_PROCESS_ACCT) += acct.o
-obj-$(CONFIG_VMCORE_INFO) += vmcore_info.o
+obj-$(CONFIG_VMCORE_INFO) += vmcore_info.o elfcorehdr.o
 obj-$(CONFIG_CRASH_RESERVE) += crash_reserve.o
 obj-$(CONFIG_KEXEC_CORE) += kexec_core.o crash_core.o
 obj-$(CONFIG_KEXEC) += kexec.o
@@ -121,7 +121,6 @@ obj-$(CONFIG_PERF_EVENTS) += events/
 
 obj-$(CONFIG_USER_RETURN_NOTIFIER) += user-return-notifier.o
 obj-$(CONFIG_PADATA) += padata.o
-obj-$(CONFIG_CRASH_DUMP) += crash_dump.o
 obj-$(CONFIG_JUMP_LABEL) += jump_label.o
 obj-$(CONFIG_CONTEXT_TRACKING) += context_tracking.o
 obj-$(CONFIG_TORTURE_TEST) += torture.o
kernel/elfcorehdr.c renamed from kernel/crash_dump.c +0 -0
similarity index 100%
rename from kernel/crash_dump.c
rename to kernel/elfcorehdr.c
kernel/kexec_internal.h +2 -0
index 74da1409cd14..2595defe8c0d 100644
@@ -4,6 +4,8 @@
 
 #include <linux/kexec.h>
 
+struct kexec_segment;
+
 struct kimage *do_kimage_alloc_init(void);
 int sanity_check_segment_list(struct kimage *image);
 void kimage_free_page_list(struct list_head *list);
Signature · 1 line
2.41.0
author
#8 lore raw

PATCH v3 4/14 crash: split crash dumping code out from kexec_core.c

In reply to Baoquan He

Currently, KEXEC_CORE select CRASH_CORE automatically because crash codes need be built in to avoid compiling error when building kexec code even though the crash dumping functionality is not enabled. E.g
--------------------
CONFIG_CRASH_CORE=y
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC=y
CONFIG_KEXEC_FILE=y
---------------------

After splitting out crashkernel reservation code and vmcoreinfo exporting code, there's only crash related code left in kernel/crash_core.c. Now move crash related codes from kexec_core.c to crash_core.c and only build it in when CONFIG_CRASH_DUMP=y.

And also wrap up crash codes inside CONFIG_CRASH_DUMP ifdeffery scope, or replace inappropriate CONFIG_KEXEC_CORE ifdef with CONFIG_CRASH_DUMP ifdef in generic kernel files.

With these changes, crash_core codes are abstracted from kexec codes and can be disabled at all if only kexec reboot feature is wanted.

---
 drivers/base/cpu.c         |   6 +-
 include/linux/crash_core.h |  61 +++++++++
 include/linux/kexec.h      |  45 +------
 init/initramfs.c           |   2 +-
 kernel/Makefile            |   3 +-
 kernel/crash_core.c        | 256 +++++++++++++++++++++++++++++++++++++
 kernel/kexec.c             |  11 +-
 kernel/kexec_core.c        | 250 ++----------------------------------
 kernel/kexec_file.c        |  13 +-
 kernel/ksysfs.c            |   4 +
 10 files changed, 359 insertions(+), 292 deletions(-)
drivers/base/cpu.c +3 -3
index 47de0f140ba6..b621a0fc75e1 100644
@@ -144,7 +144,7 @@ static DEVICE_ATTR(release, S_IWUSR, NULL, cpu_release_store);
 #endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
 #endif /* CONFIG_HOTPLUG_CPU */
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 #include <linux/kexec.h>
 
 static ssize_t crash_notes_show(struct device *dev,
@@ -189,14 +189,14 @@ static const struct attribute_group crash_note_cpu_attr_group = {
 #endif
 
 static const struct attribute_group *common_cpu_attr_groups[] = {
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 	&crash_note_cpu_attr_group,
 #endif
 	NULL
 };
 
 static const struct attribute_group *hotplugable_cpu_attr_groups[] = {
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 	&crash_note_cpu_attr_group,
 #endif
 	NULL
include/linux/crash_core.h +61 -0

71 lines not on this page · raw message

include/linux/kexec.h +1 -44

75 lines not on this page · raw message

init/initramfs.c +1 -1
index d3c623dde01a..70cdefa0911b 100644
@@ -642,7 +642,7 @@ void __weak __init free_initrd_mem(unsigned long start, unsigned long end)
 			"initrd");
 }
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_RESERVE
 static bool __init kexec_free_initrd(void)
 {
 	unsigned long crashk_start = (unsigned long)__va(crashk_res.start);
kernel/Makefile +2 -1
index 35abc65e1f1a..3c13240dfc9f 100644
@@ -70,7 +70,8 @@ obj-$(CONFIG_KALLSYMS_SELFTEST) += kallsyms_selftest.o
 obj-$(CONFIG_BSD_PROCESS_ACCT) += acct.o
 obj-$(CONFIG_VMCORE_INFO) += vmcore_info.o elfcorehdr.o
 obj-$(CONFIG_CRASH_RESERVE) += crash_reserve.o
-obj-$(CONFIG_KEXEC_CORE) += kexec_core.o crash_core.o
+obj-$(CONFIG_KEXEC_CORE) += kexec_core.o
+obj-$(CONFIG_CRASH_DUMP) += crash_core.o
 obj-$(CONFIG_KEXEC) += kexec.o
 obj-$(CONFIG_KEXEC_FILE) += kexec_file.o
 obj-$(CONFIG_KEXEC_ELF) += kexec_elf.o
kernel/crash_core.c +256 -0

283 lines not on this page · raw message

kernel/kexec.c +9 -2
index 8f35a5a42af8..bab542fc1463 100644
@@ -28,12 +28,14 @@ static int kimage_alloc_init(struct kimage **rimage, unsigned long entry,
 	struct kimage *image;
 	bool kexec_on_panic = flags & KEXEC_ON_CRASH;
 
+#ifdef CONFIG_CRASH_DUMP
 	if (kexec_on_panic) {
 		/* Verify we have a valid entry point */
 		if ((entry < phys_to_boot_phys(crashk_res.start)) ||
 		    (entry > phys_to_boot_phys(crashk_res.end)))
 			return -EADDRNOTAVAIL;
 	}
+#endif
 
 	/* Allocate and initialize a controlling structure */
 	image = do_kimage_alloc_init();
@@ -44,11 +46,13 @@ static int kimage_alloc_init(struct kimage **rimage, unsigned long entry,
 	image->nr_segments = nr_segments;
 	memcpy(image->segment, segments, nr_segments * sizeof(*segments));
 
+#ifdef CONFIG_CRASH_DUMP
 	if (kexec_on_panic) {
 		/* Enable special crash kernel control page alloc policy. */
 		image->control_page = crashk_res.start;
 		image->type = KEXEC_TYPE_CRASH;
 	}
+#endif
 
 	ret = sanity_check_segment_list(image);
 	if (ret)
@@ -99,13 +103,14 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
 	if (!kexec_trylock())
 		return -EBUSY;
 
+#ifdef CONFIG_CRASH_DUMP
 	if (flags & KEXEC_ON_CRASH) {
 		dest_image = &kexec_crash_image;
 		if (kexec_crash_image)
 			arch_kexec_unprotect_crashkres();
-	} else {
+	} else
+#endif
 		dest_image = &kexec_image;
-	}
 
 	if (nr_segments == 0) {
 		/* Uninstall image */
@@ -162,8 +167,10 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
 	image = xchg(dest_image, image);
 
 out:
+#ifdef CONFIG_CRASH_DUMP
 	if ((flags & KEXEC_ON_CRASH) && kexec_crash_image)
 		arch_kexec_protect_crashkres();
+#endif
 
 	kimage_free(image);
 out_unlock:
kernel/kexec_core.c +12 -238

331 lines not on this page · raw message

kernel/kexec_file.c +10 -3
index bef2f6f2571b..ce7ce2ae27cd 100644
@@ -285,11 +285,13 @@ kimage_file_alloc_init(struct kimage **rimage, int kernel_fd,
 	kexec_file_dbg_print = !!(flags & KEXEC_FILE_DEBUG);
 	image->file_mode = 1;
 
+#ifdef CONFIG_CRASH_DUMP
 	if (kexec_on_panic) {
 		/* Enable special crash kernel control page alloc policy. */
 		image->control_page = crashk_res.start;
 		image->type = KEXEC_TYPE_CRASH;
 	}
+#endif
 
 	ret = kimage_file_prepare_segments(image, kernel_fd, initrd_fd,
 					   cmdline_ptr, cmdline_len, flags);
@@ -349,13 +351,14 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,
 	if (!kexec_trylock())
 		return -EBUSY;
 
+#ifdef CONFIG_CRASH_DUMP
 	if (image_type == KEXEC_TYPE_CRASH) {
 		dest_image = &kexec_crash_image;
 		if (kexec_crash_image)
 			arch_kexec_unprotect_crashkres();
-	} else {
+	} else
+#endif
 		dest_image = &kexec_image;
-	}
 
 	if (flags & KEXEC_FILE_UNLOAD)
 		goto exchange;
@@ -419,8 +422,10 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,
 exchange:
 	image = xchg(dest_image, image);
 out:
+#ifdef CONFIG_CRASH_DUMP
 	if ((flags & KEXEC_FILE_ON_CRASH) && kexec_crash_image)
 		arch_kexec_protect_crashkres();
+#endif
 
 	kexec_unlock();
 	kimage_free(image);
@@ -595,12 +600,14 @@ static int kexec_walk_memblock(struct kexec_buf *kbuf,
 static int kexec_walk_resources(struct kexec_buf *kbuf,
 				int (*func)(struct resource *, void *))
 {
+#ifdef CONFIG_CRASH_DUMP
 	if (kbuf->image->type == KEXEC_TYPE_CRASH)
 		return walk_iomem_res_desc(crashk_res.desc,
 					   IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY,
 					   crashk_res.start, crashk_res.end,
 					   kbuf, func);
-	else if (kbuf->top_down)
+#endif
+	if (kbuf->top_down)
 		return walk_system_ram_res_rev(0, ULONG_MAX, kbuf, func);
 	else
 		return walk_system_ram_res(0, ULONG_MAX, kbuf, func);
kernel/ksysfs.c +4 -0
index 11526fc42bc2..fe7a517fc4ab 100644
@@ -120,6 +120,7 @@ static ssize_t kexec_loaded_show(struct kobject *kobj,
 }
 KERNEL_ATTR_RO(kexec_loaded);
 
+#ifdef CONFIG_CRASH_DUMP
 static ssize_t kexec_crash_loaded_show(struct kobject *kobj,
 				       struct kobj_attribute *attr, char *buf)
 {
@@ -152,6 +153,7 @@ static ssize_t kexec_crash_size_store(struct kobject *kobj,
 }
 KERNEL_ATTR_RW(kexec_crash_size);
 
+#endif /* CONFIG_CRASH_DUMP*/
 #endif /* CONFIG_KEXEC_CORE */
 
 #ifdef CONFIG_VMCORE_INFO
@@ -262,9 +264,11 @@ static struct attribute * kernel_attrs[] = {
 #endif
 #ifdef CONFIG_KEXEC_CORE
 	&kexec_loaded_attr.attr,
+#ifdef CONFIG_CRASH_DUMP
 	&kexec_crash_loaded_attr.attr,
 	&kexec_crash_size_attr.attr,
 #endif
+#endif
 #ifdef CONFIG_VMCORE_INFO
 	&vmcoreinfo_attr.attr,
 #ifdef CONFIG_CRASH_HOTPLUG
Signature · 1 line
2.41.0
author
#9 lore raw

PATCH v3 5/14 crash: clean up kdump related config items

In reply to Baoquan He

By splitting CRASH_RESERVE and VMCORE_INFO out from CRASH_CORE, cleaning up the dependency of FA_DMUMP on CRASH_DUMP, and moving crash codes from kexec_core.c to crash_core.c, now we can rearrange CRASH_DUMP to depend on KEXEC_CORE, and make CRASH_DUMP select CRASH_RESERVE and VMCORE_INFO.

KEXEC_CORE won't select CRASH_RESERVE and VMCORE_INFO any more because KEXEC_CORE enables codes which allocate control pages, copy kexec/kdump segments, and prepare for switching. These codes are shared by both kexec reboot and crash dumping.

Doing this makes codes and the corresponding config items more logical (the right item depends on or is selected by the left item).

PROC_KCORE -----------> VMCORE_INFO

           |----------> VMCORE_INFO
FA_DUMP----|
           |----------> CRASH_RESERVE
                                                ---->VMCORE_INFO
                                               /
                                               |---->CRASH_RESERVE
KEXEC      --|                                /|
             |--> KEXEC_CORE--> CRASH_DUMP-->/-|---->PROC_VMCORE
KEXEC_FILE --|                               \ |
                                               \---->CRASH_HOTPLUG
KEXEC      --|
             |--> KEXEC_CORE--> kexec reboot
KEXEC_FILE --|
---
 kernel/Kconfig.kexec | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/Kconfig.kexec +4 -3
index 8faf27043432..6c34e63c88ff 100644
@@ -9,8 +9,6 @@ config VMCORE_INFO
 	bool
 
 config KEXEC_CORE
-	select VMCORE_INFO
-	select CRASH_RESERVE
 	bool
 
 config KEXEC_ELF
@@ -99,8 +97,11 @@ config KEXEC_JUMP
 
 config CRASH_DUMP
 	bool "kernel crash dumps"
+	default y
 	depends on ARCH_SUPPORTS_CRASH_DUMP
-	select KEXEC_CORE
+	depends on KEXEC_CORE
+	select VMCORE_INFO
+	select CRASH_RESERVE
 	help
 	  Generate crash dump after being started by kexec.
 	  This should be normally only set in special crash dump kernels
Signature · 1 line
2.41.0
#10 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Hi Baoquan,

Quoted 53 lines · Baoquan He

By splitting CRASH_RESERVE and VMCORE_INFO out from CRASH_CORE, cleaning up the dependency of FA_DMUMP on CRASH_DUMP, and moving crash codes from kexec_core.c to crash_core.c, now we can rearrange CRASH_DUMP to depend on KEXEC_CORE, and make CRASH_DUMP select CRASH_RESERVE and VMCORE_INFO.

KEXEC_CORE won't select CRASH_RESERVE and VMCORE_INFO any more because KEXEC_CORE enables codes which allocate control pages, copy kexec/kdump segments, and prepare for switching. These codes are shared by both kexec reboot and crash dumping.

Doing this makes codes and the corresponding config items more logical (the right item depends on or is selected by the left item).

PROC_KCORE -----------> VMCORE_INFO

           |----------> VMCORE_INFO
FA_DUMP----|
           |----------> CRASH_RESERVE
                                                ---->VMCORE_INFO
                                               /
                                               |---->CRASH_RESERVE
KEXEC      --|                                /|
             |--> KEXEC_CORE--> CRASH_DUMP-->/-|---->PROC_VMCORE
KEXEC_FILE --|                               \ |
                                               \---->CRASH_HOTPLUG
KEXEC      --|
             |--> KEXEC_CORE--> kexec reboot
KEXEC_FILE --|
---
 kernel/Kconfig.kexec | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/Kconfig.kexec +0 -2
index 8faf27043432..6c34e63c88ff 100644
@@ -9,8 +9,6 @@ config VMCORE_INFO
 	bool
 
 config KEXEC_CORE
-	select VMCORE_INFO
-	select CRASH_RESERVE
 	bool
 
 config KEXEC_ELF
@@ -99,8 +97,11 @@ config KEXEC_JUMP
 
 config CRASH_DUMP

kernel/Kconfig.kexec config KEXEC_JUMP · line 107

Quoted 10 lines
kernel/Kconfig.kexec +4 -1
 	bool "kernel crash dumps"
+	default y
 	depends on ARCH_SUPPORTS_CRASH_DUMP
-	select KEXEC_CORE
+	depends on KEXEC_CORE
+	select VMCORE_INFO
+	select CRASH_RESERVE
 	help
 	  Generate crash dump after being started by kexec.
 	  This should be normally only set in special crash dump kernels

The change to enable CONFIG_CRASH_DUMP by default apparently broke the boot on 32-bit Power Macintosh systems which fail after GRUB with:

"Error: You can't boot a kdump kernel from OF!"

We may have to turn this off for 32-bit Power Macintosh systems.

See this thread on debian-powerpc ML: https://lists.debian.org/debian-powerpc/2024/07/msg00001.html

Thanks,
Adrian

Signature · 4 lines
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913
author
#11 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Quoted 70 lines · John Paul Adrian Glaubitz

Hi Baoquan,

Quoted 53 lines · Baoquan He

By splitting CRASH_RESERVE and VMCORE_INFO out from CRASH_CORE, cleaning up the dependency of FA_DMUMP on CRASH_DUMP, and moving crash codes from kexec_core.c to crash_core.c, now we can rearrange CRASH_DUMP to depend on KEXEC_CORE, and make CRASH_DUMP select CRASH_RESERVE and VMCORE_INFO.

KEXEC_CORE won't select CRASH_RESERVE and VMCORE_INFO any more because KEXEC_CORE enables codes which allocate control pages, copy kexec/kdump segments, and prepare for switching. These codes are shared by both kexec reboot and crash dumping.

Doing this makes codes and the corresponding config items more logical (the right item depends on or is selected by the left item).

PROC_KCORE -----------> VMCORE_INFO

           |----------> VMCORE_INFO
FA_DUMP----|
           |----------> CRASH_RESERVE
                                                ---->VMCORE_INFO
                                               /
                                               |---->CRASH_RESERVE
KEXEC      --|                                /|
             |--> KEXEC_CORE--> CRASH_DUMP-->/-|---->PROC_VMCORE
KEXEC_FILE --|                               \ |
                                               \---->CRASH_HOTPLUG
KEXEC      --|
             |--> KEXEC_CORE--> kexec reboot
KEXEC_FILE --|
---
 kernel/Kconfig.kexec | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/Kconfig.kexec +0 -2
index 8faf27043432..6c34e63c88ff 100644
@@ -9,8 +9,6 @@ config VMCORE_INFO
 	bool
 
 config KEXEC_CORE
-	select VMCORE_INFO
-	select CRASH_RESERVE
 	bool
 
 config KEXEC_ELF
@@ -99,8 +97,11 @@ config KEXEC_JUMP
 
 config CRASH_DUMP

kernel/Kconfig.kexec config KEXEC_JUMP · line 107

Quoted 10 lines
kernel/Kconfig.kexec +4 -1
 	bool "kernel crash dumps"
+	default y
 	depends on ARCH_SUPPORTS_CRASH_DUMP
-	select KEXEC_CORE
+	depends on KEXEC_CORE
+	select VMCORE_INFO
+	select CRASH_RESERVE
 	help
 	  Generate crash dump after being started by kexec.
 	  This should be normally only set in special crash dump kernels

The change to enable CONFIG_CRASH_DUMP by default apparently broke the boot on 32-bit Power Macintosh systems which fail after GRUB with:

Quoted 5 lines

"Error: You can't boot a kdump kernel from OF!"

We may have to turn this off for 32-bit Power Macintosh systems.

See this thread on debian-powerpc ML: https://lists.debian.org/debian-powerpc/2024/07/msg00001.html

If so, fix need be made.

We may need change in ARCH_SUPPORTS_CRASH_DUMP of ppc, can you or anyone post a patch? I don't know how to identify 32-bit Power Macintosh.

arch/powerpc/Kconfig:
===
config ARCH_SUPPORTS_CRASH_DUMP

        def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)

config ARCH_SELECTS_CRASH_DUMP
def_bool y
depends on CRASH_DUMP
select RELOCATABLE if PPC64 || 44x || PPC_85xx
......
config PHYSICAL_START

        hex "Physical address where the kernel is loaded" if PHYSICAL_START_BOOL
        default "0x02000000" if PPC_BOOK3S && CRASH_DUMP && !NONSTATIC_KERNEL
        default "0x00000000"
#12 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Hi Baoquan,

Quoted 25 lines · Baoquan He
Quoted 8 lines

The change to enable CONFIG_CRASH_DUMP by default apparently broke the boot on 32-bit Power Macintosh systems which fail after GRUB with:

"Error: You can't boot a kdump kernel from OF!"

We may have to turn this off for 32-bit Power Macintosh systems.

See this thread on debian-powerpc ML: https://lists.debian.org/debian-powerpc/2024/07/msg00001.html

If so, fix need be made.

We may need change in ARCH_SUPPORTS_CRASH_DUMP of ppc, can you or anyone post a patch? I don't know how to identify 32-bit Power Macintosh.

arch/powerpc/Kconfig:
===
config ARCH_SUPPORTS_CRASH_DUMP

        def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)

config ARCH_SELECTS_CRASH_DUMP
def_bool y
depends on CRASH_DUMP
select RELOCATABLE if PPC64 || 44x || PPC_85xx
......
config PHYSICAL_START

Quoted 3 lines
        hex "Physical address where the kernel is loaded" if PHYSICAL_START_BOOL
        default "0x02000000" if PPC_BOOK3S && CRASH_DUMP && !NONSTATIC_KERNEL
        default "0x00000000"

I think the architecture does support crash dumps, but I think the kernel has to be booted from kexec in this case. Booting a kernel with CRASH_DUMP enabled won't work from Open Firmware. So, I think CRASH_DUMP should just be disabled for PPC_BOOK3S_32 by default and users who want to use it on these systems, will have to enable it explicitly.

Adrian

Signature · 4 lines
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913
author
#13 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Quoted 32 lines · John Paul Adrian Glaubitz

Hi Baoquan,

Quoted 25 lines · Baoquan He
Quoted 8 lines

The change to enable CONFIG_CRASH_DUMP by default apparently broke the boot on 32-bit Power Macintosh systems which fail after GRUB with:

"Error: You can't boot a kdump kernel from OF!"

We may have to turn this off for 32-bit Power Macintosh systems.

See this thread on debian-powerpc ML: https://lists.debian.org/debian-powerpc/2024/07/msg00001.html

If so, fix need be made.

We may need change in ARCH_SUPPORTS_CRASH_DUMP of ppc, can you or anyone post a patch? I don't know how to identify 32-bit Power Macintosh.

arch/powerpc/Kconfig:
===
config ARCH_SUPPORTS_CRASH_DUMP

        def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)

config ARCH_SELECTS_CRASH_DUMP
def_bool y
depends on CRASH_DUMP
select RELOCATABLE if PPC64 || 44x || PPC_85xx
......
config PHYSICAL_START

Quoted 3 lines
        hex "Physical address where the kernel is loaded" if PHYSICAL_START_BOOL
        default "0x02000000" if PPC_BOOK3S && CRASH_DUMP && !NONSTATIC_KERNEL
        default "0x00000000"
Quoted 5 lines

I think the architecture does support crash dumps, but I think the kernel has to be booted from kexec in this case. Booting a kernel with CRASH_DUMP enabled won't work from Open Firmware. So, I think CRASH_DUMP should just be disabled for PPC_BOOK3S_32 by default and users who want to use it on these systems, will have to enable it explicitly.

If so, below patch possiblly can fix it. Can you help check if it's OK?

From dd5318dc5dcd66521b31214f0e5921f258532ef8 Mon Sep 17 00:00:00 2001
From: Baoquan He <[email protected]>
Date: Fri, 23 Aug 2024 07:37:38 +0800
Subject: [PATCH] powerpc/crash: do not default to enable CRASH_DUMP for
PPC_BOOK3S_32 system
Content-type: text/plain

Recently it's reported that PowerPC macMini system failed to boot up. It's because CONFIG_CRASH_DUMP=y is set by default on the system since kernel 6.9, and that makes CONFIG_PHYSICAL_START not equaling 0 any more and causes failure of normal kernel bootup.

The link of error report can be found here:

https://lists.debian.org/debian-powerpc/2024/07/msg00001.html

And copy the code snippet here for reference:
arch/powerpc/Kconfig:
==================
config KERNEL_START
hex "Virtual address of kernel base" if KERNEL_START_BOOL default PAGE_OFFSET if PAGE_OFFSET_BOOL
default "0xc2000000" if CRASH_DUMP && !NONSTATIC_KERNEL default "0xc0000000"

So let's stop enabling CRASH_DUMP by default on PPC_BOOK3S_32.

---
 arch/powerpc/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/powerpc/Kconfig +1 -1
index d7b09b064a8a..dc5ca58be1d6 100644
@@ -680,7 +680,7 @@ config RELOCATABLE_TEST
 	  relocation code.
 
 config ARCH_SUPPORTS_CRASH_DUMP
-	def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)
+	def_bool PPC64 || PPC_85xx || (44x && !SMP)
 
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
Signature · 1 line
2.41.0
#14 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Quoted 1 line · Baoquan He

If so, below patch possiblly can fix it. Can you help check if it's OK?

That removes the possibility of enabling CRASH_DUMP on PPC_BOOK3S_32, even when booting via other mechanisms. Maybe it would be best to just make it not-default? Please take a look at this patch:

From d6e5fe3a45f46f1aa01914648c443291d956de9e Mon Sep 17 00:00:00 2001
From: Dave Vasilevsky <[email protected]>
Date: Thu, 22 Aug 2024 20:13:46 -0400
Subject: [PATCH] powerpc: Default to CRASH_DUMP=n when Open Firmware boot is
likely
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Open Firmware is unable to boot a kernel where PHYSICAL_START is non-zero, which occurs when CRASH_DUMP is on.

On PPC_BOOK3S_32, the most common way of booting is Open Firmware, so most users probably don't want CRASH_DUMP. Users booting via some other mechanism can turn it on explicitly.

---
 arch/arm/Kconfig       | 3 +++
 arch/arm64/Kconfig     | 3 +++
 arch/loongarch/Kconfig | 3 +++
 arch/mips/Kconfig      | 3 +++
 arch/powerpc/Kconfig   | 4 ++++
 arch/riscv/Kconfig     | 3 +++
 arch/s390/Kconfig      | 3 +++
 arch/sh/Kconfig        | 3 +++
 arch/x86/Kconfig       | 3 +++
 kernel/Kconfig.kexec   | 2 +-
 10 files changed, 29 insertions(+), 1 deletion(-)
arch/arm/Kconfig +3 -0
index 54b2bb817a7f..200995052690 100644
@@ -1597,6 +1597,9 @@ config ATAGS_PROC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config AUTO_ZRELADDR
 	bool "Auto calculation of the decompressed kernel image address" if !ARCH_MULTIPLATFORM
 	default !(ARCH_FOOTBRIDGE || ARCH_RPC || ARCH_SA1100)
arch/arm64/Kconfig +3 -0
index a2f8ff354ca6..43e08cc8204f 100644
@@ -1558,6 +1558,9 @@ config ARCH_DEFAULT_KEXEC_IMAGE_VERIFY_SIG
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
 	def_bool CRASH_RESERVE
 
arch/loongarch/Kconfig +3 -0
index 70f169210b52..ce232ddcd27d 100644
@@ -599,6 +599,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
 	depends on CRASH_DUMP
arch/mips/Kconfig +3 -0
index 60077e576935..b547f4304d0c 100644
@@ -2881,6 +2881,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config PHYSICAL_START
 	hex "Physical address where the kernel is loaded"
 	default "0xffffffff84000000"
arch/powerpc/Kconfig +4 -0
index d7b09b064a8a..0f3c1f958eac 100644
@@ -682,6 +682,10 @@ config RELOCATABLE_TEST
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)
 
+config ARCH_DEFAULT_CRASH_DUMP
+	bool
+	default y if !PPC_BOOK3S_32
+
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
 	depends on CRASH_DUMP
arch/riscv/Kconfig +3 -0
index 0f3cd7c3a436..eb247b5ee569 100644
@@ -880,6 +880,9 @@ config ARCH_SUPPORTS_KEXEC_PURGATORY
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
 	def_bool CRASH_RESERVE
 
arch/s390/Kconfig +3 -0
index a822f952f64a..05a1fb408471 100644
@@ -275,6 +275,9 @@ config ARCH_SUPPORTS_CRASH_DUMP
 	  This option also enables s390 zfcpdump.
 	  See also <file:Documentation/arch/s390/zfcpdump.rst>
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 menu "Processor type and features"
 
 config HAVE_MARCH_Z10_FEATURES
arch/sh/Kconfig +3 -0
index 1aa3c4a0c5b2..3a6338962636 100644
@@ -549,6 +549,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool BROKEN_ON_SMP
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SUPPORTS_KEXEC_JUMP
 	def_bool y
 
arch/x86/Kconfig +3 -0
index 007bab9f2a0e..aa4666bb9e9c 100644
@@ -2087,6 +2087,9 @@ config ARCH_SUPPORTS_KEXEC_JUMP
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool X86_64 || (X86_32 && HIGHMEM)
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SUPPORTS_CRASH_HOTPLUG
 	def_bool y
 
kernel/Kconfig.kexec +1 -1
index 6c34e63c88ff..4d111f871951 100644
@@ -97,7 +97,7 @@ config KEXEC_JUMP
 
 config CRASH_DUMP
 	bool "kernel crash dumps"
-	default y
+	default ARCH_DEFAULT_CRASH_DUMP
 	depends on ARCH_SUPPORTS_CRASH_DUMP
 	depends on KEXEC_CORE
 	select VMCORE_INFO
Signature · 1 line
2.34.1
author
#15 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Quoted 5 lines · Dave Vasilevsky
Quoted 1 line · Baoquan He

If so, below patch possiblly can fix it. Can you help check if it's OK?

That removes the possibility of enabling CRASH_DUMP on PPC_BOOK3S_32, even when booting via other mechanisms. Maybe it would be best to just make it not-default? Please take a look at this patch:

This is a good mimic of ARCH_DEFAULT_KEXEC_IMAGE_VERIFY_SIG and the correspondent KEXEC_IMAGE_VERIFY_SIG. It looks good to me, as long as no one complain we introduce too many knobs.

Can you post this formally so that people can review it?

Quoted 176 lines

From d6e5fe3a45f46f1aa01914648c443291d956de9e Mon Sep 17 00:00:00 2001
From: Dave Vasilevsky <[email protected]>
Date: Thu, 22 Aug 2024 20:13:46 -0400
Subject: [PATCH] powerpc: Default to CRASH_DUMP=n when Open Firmware boot is
likely
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Open Firmware is unable to boot a kernel where PHYSICAL_START is non-zero, which occurs when CRASH_DUMP is on.

On PPC_BOOK3S_32, the most common way of booting is Open Firmware, so most users probably don't want CRASH_DUMP. Users booting via some other mechanism can turn it on explicitly.

---
 arch/arm/Kconfig       | 3 +++
 arch/arm64/Kconfig     | 3 +++
 arch/loongarch/Kconfig | 3 +++
 arch/mips/Kconfig      | 3 +++
 arch/powerpc/Kconfig   | 4 ++++
 arch/riscv/Kconfig     | 3 +++
 arch/s390/Kconfig      | 3 +++
 arch/sh/Kconfig        | 3 +++
 arch/x86/Kconfig       | 3 +++
 kernel/Kconfig.kexec   | 2 +-
 10 files changed, 29 insertions(+), 1 deletion(-)
arch/arm/Kconfig +3 -0
index 54b2bb817a7f..200995052690 100644
@@ -1597,6 +1597,9 @@ config ATAGS_PROC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config AUTO_ZRELADDR
 	bool "Auto calculation of the decompressed kernel image address" if !ARCH_MULTIPLATFORM
 	default !(ARCH_FOOTBRIDGE || ARCH_RPC || ARCH_SA1100)
arch/arm64/Kconfig +3 -0
index a2f8ff354ca6..43e08cc8204f 100644
@@ -1558,6 +1558,9 @@ config ARCH_DEFAULT_KEXEC_IMAGE_VERIFY_SIG
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
 	def_bool CRASH_RESERVE
 
arch/loongarch/Kconfig +3 -0
index 70f169210b52..ce232ddcd27d 100644
@@ -599,6 +599,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
 	depends on CRASH_DUMP
arch/mips/Kconfig +3 -0
index 60077e576935..b547f4304d0c 100644
@@ -2881,6 +2881,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config PHYSICAL_START
 	hex "Physical address where the kernel is loaded"
 	default "0xffffffff84000000"
arch/powerpc/Kconfig +4 -0
index d7b09b064a8a..0f3c1f958eac 100644
@@ -682,6 +682,10 @@ config RELOCATABLE_TEST
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)
 
+config ARCH_DEFAULT_CRASH_DUMP
+	bool
+	default y if !PPC_BOOK3S_32
+
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
 	depends on CRASH_DUMP
arch/riscv/Kconfig +3 -0
index 0f3cd7c3a436..eb247b5ee569 100644
@@ -880,6 +880,9 @@ config ARCH_SUPPORTS_KEXEC_PURGATORY
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
 	def_bool CRASH_RESERVE
 
arch/s390/Kconfig +3 -0
index a822f952f64a..05a1fb408471 100644
@@ -275,6 +275,9 @@ config ARCH_SUPPORTS_CRASH_DUMP
 	  This option also enables s390 zfcpdump.
 	  See also <file:Documentation/arch/s390/zfcpdump.rst>
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 menu "Processor type and features"
 
 config HAVE_MARCH_Z10_FEATURES
arch/sh/Kconfig +3 -0
index 1aa3c4a0c5b2..3a6338962636 100644
@@ -549,6 +549,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool BROKEN_ON_SMP
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SUPPORTS_KEXEC_JUMP
 	def_bool y
 
arch/x86/Kconfig +3 -0
index 007bab9f2a0e..aa4666bb9e9c 100644
@@ -2087,6 +2087,9 @@ config ARCH_SUPPORTS_KEXEC_JUMP
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool X86_64 || (X86_32 && HIGHMEM)
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SUPPORTS_CRASH_HOTPLUG
 	def_bool y
 
kernel/Kconfig.kexec +1 -1
index 6c34e63c88ff..4d111f871951 100644
@@ -97,7 +97,7 @@ config KEXEC_JUMP
 
 config CRASH_DUMP
 	bool "kernel crash dumps"
-	default y
+	default ARCH_DEFAULT_CRASH_DUMP
 	depends on ARCH_SUPPORTS_CRASH_DUMP
 	depends on KEXEC_CORE
 	select VMCORE_INFO
Signature · 1 line
2.34.1
#16 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

In reply to Dave Vasilevsky
Quoted 164 lines · Dave Vasilevsky

From d6e5fe3a45f46f1aa01914648c443291d956de9e Mon Sep 17 00:00:00 2001
From: Dave Vasilevsky <[email protected]>
Date: Thu, 22 Aug 2024 20:13:46 -0400
Subject: [PATCH] powerpc: Default to CRASH_DUMP=n when Open Firmware boot is
likely
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Open Firmware is unable to boot a kernel where PHYSICAL_START is non-zero, which occurs when CRASH_DUMP is on.

On PPC_BOOK3S_32, the most common way of booting is Open Firmware, so most users probably don't want CRASH_DUMP. Users booting via some other mechanism can turn it on explicitly.

---
 arch/arm/Kconfig       | 3 +++
 arch/arm64/Kconfig     | 3 +++
 arch/loongarch/Kconfig | 3 +++
 arch/mips/Kconfig      | 3 +++
 arch/powerpc/Kconfig   | 4 ++++
 arch/riscv/Kconfig     | 3 +++
 arch/s390/Kconfig      | 3 +++
 arch/sh/Kconfig        | 3 +++
 arch/x86/Kconfig       | 3 +++
 kernel/Kconfig.kexec   | 2 +-
 10 files changed, 29 insertions(+), 1 deletion(-)
arch/arm/Kconfig +3 -0
index 54b2bb817a7f..200995052690 100644
@@ -1597,6 +1597,9 @@ config ATAGS_PROC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config AUTO_ZRELADDR
 	bool "Auto calculation of the decompressed kernel image address" if !ARCH_MULTIPLATFORM
 	default !(ARCH_FOOTBRIDGE || ARCH_RPC || ARCH_SA1100)
arch/arm64/Kconfig +3 -0
index a2f8ff354ca6..43e08cc8204f 100644
@@ -1558,6 +1558,9 @@ config ARCH_DEFAULT_KEXEC_IMAGE_VERIFY_SIG
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
 	def_bool CRASH_RESERVE
 
arch/loongarch/Kconfig +3 -0
index 70f169210b52..ce232ddcd27d 100644
@@ -599,6 +599,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
 	depends on CRASH_DUMP
arch/mips/Kconfig +3 -0
index 60077e576935..b547f4304d0c 100644
@@ -2881,6 +2881,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config PHYSICAL_START
 	hex "Physical address where the kernel is loaded"
 	default "0xffffffff84000000"
arch/powerpc/Kconfig +4 -0
index d7b09b064a8a..0f3c1f958eac 100644
@@ -682,6 +682,10 @@ config RELOCATABLE_TEST
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool PPC64 || PPC_BOOK3S_32 || PPC_85xx || (44x && !SMP)
 
+config ARCH_DEFAULT_CRASH_DUMP
+	bool
+	default y if !PPC_BOOK3S_32
+
 config ARCH_SELECTS_CRASH_DUMP
 	def_bool y
 	depends on CRASH_DUMP
arch/riscv/Kconfig +3 -0
index 0f3cd7c3a436..eb247b5ee569 100644
@@ -880,6 +880,9 @@ config ARCH_SUPPORTS_KEXEC_PURGATORY
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool y
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_HAS_GENERIC_CRASHKERNEL_RESERVATION
 	def_bool CRASH_RESERVE
 
arch/s390/Kconfig +3 -0
index a822f952f64a..05a1fb408471 100644
@@ -275,6 +275,9 @@ config ARCH_SUPPORTS_CRASH_DUMP
 	  This option also enables s390 zfcpdump.
 	  See also <file:Documentation/arch/s390/zfcpdump.rst>
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 menu "Processor type and features"
 
 config HAVE_MARCH_Z10_FEATURES
arch/sh/Kconfig +3 -0
index 1aa3c4a0c5b2..3a6338962636 100644
@@ -549,6 +549,9 @@ config ARCH_SUPPORTS_KEXEC
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool BROKEN_ON_SMP
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SUPPORTS_KEXEC_JUMP
 	def_bool y
 
arch/x86/Kconfig +3 -0
index 007bab9f2a0e..aa4666bb9e9c 100644
@@ -2087,6 +2087,9 @@ config ARCH_SUPPORTS_KEXEC_JUMP
 config ARCH_SUPPORTS_CRASH_DUMP
 	def_bool X86_64 || (X86_32 && HIGHMEM)
 
+config ARCH_DEFAULT_CRASH_DUMP
+	def_bool y
+
 config ARCH_SUPPORTS_CRASH_HOTPLUG
 	def_bool y
 
kernel/Kconfig.kexec +0 -0
index 6c34e63c88ff..4d111f871951 100644

kernel/Kconfig.kexec config KEXEC_JUMP · line 103

Quoted 8 lines
kernel/Kconfig.kexec +1 -1
@@ -97,7 +97,7 @@ config KEXEC_JUMP
 
 config CRASH_DUMP
 	bool "kernel crash dumps"
-	default y
+	default ARCH_DEFAULT_CRASH_DUMP
 	depends on ARCH_SUPPORTS_CRASH_DUMP
 	depends on KEXEC_CORE
 	select VMCORE_INFO

It should be disabled on m68k and sh by default as well.

Adrian

Signature · 4 lines
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913
#17 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Quoted 1 line · John Paul Adrian Glaubitz

It should be disabled on m68k and sh by default as well.

Sure, I can change that. What's the reasoning, so I can explain in my commit message?

-Dave

#18 lore raw

Re: [PATCH linux-next v3 05/14] crash: clean up kdump related config items

Quoted 4 lines · Dave Vasilevsky
Quoted 1 line · John Paul Adrian Glaubitz

It should be disabled on m68k and sh by default as well.

Sure, I can change that. What's the reasoning, so I can explain in my commit message?

Oh I don't think m68k even has ARCH_SUPPORTS_CRASH_DUMP, so it will always be off. My question still stands for sh though.

-Dave

author
#19 lore raw

PATCH v3 6/14 x86, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on x86 with some adjustments.

Here, also change some ifdefs or IS_ENABLED() check to more appropriate ones, e,g
- #ifdef CONFIG_KEXEC_CORE -> #ifdef CONFIG_CRASH_DUMP
- (!IS_ENABLED(CONFIG_KEXEC_CORE)) - > (!IS_ENABLED(CONFIG_CRASH_RESERVE))

---
 arch/x86/kernel/Makefile           | 4 ++--
 arch/x86/kernel/cpu/mshyperv.c     | 4 ++++
 arch/x86/kernel/kexec-bzimage64.c  | 4 ++++
 arch/x86/kernel/kvm.c              | 4 ++--
 arch/x86/kernel/machine_kexec_64.c | 3 +++
 arch/x86/kernel/reboot.c           | 2 +-
 arch/x86/kernel/setup.c            | 2 +-
 arch/x86/kernel/smp.c              | 2 +-
 arch/x86/xen/enlighten_hvm.c       | 4 ++++
 9 files changed, 22 insertions(+), 7 deletions(-)
arch/x86/kernel/Makefile +2 -2
index 913d4022131e..3668b1edef2d 100644
@@ -100,9 +100,9 @@ obj-$(CONFIG_TRACING)		+= trace.o
 obj-$(CONFIG_RETHOOK)		+= rethook.o
 obj-$(CONFIG_VMCORE_INFO)	+= vmcore_info_$(BITS).o
 obj-$(CONFIG_KEXEC_CORE)	+= machine_kexec_$(BITS).o
-obj-$(CONFIG_KEXEC_CORE)	+= relocate_kernel_$(BITS).o crash.o
+obj-$(CONFIG_KEXEC_CORE)	+= relocate_kernel_$(BITS).o
 obj-$(CONFIG_KEXEC_FILE)	+= kexec-bzimage64.o
-obj-$(CONFIG_CRASH_DUMP)	+= crash_dump_$(BITS).o
+obj-$(CONFIG_CRASH_DUMP)	+= crash_dump_$(BITS).o crash.o
 obj-y				+= kprobes/
 obj-$(CONFIG_MODULES)		+= module.o
 obj-$(CONFIG_X86_32)		+= doublefault_32.o
arch/x86/kernel/cpu/mshyperv.c +4 -0
index 01fa06dd06b6..f8163a59026b 100644
@@ -210,6 +210,7 @@ static void hv_machine_shutdown(void)
 		hyperv_cleanup();
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
 {
 	if (hv_crash_handler)
@@ -221,6 +222,7 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
+#endif
 #endif /* CONFIG_KEXEC_CORE */
 #endif /* CONFIG_HYPERV */
 
@@ -497,7 +499,9 @@ static void __init ms_hyperv_init_platform(void)
 
 #if IS_ENABLED(CONFIG_HYPERV) && defined(CONFIG_KEXEC_CORE)
 	machine_ops.shutdown = hv_machine_shutdown;
+#ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = hv_machine_crash_shutdown;
+#endif
 #endif
 	if (ms_hyperv.features & HV_ACCESS_TSC_INVARIANT) {
 		/*
arch/x86/kernel/kexec-bzimage64.c +4 -0
index 2a422e00ed4b..b55737b83a84 100644
@@ -263,11 +263,13 @@ setup_boot_parameters(struct kimage *image, struct boot_params *params,
 	memset(&params->hd0_info, 0, sizeof(params->hd0_info));
 	memset(&params->hd1_info, 0, sizeof(params->hd1_info));
 
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH) {
 		ret = crash_setup_memmap_entries(image, params);
 		if (ret)
 			return ret;
 	} else
+#endif
 		setup_e820_entries(params);
 
 	nr_e820_entries = params->e820_entries;
@@ -433,12 +435,14 @@ static void *bzImage64_load(struct kimage *image, char *kernel,
 		return ERR_PTR(-EINVAL);
 	}
 
+#ifdef CONFIG_CRASH_DUMP
 	/* Allocate and load backup region */
 	if (image->type == KEXEC_TYPE_CRASH) {
 		ret = crash_load_segments(image);
 		if (ret)
 			return ERR_PTR(ret);
 	}
+#endif
 
 	/*
 	 * Load purgatory. For 64bit entry point, purgatory  code can be
arch/x86/kernel/kvm.c +2 -2
index dfe9945b9bec..acfc2d3183bc 100644
@@ -769,7 +769,7 @@ static struct notifier_block kvm_pv_reboot_nb = {
  * won't be valid. In cases like kexec, in which you install a new kernel, this
  * means a random memory location will be kept being written.
  */
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 static void kvm_crash_shutdown(struct pt_regs *regs)
 {
 	kvm_guest_cpu_offline(true);
@@ -852,7 +852,7 @@ static void __init kvm_guest_init(void)
 	kvm_guest_cpu_init();
 #endif
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = kvm_crash_shutdown;
 #endif
 
arch/x86/kernel/machine_kexec_64.c +3 -0
index bc0a5348b4a6..b180d8e497c3 100644
@@ -508,6 +508,8 @@ int arch_kimage_file_post_load_cleanup(struct kimage *image)
 }
 #endif /* CONFIG_KEXEC_FILE */
 
+#ifdef CONFIG_CRASH_DUMP
+
 static int
 kexec_mark_range(unsigned long start, unsigned long end, bool protect)
 {
@@ -552,6 +554,7 @@ void arch_kexec_unprotect_crashkres(void)
 {
 	kexec_mark_crashkres(false);
 }
+#endif
 
 /*
  * During a traditional boot under SME, SME will encrypt the kernel,
arch/x86/kernel/reboot.c +1 -1
index 830425e6d38e..1287b0d5962f 100644
@@ -796,7 +796,7 @@ struct machine_ops machine_ops __ro_after_init = {
 	.emergency_restart = native_machine_emergency_restart,
 	.restart = native_machine_restart,
 	.halt = native_machine_halt,
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 	.crash_shutdown = native_machine_crash_shutdown,
 #endif
 };
arch/x86/kernel/setup.c +1 -1
index 84201071dfac..899d839a2954 100644
@@ -471,7 +471,7 @@ static void __init arch_reserve_crashkernel(void)
 	bool high = false;
 	int ret;
 
-	if (!IS_ENABLED(CONFIG_KEXEC_CORE))
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
 		return;
 
 	ret = parse_crashkernel(cmdline, memblock_phys_mem_size(),
arch/x86/kernel/smp.c +1 -1
index 2908e063d7d8..18266cc3d98c 100644
@@ -286,7 +286,7 @@ struct smp_ops smp_ops = {
 	.smp_cpus_done		= native_smp_cpus_done,
 
 	.stop_other_cpus	= native_stop_other_cpus,
-#if defined(CONFIG_KEXEC_CORE)
+#if defined(CONFIG_CRASH_DUMP)
 	.crash_stop_other_cpus	= kdump_nmi_shootdown_cpus,
 #endif
 	.smp_send_reschedule	= native_smp_send_reschedule,
arch/x86/xen/enlighten_hvm.c +4 -0
index 3f8c34707c50..09e3db7ff990 100644
@@ -149,12 +149,14 @@ static void xen_hvm_shutdown(void)
 		xen_reboot(SHUTDOWN_soft_reset);
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void xen_hvm_crash_shutdown(struct pt_regs *regs)
 {
 	native_machine_crash_shutdown(regs);
 	xen_reboot(SHUTDOWN_soft_reset);
 }
 #endif
+#endif
 
 static int xen_cpu_up_prepare_hvm(unsigned int cpu)
 {
@@ -236,8 +238,10 @@ static void __init xen_hvm_guest_init(void)
 
 #ifdef CONFIG_KEXEC_CORE
 	machine_ops.shutdown = xen_hvm_shutdown;
+#ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = xen_hvm_crash_shutdown;
 #endif
+#endif
 }
 
 static __init int xen_parse_nopv(char *arg)
Signature · 1 line
2.41.0
#20 lore raw

RE: [PATCH linux-next v3 06/14] x86, crash: wrap crash dumping code into crash related ifdefs

From: Baoquan He <[email protected]> Sent: Tuesday, January 23, 2024 9:13 PM

Quoted 54 lines

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on x86 with some adjustments.

Here, also change some ifdefs or IS_ENABLED() check to more appropriate ones, e,g
- #ifdef CONFIG_KEXEC_CORE -> #ifdef CONFIG_CRASH_DUMP
- (!IS_ENABLED(CONFIG_KEXEC_CORE)) - >
(!IS_ENABLED(CONFIG_CRASH_RESERVE))

---
 arch/x86/kernel/Makefile           | 4 ++--
 arch/x86/kernel/cpu/mshyperv.c     | 4 ++++
 arch/x86/kernel/kexec-bzimage64.c  | 4 ++++
 arch/x86/kernel/kvm.c              | 4 ++--
 arch/x86/kernel/machine_kexec_64.c | 3 +++
 arch/x86/kernel/reboot.c           | 2 +-
 arch/x86/kernel/setup.c            | 2 +-
 arch/x86/kernel/smp.c              | 2 +-
 arch/x86/xen/enlighten_hvm.c       | 4 ++++
 9 files changed, 22 insertions(+), 7 deletions(-)
arch/x86/kernel/Makefile +2 -2
index 913d4022131e..3668b1edef2d 100644
@@ -100,9 +100,9 @@ obj-$(CONFIG_TRACING)		+= trace.o
 obj-$(CONFIG_RETHOOK)		+= rethook.o
 obj-$(CONFIG_VMCORE_INFO)	+= vmcore_info_$(BITS).o
 obj-$(CONFIG_KEXEC_CORE)	+= machine_kexec_$(BITS).o
-obj-$(CONFIG_KEXEC_CORE)	+= relocate_kernel_$(BITS).o crash.o
+obj-$(CONFIG_KEXEC_CORE)	+= relocate_kernel_$(BITS).o
 obj-$(CONFIG_KEXEC_FILE)	+= kexec-bzimage64.o
-obj-$(CONFIG_CRASH_DUMP)	+= crash_dump_$(BITS).o
+obj-$(CONFIG_CRASH_DUMP)	+= crash_dump_$(BITS).o crash.o
 obj-y				+= kprobes/
 obj-$(CONFIG_MODULES)		+= module.o
 obj-$(CONFIG_X86_32)		+= doublefault_32.o
arch/x86/kernel/cpu/mshyperv.c +1 -0
index 01fa06dd06b6..f8163a59026b 100644
@@ -210,6 +210,7 @@ static void hv_machine_shutdown(void)
 		hyperv_cleanup();
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
 {
 	if (hv_crash_handler)

arch/x86/kernel/cpu/mshyperv.c static void hv_machine_crash_shutdown(struct · line 227

Quoted 6 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
@@ -221,6 +222,7 @@ static void hv_machine_crash_shutdown(struct
 pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
+#endif
 #endif /* CONFIG_KEXEC_CORE */

Note that the #ifdef CONFIG_CRASH_DUMP is nested inside #ifdef CONFIG_KEXEC_CODE here, and in the other Hyper-V code
just below. It's also nested in xen_hvm_guest_init() at the bottom
of this patch. But the KVM case of setting crash_shutdown is not nested -- you changed #ifdef CONFIG_KEXEC_CORE to #ifdef CONFIG_CRASH_DUMP.

I think both approaches work because CONFIG_CRASH_DUMP implies CONFIG_KEXEC_CORE, but I wonder if it would be better to *not* nest in all cases. I'd like to see the cases be consistent so in the future someone doesn't wonder why there's a difference (unless there's a reason for the difference that I missed).

Quoted 97 lines
arch/x86/kernel/cpu/mshyperv.c +2 -0
 #endif /* CONFIG_HYPERV */
@@ -497,7 +499,9 @@ static void __init ms_hyperv_init_platform(void)
 
 #if IS_ENABLED(CONFIG_HYPERV) && defined(CONFIG_KEXEC_CORE)
 	machine_ops.shutdown = hv_machine_shutdown;
+#ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = hv_machine_crash_shutdown;
+#endif
 #endif
 	if (ms_hyperv.features & HV_ACCESS_TSC_INVARIANT) {
 		/*
arch/x86/kernel/kexec- renamed from arch/x86/kernel/kexec-bzimage64.c +0 -0

bzimage64.c
index 2a422e00ed4b..b55737b83a84 100644

arch/x86/kernel/kexec-bzimage64.c +2 -0
@@ -263,11 +263,13 @@ setup_boot_parameters(struct kimage *image,
 struct boot_params *params,
 	memset(&params->hd0_info, 0, sizeof(params->hd0_info));
 	memset(&params->hd1_info, 0, sizeof(params->hd1_info));
 
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH) {
 		ret = crash_setup_memmap_entries(image, params);
 		if (ret)
 			return ret;
 	} else
+#endif
 		setup_e820_entries(params);
 
arch/x86/kernel/kexec-bzimage64.c +2 -0
 	nr_e820_entries = params->e820_entries;
@@ -433,12 +435,14 @@ static void *bzImage64_load(struct kimage *image, char *kernel,
 		return ERR_PTR(-EINVAL);
 	}
 
+#ifdef CONFIG_CRASH_DUMP
 	/* Allocate and load backup region */
 	if (image->type == KEXEC_TYPE_CRASH) {
 		ret = crash_load_segments(image);
 		if (ret)
 			return ERR_PTR(ret);
 	}
+#endif
 
 	/*
 	 * Load purgatory. For 64bit entry point, purgatory  code can be
arch/x86/kernel/kvm.c +2 -2
index dfe9945b9bec..acfc2d3183bc 100644
@@ -769,7 +769,7 @@ static struct notifier_block kvm_pv_reboot_nb = {
  * won't be valid. In cases like kexec, in which you install a new kernel, this
  * means a random memory location will be kept being written.
  */
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 static void kvm_crash_shutdown(struct pt_regs *regs)
 {
 	kvm_guest_cpu_offline(true);
@@ -852,7 +852,7 @@ static void __init kvm_guest_init(void)
 	kvm_guest_cpu_init();
 #endif
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = kvm_crash_shutdown;
 #endif
 
arch/x86/kernel/machine_kexec_64.c +2 -0
index bc0a5348b4a6..b180d8e497c3 100644
@@ -508,6 +508,8 @@ int arch_kimage_file_post_load_cleanup(struct
 kimage *image)
 }
 #endif /* CONFIG_KEXEC_FILE */
 
+#ifdef CONFIG_CRASH_DUMP
+
 static int
 kexec_mark_range(unsigned long start, unsigned long end, bool protect)

{

arch/x86/kernel/machine_kexec_64.c +1 -0
@@ -552,6 +554,7 @@ void arch_kexec_unprotect_crashkres(void)
 {
 	kexec_mark_crashkres(false);
 }
+#endif
 
 /*
  * During a traditional boot under SME, SME will encrypt the kernel,
arch/x86/kernel/reboot.c +0 -0
index 830425e6d38e..1287b0d5962f 100644

arch/x86/kernel/reboot.c struct machine_ops machine_ops __ro_after_init = { · line 802

Quoted 8 lines
arch/x86/kernel/reboot.c +1 -1
@@ -796,7 +796,7 @@ struct machine_ops machine_ops __ro_after_init = {
 	.emergency_restart = native_machine_emergency_restart,
 	.restart = native_machine_restart,
 	.halt = native_machine_halt,
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 	.crash_shutdown = native_machine_crash_shutdown,
 #endif
 };

Also in arch/x86/kernel/reboot.c, should the function machine_crash_shutdown() be updated with #ifdef CONFIG_CRASH_SHUTDOWN instead of #ifdef CONFIG_KEXEC_CORE?

Michael

Quoted 59 lines
arch/x86/kernel/setup.c +1 -1
index 84201071dfac..899d839a2954 100644
@@ -471,7 +471,7 @@ static void __init arch_reserve_crashkernel(void)
 	bool high = false;
 	int ret;
 
-	if (!IS_ENABLED(CONFIG_KEXEC_CORE))
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
 		return;
 
 	ret = parse_crashkernel(cmdline, memblock_phys_mem_size(),
arch/x86/kernel/smp.c +1 -1
index 2908e063d7d8..18266cc3d98c 100644
@@ -286,7 +286,7 @@ struct smp_ops smp_ops = {
 	.smp_cpus_done		= native_smp_cpus_done,
 
 	.stop_other_cpus	= native_stop_other_cpus,
-#if defined(CONFIG_KEXEC_CORE)
+#if defined(CONFIG_CRASH_DUMP)
 	.crash_stop_other_cpus	= kdump_nmi_shootdown_cpus,
 #endif
 	.smp_send_reschedule	= native_smp_send_reschedule,
arch/x86/xen/enlighten_hvm.c +4 -1
index 3f8c34707c50..09e3db7ff990 100644
@@ -149,12 +149,14 @@ static void xen_hvm_shutdown(void)
 		xen_reboot(SHUTDOWN_soft_reset);
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void xen_hvm_crash_shutdown(struct pt_regs *regs)
 {
 	native_machine_crash_shutdown(regs);
 	xen_reboot(SHUTDOWN_soft_reset);
 }
 #endif
+#endif
 
 static int xen_cpu_up_prepare_hvm(unsigned int cpu)
 {
@@ -236,8 +238,10 @@ static void __init xen_hvm_guest_init(void)
 
 #ifdef CONFIG_KEXEC_CORE
 	machine_ops.shutdown = xen_hvm_shutdown;
+#ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = xen_hvm_crash_shutdown;
 #endif
+#endif
 }
 
 static __init int xen_parse_nopv(char *arg)
--

2.41.0

author
#21 lore raw

Re: [PATCH linux-next v3 06/14] x86, crash: wrap crash dumping code into crash related ifdefs

Quoted 28 lines · Michael Kelley
Quoted 14 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
index 01fa06dd06b6..f8163a59026b 100644
@@ -210,6 +210,7 @@ static void hv_machine_shutdown(void)
 		hyperv_cleanup();
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
 {
 	if (hv_crash_handler)

arch/x86/kernel/cpu/mshyperv.c static void hv_machine_crash_shutdown(struct · line 227

Quoted 6 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
@@ -221,6 +222,7 @@ static void hv_machine_crash_shutdown(struct
 pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
+#endif
 #endif /* CONFIG_KEXEC_CORE */

Note that the #ifdef CONFIG_CRASH_DUMP is nested inside #ifdef CONFIG_KEXEC_CODE here, and in the other Hyper-V code
just below. It's also nested in xen_hvm_guest_init() at the bottom
of this patch. But the KVM case of setting crash_shutdown is not nested -- you changed #ifdef CONFIG_KEXEC_CORE to #ifdef CONFIG_CRASH_DUMP.

Quoted 5 lines

I think both approaches work because CONFIG_CRASH_DUMP implies CONFIG_KEXEC_CORE, but I wonder if it would be better to *not* nest in all cases. I'd like to see the cases be consistent so in the future someone doesn't wonder why there's a difference (unless there's a reason for the difference that I missed).

I agree with you, it's a great suggestion. Thanks.

Do you think below draft patch includes all changes you are concerned about?

arch/x86/kernel/cpu/mshyperv.c +6 -4
index f8163a59026b..2e8cd5a4ae85 100644
@@ -209,6 +209,7 @@ static void hv_machine_shutdown(void)
 	if (kexec_in_progress)
 		hyperv_cleanup();
 }
+#endif /* CONFIG_KEXEC_CORE */
 
 #ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
@@ -222,8 +223,7 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
-#endif
-#endif /* CONFIG_KEXEC_CORE */
+#endif /* CONFIG_CRASH_DUMP */
 #endif /* CONFIG_HYPERV */
 
 static uint32_t  __init ms_hyperv_platform(void)
@@ -497,9 +497,11 @@ static void __init ms_hyperv_init_platform(void)
 	no_timer_check = 1;
 #endif
 
-#if IS_ENABLED(CONFIG_HYPERV) && defined(CONFIG_KEXEC_CORE)
+#if IS_ENABLED(CONFIG_HYPERV)
+#if defined(CONFIG_KEXEC_CORE)
 	machine_ops.shutdown = hv_machine_shutdown;
-#ifdef CONFIG_CRASH_DUMP
+#endif
+#if defined(CONFIG_CRASH_DUMP)
 	machine_ops.crash_shutdown = hv_machine_crash_shutdown;
 #endif
 #endif
arch/x86/kernel/reboot.c +1 -1
index 1287b0d5962f..f3130f762784 100644
@@ -826,7 +826,7 @@ void machine_halt(void)
 	machine_ops.halt();
 }
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 void machine_crash_shutdown(struct pt_regs *regs)
 {
 	machine_ops.crash_shutdown(regs);
#22 lore raw

RE: [PATCH linux-next v3 06/14] x86, crash: wrap crash dumping code into crash related ifdefs

From: Baoquan He <[email protected]> Sent: Wednesday, January 24, 2024 8:10 PM

Quoted 35 lines · Michael Kelley
Quoted 27 lines
Quoted 14 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
index 01fa06dd06b6..f8163a59026b 100644
@@ -210,6 +210,7 @@ static void hv_machine_shutdown(void)
 		hyperv_cleanup();
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
 {
 	if (hv_crash_handler)

arch/x86/kernel/cpu/mshyperv.c static void hv_machine_crash_shutdown(struct pt_regs *regs) · line 226

Quoted 5 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
@@ -221,6 +222,7 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
+#endif
 #endif /* CONFIG_KEXEC_CORE */

Note that the #ifdef CONFIG_CRASH_DUMP is nested inside #ifdef CONFIG_KEXEC_CODE here, and in the other Hyper-V code
just below. It's also nested in xen_hvm_guest_init() at the bottom
of this patch. But the KVM case of setting crash_shutdown is not nested -- you changed #ifdef CONFIG_KEXEC_CORE to #ifdef CONFIG_CRASH_DUMP.

Quoted 5 lines

I think both approaches work because CONFIG_CRASH_DUMP implies CONFIG_KEXEC_CORE, but I wonder if it would be better to *not* nest in all cases. I'd like to see the cases be consistent so in the future someone doesn't wonder why there's a difference (unless there's a reason for the difference that I missed).

Quoted 4 lines

I agree with you, it's a great suggestion. Thanks.

Do you think below draft patch includes all changes you are concerned about?

Yes, these changes look good as a delta to your original patch.

But also look at xen_hvm_guest_init(). It looks like your original patch does nesting there as well, and it could probably be "un-nested".

Michael

Quoted 52 lines
arch/x86/kernel/cpu/mshyperv.c +2 -2
index f8163a59026b..2e8cd5a4ae85 100644
@@ -209,6 +209,7 @@ static void hv_machine_shutdown(void)
 	if (kexec_in_progress)
 		hyperv_cleanup();
 }
+#endif /* CONFIG_KEXEC_CORE */
 
 #ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
@@ -222,8 +223,7 @@ static void hv_machine_crash_shutdown(struct
 pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
-#endif
-#endif /* CONFIG_KEXEC_CORE */
+#endif /* CONFIG_CRASH_DUMP */
 #endif /* CONFIG_HYPERV */
 

static uint32_t __init ms_hyperv_platform(void)

arch/x86/kernel/cpu/mshyperv.c +4 -2
@@ -497,9 +497,11 @@ static void __init ms_hyperv_init_platform(void)
 	no_timer_check = 1;
 #endif
 
-#if IS_ENABLED(CONFIG_HYPERV) && defined(CONFIG_KEXEC_CORE)
+#if IS_ENABLED(CONFIG_HYPERV)
+#if defined(CONFIG_KEXEC_CORE)
 	machine_ops.shutdown = hv_machine_shutdown;
-#ifdef CONFIG_CRASH_DUMP
+#endif
+#if defined(CONFIG_CRASH_DUMP)
 	machine_ops.crash_shutdown = hv_machine_crash_shutdown;
 #endif
 #endif
arch/x86/kernel/reboot.c +1 -1
index 1287b0d5962f..f3130f762784 100644
@@ -826,7 +826,7 @@ void machine_halt(void)
 	machine_ops.halt();
 }
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 void machine_crash_shutdown(struct pt_regs *regs)
 {
 	machine_ops.crash_shutdown(regs);
author
#23 lore raw

Re: [PATCH linux-next v3 06/14] x86, crash: wrap crash dumping code into crash related ifdefs

Quoted 38 lines · Michael Kelley

From: Baoquan He <[email protected]> Sent: Wednesday, January 24, 2024 8:10 PM

Quoted 37 lines · Michael Kelley
Quoted 27 lines
Quoted 14 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
index 01fa06dd06b6..f8163a59026b 100644
@@ -210,6 +210,7 @@ static void hv_machine_shutdown(void)
 		hyperv_cleanup();
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
 {
 	if (hv_crash_handler)

arch/x86/kernel/cpu/mshyperv.c static void hv_machine_crash_shutdown(struct pt_regs *regs) · line 226

Quoted 5 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
@@ -221,6 +222,7 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
+#endif
 #endif /* CONFIG_KEXEC_CORE */

Note that the #ifdef CONFIG_CRASH_DUMP is nested inside #ifdef CONFIG_KEXEC_CODE here, and in the other Hyper-V code
just below. It's also nested in xen_hvm_guest_init() at the bottom
of this patch. But the KVM case of setting crash_shutdown is not nested -- you changed #ifdef CONFIG_KEXEC_CORE to #ifdef CONFIG_CRASH_DUMP.

Quoted 5 lines

I think both approaches work because CONFIG_CRASH_DUMP implies CONFIG_KEXEC_CORE, but I wonder if it would be better to *not* nest in all cases. I'd like to see the cases be consistent so in the future someone doesn't wonder why there's a difference (unless there's a reason for the difference that I missed).

I agree with you, it's a great suggestion. Thanks.

Quoted 7 lines
Quoted 2 lines

Do you think below draft patch includes all changes you are concerned about?

Yes, these changes look good as a delta to your original patch.

But also look at xen_hvm_guest_init(). It looks like your original patch does nesting there as well, and it could probably be "un-nested".

Right. I checked them all in arch/x86 this time, hope nothing is missed again. I can post a v4 to update this x86 patch later if no other concern. Thanks.

arch/x86/kernel/cpu/mshyperv.c +6 -4
index f8163a59026b..2e8cd5a4ae85 100644
@@ -209,6 +209,7 @@ static void hv_machine_shutdown(void)
 	if (kexec_in_progress)
 		hyperv_cleanup();
 }
+#endif /* CONFIG_KEXEC_CORE */
 
 #ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
@@ -222,8 +223,7 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
-#endif
-#endif /* CONFIG_KEXEC_CORE */
+#endif /* CONFIG_CRASH_DUMP */
 #endif /* CONFIG_HYPERV */
 
 static uint32_t  __init ms_hyperv_platform(void)
@@ -497,9 +497,11 @@ static void __init ms_hyperv_init_platform(void)
 	no_timer_check = 1;
 #endif
 
-#if IS_ENABLED(CONFIG_HYPERV) && defined(CONFIG_KEXEC_CORE)
+#if IS_ENABLED(CONFIG_HYPERV)
+#if defined(CONFIG_KEXEC_CORE)
 	machine_ops.shutdown = hv_machine_shutdown;
-#ifdef CONFIG_CRASH_DUMP
+#endif
+#if defined(CONFIG_CRASH_DUMP)
 	machine_ops.crash_shutdown = hv_machine_crash_shutdown;
 #endif
 #endif
arch/x86/kernel/reboot.c +1 -1
index 1287b0d5962f..f3130f762784 100644
@@ -826,7 +826,7 @@ void machine_halt(void)
 	machine_ops.halt();
 }
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 void machine_crash_shutdown(struct pt_regs *regs)
 {
 	machine_ops.crash_shutdown(regs);
arch/x86/xen/enlighten_hvm.c +2 -2
index 09e3db7ff990..0b367c1e086d 100644
@@ -148,6 +148,7 @@ static void xen_hvm_shutdown(void)
 	if (kexec_in_progress)
 		xen_reboot(SHUTDOWN_soft_reset);
 }
+#endif
 
 #ifdef CONFIG_CRASH_DUMP
 static void xen_hvm_crash_shutdown(struct pt_regs *regs)
@@ -156,7 +157,6 @@ static void xen_hvm_crash_shutdown(struct pt_regs *regs)
 	xen_reboot(SHUTDOWN_soft_reset);
 }
 #endif
-#endif
 
 static int xen_cpu_up_prepare_hvm(unsigned int cpu)
 {
@@ -238,10 +238,10 @@ static void __init xen_hvm_guest_init(void)
 
 #ifdef CONFIG_KEXEC_CORE
 	machine_ops.shutdown = xen_hvm_shutdown;
+#endif
 #ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = xen_hvm_crash_shutdown;
 #endif
-#endif
 }
 
 static __init int xen_parse_nopv(char *arg)
#24 lore raw

RE: [PATCH linux-next v3 06/14] x86, crash: wrap crash dumping code into crash related ifdefs

From: Baoquan He <[email protected]> Sent: Thursday, January 25, 2024 1:17 AM

Quoted 44 lines · Michael Kelley
Quoted 1 line

From: Baoquan He <[email protected]> Sent: Wednesday, January 24, 2024

8:10 PM

Quoted 40 lines
Quoted 39 lines · Michael Kelley
Quoted 27 lines
Quoted 14 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
index 01fa06dd06b6..f8163a59026b 100644
@@ -210,6 +210,7 @@ static void hv_machine_shutdown(void)
 		hyperv_cleanup();
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
 {
 	if (hv_crash_handler)

arch/x86/kernel/cpu/mshyperv.c static void hv_machine_crash_shutdown(struct pt_regs *regs) · line 226

Quoted 5 lines
arch/x86/kernel/cpu/mshyperv.c +1 -0
@@ -221,6 +222,7 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
+#endif
 #endif /* CONFIG_KEXEC_CORE */

Note that the #ifdef CONFIG_CRASH_DUMP is nested inside #ifdef CONFIG_KEXEC_CODE here, and in the other Hyper-V code
just below. It's also nested in xen_hvm_guest_init() at the bottom
of this patch. But the KVM case of setting crash_shutdown is not nested -- you changed #ifdef CONFIG_KEXEC_CORE to #ifdef CONFIG_CRASH_DUMP.

Quoted 5 lines

I think both approaches work because CONFIG_CRASH_DUMP implies CONFIG_KEXEC_CORE, but I wonder if it would be better to *not* nest in all cases. I'd like to see the cases be consistent so in the future someone doesn't wonder why there's a difference (unless there's a reason for the difference that I missed).

I agree with you, it's a great suggestion. Thanks.

Do you think below draft patch includes all changes you are concerned about?

Quoted 8 lines
Quoted 4 lines

Yes, these changes look good as a delta to your original patch.

But also look at xen_hvm_guest_init(). It looks like your original patch does nesting there as well, and it could probably be "un-nested".

Right. I checked them all in arch/x86 this time, hope nothing is missed again. I can post a v4 to update this x86 patch later if no other concern. Thanks.

Yes -- everything looks good to me now.

Michael

Quoted 85 lines
arch/x86/kernel/cpu/mshyperv.c +2 -2
index f8163a59026b..2e8cd5a4ae85 100644
@@ -209,6 +209,7 @@ static void hv_machine_shutdown(void)
 	if (kexec_in_progress)
 		hyperv_cleanup();
 }
+#endif /* CONFIG_KEXEC_CORE */
 
 #ifdef CONFIG_CRASH_DUMP
 static void hv_machine_crash_shutdown(struct pt_regs *regs)
@@ -222,8 +223,7 @@ static void hv_machine_crash_shutdown(struct
 pt_regs *regs)
 	/* Disable the hypercall page when there is only 1 active CPU. */
 	hyperv_cleanup();
 }
-#endif
-#endif /* CONFIG_KEXEC_CORE */
+#endif /* CONFIG_CRASH_DUMP */
 #endif /* CONFIG_HYPERV */
 

static uint32_t __init ms_hyperv_platform(void)

arch/x86/kernel/cpu/mshyperv.c +4 -2
@@ -497,9 +497,11 @@ static void __init ms_hyperv_init_platform(void)
 	no_timer_check = 1;
 #endif
 
-#if IS_ENABLED(CONFIG_HYPERV) && defined(CONFIG_KEXEC_CORE)
+#if IS_ENABLED(CONFIG_HYPERV)
+#if defined(CONFIG_KEXEC_CORE)
 	machine_ops.shutdown = hv_machine_shutdown;
-#ifdef CONFIG_CRASH_DUMP
+#endif
+#if defined(CONFIG_CRASH_DUMP)
 	machine_ops.crash_shutdown = hv_machine_crash_shutdown;
 #endif
 #endif
arch/x86/kernel/reboot.c +1 -1
index 1287b0d5962f..f3130f762784 100644
@@ -826,7 +826,7 @@ void machine_halt(void)
 	machine_ops.halt();
 }
 
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_DUMP
 void machine_crash_shutdown(struct pt_regs *regs)
 {
 	machine_ops.crash_shutdown(regs);
arch/x86/xen/enlighten_hvm.c +1 -1
index 09e3db7ff990..0b367c1e086d 100644
@@ -148,6 +148,7 @@ static void xen_hvm_shutdown(void)
 	if (kexec_in_progress)
 		xen_reboot(SHUTDOWN_soft_reset);
 }
+#endif
 
 #ifdef CONFIG_CRASH_DUMP
 static void xen_hvm_crash_shutdown(struct pt_regs *regs)
@@ -156,7 +157,6 @@ static void xen_hvm_crash_shutdown(struct pt_regs
 *regs)
 	xen_reboot(SHUTDOWN_soft_reset);
 }
 #endif
-#endif
 
 static int xen_cpu_up_prepare_hvm(unsigned int cpu)

{

arch/x86/xen/enlighten_hvm.c +1 -1
@@ -238,10 +238,10 @@ static void __init xen_hvm_guest_init(void)
 
 #ifdef CONFIG_KEXEC_CORE
 	machine_ops.shutdown = xen_hvm_shutdown;
+#endif
 #ifdef CONFIG_CRASH_DUMP
 	machine_ops.crash_shutdown = xen_hvm_crash_shutdown;
 #endif
-#endif
 }
 
 static __init int xen_parse_nopv(char *arg)
author
#25 lore raw

PATCH v3 7/14 arm64, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on arm64 with some adjustments.

Here wrap up crash dumping codes with CONFIG_CRASH_DUMP ifdeffery.

---
 arch/arm64/include/asm/kexec.h         |  2 +-
 arch/arm64/kernel/machine_kexec.c      |  2 +-
 arch/arm64/kernel/machine_kexec_file.c | 10 ++++++++--
 arch/arm64/mm/init.c                   |  2 +-
 4 files changed, 11 insertions(+), 5 deletions(-)
arch/arm64/include/asm/kexec.h +1 -1
index 9ac9572a3bbe..4d9cc7a76d9c 100644
@@ -80,7 +80,7 @@ static inline void crash_setup_regs(struct pt_regs *newregs,
 	}
 }
 
-#if defined(CONFIG_KEXEC_CORE) && defined(CONFIG_HIBERNATION)
+#if defined(CONFIG_CRASH_DUMP) && defined(CONFIG_HIBERNATION)
 extern bool crash_is_nosave(unsigned long pfn);
 extern void crash_prepare_suspend(void);
 extern void crash_post_resume(void);
arch/arm64/kernel/machine_kexec.c +1 -1
index b38aae5b488d..82e2203d86a3 100644
@@ -255,7 +255,7 @@ void machine_crash_shutdown(struct pt_regs *regs)
 	pr_info("Starting crashdump kernel...\n");
 }
 
-#ifdef CONFIG_HIBERNATION
+#if defined(CONFIG_CRASH_DUMP) && defined(CONFIG_HIBERNATION)
 /*
  * To preserve the crash dump kernel image, the relevant memory segments
  * should be mapped again around the hibernation.
arch/arm64/kernel/machine_kexec_file.c +8 -2
index 0e017358f4ba..af1ca875c52c 100644
@@ -39,6 +39,7 @@ int arch_kimage_file_post_load_cleanup(struct kimage *image)
 	return kexec_image_post_load_cleanup_default(image);
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static int prepare_elf_headers(void **addr, unsigned long *sz)
 {
 	struct crash_mem *cmem;
@@ -80,6 +81,7 @@ static int prepare_elf_headers(void **addr, unsigned long *sz)
 	kfree(cmem);
 	return ret;
 }
+#endif
 
 /*
  * Tries to add the initrd and DTB to the image. If it is not possible to find
@@ -93,8 +95,8 @@ int load_other_segments(struct kimage *image,
 			char *cmdline)
 {
 	struct kexec_buf kbuf;
-	void *headers, *dtb = NULL;
-	unsigned long headers_sz, initrd_load_addr = 0, dtb_len,
+	void *dtb = NULL;
+	unsigned long initrd_load_addr = 0, dtb_len,
 		      orig_segments = image->nr_segments;
 	int ret = 0;
 
@@ -102,7 +104,10 @@ int load_other_segments(struct kimage *image,
 	/* not allocate anything below the kernel */
 	kbuf.buf_min = kernel_load_addr + kernel_size;
 
+#ifdef CONFIG_CRASH_DUMP
 	/* load elf core header */
+	void *headers;
+	unsigned long headers_sz;
 	if (image->type == KEXEC_TYPE_CRASH) {
 		ret = prepare_elf_headers(&headers, &headers_sz);
 		if (ret) {
@@ -130,6 +135,7 @@ int load_other_segments(struct kimage *image,
 		kexec_dprintk("Loaded elf core header at 0x%lx bufsz=0x%lx memsz=0x%lx\n",
 			      image->elf_load_addr, kbuf.bufsz, kbuf.memsz);
 	}
+#endif
 
 	/* load initrd */
 	if (initrd) {
arch/arm64/mm/init.c +1 -1
index 74c1db8ce271..c1f6213e77f3 100644
@@ -100,7 +100,7 @@ static void __init arch_reserve_crashkernel(void)
 	bool high = false;
 	int ret;
 
-	if (!IS_ENABLED(CONFIG_KEXEC_CORE))
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
 		return;
 
 	ret = parse_crashkernel(cmdline, memblock_phys_mem_size(),
Signature · 1 line
2.41.0
author
#26 lore raw

PATCH v3 8/14 ppc, crash: enforce KEXEC and KEXEC_FILE to select CRASH_DUMP

In reply to Baoquan He

In PowerPC, the crash dumping and kexec reboot share code in arch_kexec_locate_mem_hole(), in which struct crash_mem is used.

Here enfoce enforce KEXEC and KEXEC_FILE to select CRASH_DUMP for now.

---
 arch/powerpc/Kconfig | 5 +++++
 1 file changed, 5 insertions(+)
arch/powerpc/Kconfig +5 -0
index e66fd9923250..31f013e636e3 100644
@@ -608,6 +608,10 @@ config PPC64_SUPPORTS_MEMORY_FAILURE
 config ARCH_SUPPORTS_KEXEC
 	def_bool PPC_BOOK3S || PPC_E500 || (44x && !SMP)
 
+config ARCH_SELECTS_KEXEC
+	def_bool y
+	select CRASH_DUMP
+
 config ARCH_SUPPORTS_KEXEC_FILE
 	def_bool PPC64
 
@@ -618,6 +622,7 @@ config ARCH_SELECTS_KEXEC_FILE
 	def_bool y
 	depends on KEXEC_FILE
 	select KEXEC_ELF
+	select CRASH_DUMP
 	select HAVE_IMA_KEXEC if IMA
 
 config PPC64_BIG_ENDIAN_ELF_ABI_V2
Signature · 1 line
2.41.0
author
#27 lore raw

PATCH v3 9/14 s390, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on s390 with some adjustments.

Here wrap up crash dumping codes with CONFIG_CRASH_DUMP ifdeffery.

---
 arch/s390/kernel/kexec_elf.c          |  2 ++
 arch/s390/kernel/kexec_image.c        |  2 ++
 arch/s390/kernel/machine_kexec_file.c | 10 ++++++++++
 3 files changed, 14 insertions(+)
arch/s390/kernel/kexec_elf.c +2 -0
index 9da6fa30c447..4d364de43799 100644
@@ -40,8 +40,10 @@ static int kexec_file_add_kernel_elf(struct kimage *image,
 		buf.bufsz = phdr->p_filesz;
 
 		buf.mem = ALIGN(phdr->p_paddr, phdr->p_align);
+#ifdef CONFIG_CRASH_DUMP
 		if (image->type == KEXEC_TYPE_CRASH)
 			buf.mem += crashk_res.start;
+#endif
 		buf.memsz = phdr->p_memsz;
 		data->memsz = ALIGN(data->memsz, phdr->p_align) + buf.memsz;
 
arch/s390/kernel/kexec_image.c +2 -0
index af23eff5774d..a32ce8bea745 100644
@@ -24,8 +24,10 @@ static int kexec_file_add_kernel_image(struct kimage *image,
 	buf.bufsz = image->kernel_buf_len;
 
 	buf.mem = 0;
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH)
 		buf.mem += crashk_res.start;
+#endif
 	buf.memsz = buf.bufsz;
 
 	data->kernel_buf = image->kernel_buf;
arch/s390/kernel/machine_kexec_file.c +10 -0
index 8d207b82d9fe..c2bac14dd668 100644
@@ -105,6 +105,7 @@ static int kexec_file_update_purgatory(struct kimage *image,
 	if (ret)
 		return ret;
 
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH) {
 		u64 crash_size;
 
@@ -121,6 +122,7 @@ static int kexec_file_update_purgatory(struct kimage *image,
 						     sizeof(crash_size),
 						     false);
 	}
+#endif
 	return ret;
 }
 
@@ -134,8 +136,10 @@ static int kexec_file_add_purgatory(struct kimage *image,
 
 	data->memsz = ALIGN(data->memsz, PAGE_SIZE);
 	buf.mem = data->memsz;
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH)
 		buf.mem += crashk_res.start;
+#endif
 
 	ret = kexec_load_purgatory(image, &buf);
 	if (ret)
@@ -158,8 +162,10 @@ static int kexec_file_add_initrd(struct kimage *image,
 
 	data->memsz = ALIGN(data->memsz, PAGE_SIZE);
 	buf.mem = data->memsz;
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH)
 		buf.mem += crashk_res.start;
+#endif
 	buf.memsz = buf.bufsz;
 
 	data->parm->initrd_start = data->memsz;
@@ -223,8 +229,10 @@ static int kexec_file_add_ipl_report(struct kimage *image,
 		data->kernel_buf + offsetof(struct lowcore, ipl_parmblock_ptr);
 	*lc_ipl_parmblock_ptr = (__u32)buf.mem;
 
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH)
 		buf.mem += crashk_res.start;
+#endif
 
 	ret = kexec_add_buffer(&buf);
 out:
@@ -268,10 +276,12 @@ void *kexec_file_add_components(struct kimage *image,
 	memcpy(data.parm->command_line, image->cmdline_buf,
 	       image->cmdline_buf_len);
 
+#ifdef CONFIG_CRASH_DUMP
 	if (image->type == KEXEC_TYPE_CRASH) {
 		data.parm->oldmem_base = crashk_res.start;
 		data.parm->oldmem_size = crashk_res.end - crashk_res.start + 1;
 	}
+#endif
 
 	if (image->initrd_buf) {
 		ret = kexec_file_add_initrd(image, &data);
Signature · 1 line
2.41.0
author
#28 lore raw

PATCH v3 10/14 sh, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on SuperH with some adjustments.

wrap up crash dumping codes with CONFIG_CRASH_DUMP ifdeffery, and use IS_ENABLED(CONFIG_CRASH_RESERVE) check to decide if compiling in the crashkernel reservation code.

---
 arch/sh/kernel/machine_kexec.c | 3 +++
 arch/sh/kernel/setup.c         | 2 +-
 2 files changed, 4 insertions(+), 1 deletion(-)
arch/sh/kernel/machine_kexec.c +3 -0
index fa3a7b36190a..8daa8a6e6fa6 100644
@@ -153,6 +153,9 @@ void __init reserve_crashkernel(void)
 	unsigned long long crash_size, crash_base;
 	int ret;
 
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
+		return;
+
 	ret = parse_crashkernel(boot_command_line, memblock_phys_mem_size(),
 			&crash_size, &crash_base, NULL, NULL);
 	if (ret == 0 && crash_size > 0) {
arch/sh/kernel/setup.c +1 -1
index d3175f09b3aa..620e5cf8ae1e 100644
@@ -220,7 +220,7 @@ void __init __add_active_range(unsigned int nid, unsigned long start_pfn,
 	request_resource(res, &code_resource);
 	request_resource(res, &data_resource);
 	request_resource(res, &bss_resource);
-#ifdef CONFIG_KEXEC_CORE
+#ifdef CONFIG_CRASH_RESERVE
 	request_resource(res, &crashk_res);
 #endif
 
Signature · 1 line
2.41.0
#29 lore raw

Re: [PATCH linux-next v3 10/14] sh, crash: wrap crash dumping code into crash related ifdefs

Hello Baoquan,

Quoted 7 lines · Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on SuperH with some adjustments.

wrap up crash dumping codes with CONFIG_CRASH_DUMP ifdeffery, and use IS_ENABLED(CONFIG_CRASH_RESERVE) check to decide if compiling in the crashkernel reservation code.

Comparing this to the patches, it seems you missed the first word "Here". Either amend that or write the word "wrap" capitalized.

I would omit "Here" as it's not necessary and just start the sentence with "Wrap".

Adrian

Signature · 4 lines
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913
author
#30 lore raw

Re: [PATCH linux-next v3 10/14] sh, crash: wrap crash dumping code into crash related ifdefs

Quoted 11 lines · John Paul Adrian Glaubitz

Hello Baoquan,

Quoted 7 lines · Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on SuperH with some adjustments.

wrap up crash dumping codes with CONFIG_CRASH_DUMP ifdeffery, and use IS_ENABLED(CONFIG_CRASH_RESERVE) check to decide if compiling in the crashkernel reservation code.

Quoted 5 lines

Comparing this to the patches, it seems you missed the first word "Here". Either amend that or write the word "wrap" capitalized.

I would omit "Here" as it's not necessary and just start the sentence with "Wrap".

You are right, thanks for careful checking. I will see if I need repost to include this update.

Quoted 12 lines
Signature · 10 lines
-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
author
#31 lore raw

PATCH v3 11/14 mips, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on mips with some adjustments.

Here use IS_ENABLED(CONFIG_CRASH_RESERVE) check to decide if compiling in the crashkernel reservation code.

---
 arch/mips/kernel/setup.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)
arch/mips/kernel/setup.c +6 -11
index 9c30de151597..12a1a4ffb602 100644
@@ -442,8 +442,6 @@ static void __init mips_reserve_vmcore(void)
 #endif
 }
 
-#ifdef CONFIG_KEXEC
-
 /* 64M alignment for crash kernel regions */
 #define CRASH_ALIGN	SZ_64M
 #define CRASH_ADDR_MAX	SZ_512M
@@ -454,6 +452,9 @@ static void __init mips_parse_crashkernel(void)
 	unsigned long long crash_size, crash_base;
 	int ret;
 
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
+		return;
+
 	total_mem = memblock_phys_mem_size();
 	ret = parse_crashkernel(boot_command_line, total_mem,
 				&crash_size, &crash_base,
@@ -489,6 +490,9 @@ static void __init request_crashkernel(struct resource *res)
 {
 	int ret;
 
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
+		return;
+
 	if (crashk_res.start == crashk_res.end)
 		return;
 
@@ -498,15 +502,6 @@ static void __init request_crashkernel(struct resource *res)
 			(unsigned long)(resource_size(&crashk_res) >> 20),
 			(unsigned long)(crashk_res.start  >> 20));
 }
-#else /* !defined(CONFIG_KEXEC)		*/
-static void __init mips_parse_crashkernel(void)
-{
-}
-
-static void __init request_crashkernel(struct resource *res)
-{
-}
-#endif /* !defined(CONFIG_KEXEC)  */
 
 static void __init check_kernel_sections_mem(void)
 {
Signature · 1 line
2.41.0
author
#32 lore raw

PATCH v3 12/14 riscv, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on risc-v with some adjustments.

Here wrap up crash dumping codes with CONFIG_CRASH_DUMP ifdeffery, and use IS_ENABLED(CONFIG_CRASH_RESERVE) check to decide if compiling in the crashkernel reservation code.

---
 arch/riscv/kernel/elf_kexec.c | 9 +++++++--
 arch/riscv/mm/init.c          | 2 +-
 2 files changed, 8 insertions(+), 3 deletions(-)
arch/riscv/kernel/elf_kexec.c +7 -2
index 5bd1ec3341fe..54260c16f991 100644
@@ -117,6 +117,7 @@ static int elf_find_pbase(struct kimage *image, unsigned long kernel_len,
 	return ret;
 }
 
+#ifdef CONFIG_CRASH_DUMP
 static int get_nr_ram_ranges_callback(struct resource *res, void *arg)
 {
 	unsigned int *nr_ranges = arg;
@@ -189,6 +190,7 @@ static char *setup_kdump_cmdline(struct kimage *image, char *cmdline,
 	cmdline_ptr[COMMAND_LINE_SIZE - 1] = '\0';
 	return cmdline_ptr;
 }
+#endif
 
 static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
 			    unsigned long kernel_len, char *initrd,
@@ -196,12 +198,11 @@ static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
 			    unsigned long cmdline_len)
 {
 	int ret;
+	void *fdt;
 	unsigned long old_kernel_pbase = ULONG_MAX;
 	unsigned long new_kernel_pbase = 0UL;
 	unsigned long initrd_pbase = 0UL;
-	unsigned long headers_sz;
 	unsigned long kernel_start;
-	void *fdt, *headers;
 	struct elfhdr ehdr;
 	struct kexec_buf kbuf;
 	struct kexec_elf_info elf_info;
@@ -227,8 +228,11 @@ static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
 	kbuf.buf_min = new_kernel_pbase + kernel_len;
 	kbuf.buf_max = ULONG_MAX;
 
+#ifdef CONFIG_CRASH_DUMP
 	/* Add elfcorehdr */
 	if (image->type == KEXEC_TYPE_CRASH) {
+		void *headers;
+		unsigned long headers_sz;
 		ret = prepare_elf_headers(&headers, &headers_sz);
 		if (ret) {
 			pr_err("Preparing elf core header failed\n");
@@ -264,6 +268,7 @@ static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
 		}
 		cmdline = modified_cmdline;
 	}
+#endif
 
 #ifdef CONFIG_ARCH_SUPPORTS_KEXEC_PURGATORY
 	/* Add purgatory to the image */
arch/riscv/mm/init.c +1 -1
index 32cad6a65ccd..245919dda910 100644
@@ -1358,7 +1358,7 @@ static void __init arch_reserve_crashkernel(void)
 	bool high = false;
 	int ret;
 
-	if (!IS_ENABLED(CONFIG_KEXEC_CORE))
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
 		return;
 
 	ret = parse_crashkernel(cmdline, memblock_phys_mem_size(),
Signature · 1 line
2.41.0
author
#33 lore raw

PATCH v3 13/14 arm, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on arm with some adjustments.

Here use CONFIG_CRASH_RESERVE ifdef to replace CONFIG_KEXEC ifdef.

---
v2->v3:
- Fix the lkp reported issue by using CONFIG_CRASH_RESERVE ifdef, giving up the earlier IS_ENABLED(CONFIG_CRASH_RESERVE) checking in v2.

 arch/arm/kernel/setup.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/arm/kernel/setup.c +2 -2
index ff2299ce1ad7..7b33b157fca0 100644
@@ -979,7 +979,7 @@ static int __init init_machine_late(void)
 }
 late_initcall(init_machine_late);
 
-#ifdef CONFIG_KEXEC
+#ifdef CONFIG_CRASH_RESERVE
 /*
  * The crash region must be aligned to 128MB to avoid
  * zImage relocating below the reserved region.
@@ -1066,7 +1066,7 @@ static void __init reserve_crashkernel(void)
 }
 #else
 static inline void reserve_crashkernel(void) {}
-#endif /* CONFIG_KEXEC */
+#endif /* CONFIG_CRASH_RESERVE*/
 
 void __init hyp_mode_check(void)
 {
Signature · 1 line
2.41.0
author
#34 lore raw

PATCH v3 14/14 loongarch, crash: wrap crash dumping code into crash related ifdefs

In reply to Baoquan He

Now crash codes under kernel/ folder has been split out from kexec code, crash dumping can be separated from kexec reboot in config items on loongarch with some adjustments.

Here use IS_ENABLED(CONFIG_CRASH_RESERVE) check to decide if compiling in the crashkernel reservation code.

---
v2->v3:
- Update code change regarding below commit:
commit 78de91b45860 ("LoongArch: Use generic interface to support crashkernel=X,[high,low]")

 arch/loongarch/kernel/setup.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/loongarch/kernel/setup.c +1 -1
index edf2bba80130..57d37dd9f964 100644
@@ -260,7 +260,7 @@ static void __init arch_reserve_crashkernel(void)
 	char *cmdline = boot_command_line;
 	bool high = false;
 
-	if (!IS_ENABLED(CONFIG_KEXEC_CORE))
+	if (!IS_ENABLED(CONFIG_CRASH_RESERVE))
 		return;
 
 	ret = parse_crashkernel(cmdline, memblock_phys_mem_size(),
Signature · 1 line
2.41.0
#35 lore raw
In reply to Baoquan He

Hi Baoquan,

Quoted 134 lines · Baoquan He

Motivation:
=============
Previously, LKP reported a building error. When investigating, it can't be resolved reasonablly with the present messy kdump config items.

https://lore.kernel.org/oe-kbuild-all/[email protected]/

The kdump (crash dumping) related config items could causes confusions:

Firstly,
---
CRASH_CORE enables codes including
- crashkernel reservation;
- elfcorehdr updating;
- vmcoreinfo exporting;
- crash hotplug handling;

Now fadump of powerpc, kcore dynamic debugging and kdump all selects CRASH_CORE, while fadump
- fadump needs crashkernel parsing, vmcoreinfo exporting, and accessing global variable 'elfcorehdr_addr';
- kcore only needs vmcoreinfo exporting;
- kdump needs all of the current kernel/crash_core.c.

So only enabling PROC_CORE or FA_DUMP will enable CRASH_CORE, this mislead people that we enable crash dumping, actual it's not.

Secondly,
---
It's not reasonable to allow KEXEC_CORE select CRASH_CORE.

Because KEXEC_CORE enables codes which allocate control pages, copy kexec/kdump segments, and prepare for switching. These codes are shared by both kexec reboot and kdump. We could want kexec reboot, but disable kdump. In that case, CRASH_CORE should not be selected.

--------------------
CONFIG_CRASH_CORE=y
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC=y
CONFIG_KEXEC_FILE=y
---------------------

Thirdly,
---
It's not reasonable to allow CRASH_DUMP select KEXEC_CORE.

That could make KEXEC_CORE, CRASH_DUMP are enabled independently from KEXEC or KEXEC_FILE. However, w/o KEXEC or KEXEC_FILE, the KEXEC_CORE code built in doesn't make any sense because no kernel loading or switching will happen to utilize the KEXEC_CORE code.
---------------------
CONFIG_CRASH_CORE=y
CONFIG_KEXEC_CORE=y
CONFIG_CRASH_DUMP=y
---------------------

In this case, what is worse, on arch sh and arm, KEXEC relies on MMU, while CRASH_DUMP can still be enabled when !MMU, then compiling error is seen as the lkp test robot reported in above link.

------arch/sh/Kconfig------
config ARCH_SUPPORTS_KEXEC
def_bool MMU

config ARCH_SUPPORTS_CRASH_DUMP
def_bool BROKEN_ON_SMP
---------------------------

Changes:
===========
1, split out crash_reserve.c from crash_core.c;
2, split out vmcore_infoc. from crash_core.c;
3, move crash related codes in kexec_core.c into crash_core.c; 4, remove dependency of FA_DUMP on CRASH_DUMP;
5, clean up kdump related config items;
6, wrap up crash codes in crash related ifdefs on all 8 arch-es
which support crash dumping, except of ppc;

Achievement:
===========
With above changes, I can rearrange the config item logic as below (the right item depends on or is selected by the left item):

    PROC_KCORE -----------> VMCORE_INFO

               |----------> VMCORE_INFO
    FA_DUMP----|
               |----------> CRASH_RESERVE

                                                    ---->VMCORE_INFO
                                                   /
                                                   |---->CRASH_RESERVE
    KEXEC      --|                                /|
                 |--> KEXEC_CORE--> CRASH_DUMP-->/-|---->PROC_VMCORE
    KEXEC_FILE --|                               \ |
                                                   \---->CRASH_HOTPLUG


    KEXEC      --|
                 |--> KEXEC_CORE (for kexec reboot only)
    KEXEC_FILE --|

Test
========
On all 8 architectures, including x86_64, arm64, s390x, sh, arm, mips, riscv, loongarch, I did below three cases of config item setting and building all passed. Take configs on x86_64 as exampmle here:

(1) Both CONFIG_KEXEC and KEXEC_FILE is unset, then all kexec/kdump items are unset automatically:
# Kexec and crash features
# CONFIG_KEXEC is not set
# CONFIG_KEXEC_FILE is not set

# end of Kexec and crash features

(2) set CONFIG_KEXEC_FILE and 'make olddefconfig':
---------------
# Kexec and crash features CONFIG_CRASH_RESERVE=y CONFIG_VMCORE_INFO=y
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC_FILE=y
CONFIG_CRASH_DUMP=y
CONFIG_CRASH_HOTPLUG=y CONFIG_CRASH_MAX_MEMORY_RANGES=8192

# end of Kexec and crash features
---------------

(3) unset CONFIG_CRASH_DUMP in case 2 and execute 'make olddefconfig':
------------------------
# Kexec and crash features
CONFIG_KEXEC_CORE=y
CONFIG_KEXEC_FILE=y

Quoted 8 lines
# end of Kexec and crash features
------------------------

Note:
For ppc, it needs investigation to make clear how to split out crash code in arch folder. Hope Hari and Pingfan can help have a look, see if it's doable. Now, I make it either have both kexec and crash enabled, or disable both of them altogether.

I am seeing a few build failures in my test matrix on next-20240125 that appear to be caused by this series although I have not bisected. Some reproduction steps:

$ curl -LSso .config https://git.alpinelinux.org/aports/plain/community/linux-edge/config-edge.armv7
$ make -skj"$(nproc)" ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- olddefconfig all
...
arm-linux-gnueabi-ld: arch/arm/kernel/machine_kexec.o: in function `arch_crash_save_vmcoreinfo':
machine_kexec.c:(.text+0x488): undefined reference to `vmcoreinfo_append_str'
...
$ curl -LSso .config https://github.com/archlinuxarm/PKGBUILDs/raw/master/core/linux-aarch64/config
$ make -skj"$(nproc)" ARCH=arm64 CROSS_COMPILE=aarch64-linux- olddefconfig all
...
aarch64-linux-ld: kernel/kexec_file.o: in function `kexec_walk_memblock.constprop.0':
kexec_file.c:(.text+0x314): undefined reference to `crashk_res'
aarch64-linux-ld: kernel/kexec_file.o: relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol `crashk_res' which may bind externally can not be used when making a shared object; recompile with -fPIC
kexec_file.c:(.text+0x314): dangerous relocation: unsupported relocation
aarch64-linux-ld: kexec_file.c:(.text+0x318): undefined reference to `crashk_res'
aarch64-linux-ld: drivers/of/kexec.o: in function `of_kexec_alloc_and_setup_fdt':
kexec.c:(.text+0x580): undefined reference to `crashk_res'
aarch64-linux-ld: drivers/of/kexec.o: relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol `crashk_res' which may bind externally can not be used when making a shared object; recompile with -fPIC
kexec.c:(.text+0x580): dangerous relocation: unsupported relocation
aarch64-linux-ld: kexec.c:(.text+0x584): undefined reference to `crashk_res'
aarch64-linux-ld: kexec.c:(.text+0x590): undefined reference to `crashk_res'
aarch64-linux-ld: kexec.c:(.text+0x5b0): undefined reference to `crashk_low_res'
aarch64-linux-ld: drivers/of/kexec.o: relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol `crashk_low_res' which may bind externally can not be used when making a shared object; recompile with -fPIC
kexec.c:(.text+0x5b0): dangerous relocation: unsupported relocation
aarch64-linux-ld: kexec.c:(.text+0x5b4): undefined reference to `crashk_low_res'
aarch64-linux-ld: kexec.c:(.text+0x5c0): undefined reference to `crashk_low_res'
...
$ curl -LSso .config https://git.alpinelinux.org/aports/plain/community/linux-edge/config-edge.x86_64
$ make -skj"$(nproc)" ARCH=x86_64 CROSS_COMPILE=x86_64-linux- olddefconfig all
...
x86_64-linux-ld: arch/x86/xen/mmu_pv.o: in function `paddr_vmcoreinfo_note':
mmu_pv.c:(.text+0x3af3): undefined reference to `vmcoreinfo_note'
...

Cheers,
Nathan

author
#36 lore raw

On 01/25/24 at 09:55pm, Nathan Chancellor wrote:
......

Quoted 3 lines

I am seeing a few build failures in my test matrix on next-20240125 that appear to be caused by this series although I have not bisected. Some reproduction steps:

Thanks for trying this, I have reproduced the linking failure on arm, will work out a way to fix it.

It's weird, I remember I have built these and passed.

Quoted 39 lines
$ curl -LSso .config https://git.alpinelinux.org/aports/plain/community/linux-edge/config-edge.armv7
$ make -skj"$(nproc)" ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- olddefconfig all
...
arm-linux-gnueabi-ld: arch/arm/kernel/machine_kexec.o: in function `arch_crash_save_vmcoreinfo':
machine_kexec.c:(.text+0x488): undefined reference to `vmcoreinfo_append_str'
...
$ curl -LSso .config https://github.com/archlinuxarm/PKGBUILDs/raw/master/core/linux-aarch64/config
$ make -skj"$(nproc)" ARCH=arm64 CROSS_COMPILE=aarch64-linux- olddefconfig all
...
aarch64-linux-ld: kernel/kexec_file.o: in function `kexec_walk_memblock.constprop.0':
kexec_file.c:(.text+0x314): undefined reference to `crashk_res'
aarch64-linux-ld: kernel/kexec_file.o: relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol `crashk_res' which may bind externally can not be used when making a shared object; recompile with -fPIC
kexec_file.c:(.text+0x314): dangerous relocation: unsupported relocation
aarch64-linux-ld: kexec_file.c:(.text+0x318): undefined reference to `crashk_res'
aarch64-linux-ld: drivers/of/kexec.o: in function `of_kexec_alloc_and_setup_fdt':
kexec.c:(.text+0x580): undefined reference to `crashk_res'
aarch64-linux-ld: drivers/of/kexec.o: relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol `crashk_res' which may bind externally can not be used when making a shared object; recompile with -fPIC
kexec.c:(.text+0x580): dangerous relocation: unsupported relocation
aarch64-linux-ld: kexec.c:(.text+0x584): undefined reference to `crashk_res'
aarch64-linux-ld: kexec.c:(.text+0x590): undefined reference to `crashk_res'
aarch64-linux-ld: kexec.c:(.text+0x5b0): undefined reference to `crashk_low_res'
aarch64-linux-ld: drivers/of/kexec.o: relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol `crashk_low_res' which may bind externally can not be used when making a shared object; recompile with -fPIC
kexec.c:(.text+0x5b0): dangerous relocation: unsupported relocation
aarch64-linux-ld: kexec.c:(.text+0x5b4): undefined reference to `crashk_low_res'
aarch64-linux-ld: kexec.c:(.text+0x5c0): undefined reference to `crashk_low_res'
...
$ curl -LSso .config https://git.alpinelinux.org/aports/plain/community/linux-edge/config-edge.x86_64
$ make -skj"$(nproc)" ARCH=x86_64 CROSS_COMPILE=x86_64-linux- olddefconfig all
...
x86_64-linux-ld: arch/x86/xen/mmu_pv.o: in function `paddr_vmcoreinfo_note':
mmu_pv.c:(.text+0x3af3): undefined reference to `vmcoreinfo_note'
...

Cheers,
Nathan