mm: only mark/clear KHO scratch memory when needed
Subsystems (5)
v3 cover mm: only mark/clear KHO scratch memory when needed
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when CONFIG_KEXEC_HANDOVER is enabled and only if it is a KHO boot (both checked by is_kho_boot). Add check to prevent marking a KHO scratch region unless needed.
CONFIG_MEMBLOCK_KHO_SCRATCH is also removed as its only selected by KEXEC_HANDOVER.
v2 -> v3:
- is_kho_boot already has stubs when CONFIG_KEXEC_HANDOVER=n, no need to move the functions under CONFIG_KEXEC_HANDOVER (Pratyush Yadav)
- Commit messsage improvements
v1 -> v2:
- Remove CONFIG_MEMBLOCK_KHO_SCRATCH (Kiryl Shutsemau)
- Move memblock_mark/clear_kho_scratch under alread existing CONFIG_KEXEC_HANDOVER in memblock.c.
Usama Arif (2):
mm/memblock: remove CONFIG_MEMBLOCK_KHO_SCRATCH option mm/memblock: only mark/clear KHO scratch memory when needed
include/linux/memblock.h | 2 +- kernel/Kconfig.kexec | 1 - mm/Kconfig | 4 ---- mm/memblock.c | 18 +++++++++++------- 4 files changed, 12 insertions(+), 13 deletions(-)
Signature · 1 line
2.47.3
PATCH v3 1/2 mm/memblock: remove CONFIG_MEMBLOCK_KHO_SCRATCH option
The only cofig option that selects this is CONFIG_KEXEC_HANDOVER. Replace CONFIG_MEMBLOCK_KHO_SCRATCH with CONFIG_KEXEC_HANDOVER to simplify code.
No functional change intended.
--- include/linux/memblock.h | 2 +- kernel/Kconfig.kexec | 1 - mm/Kconfig | 4 ---- mm/memblock.c | 4 ++-- 4 files changed, 3 insertions(+), 8 deletions(-)
include/linux/memblock.h +1 -1
@@ -611,7 +611,7 @@ static inline void early_memtest(phys_addr_t start, phys_addr_t end) { }static inline void memtest_report_meminfo (struct seq_file *m) { }#endif -#ifdef CONFIG_MEMBLOCK_KHO_SCRATCH +#ifdef CONFIG_KEXEC_HANDOVER void memblock_set_kho_scratch_only (void );void memblock_clear_kho_scratch_only (void );void memmap_init_kho_scratch_pages (void );
kernel/Kconfig.kexec +0 -1
@@ -98,7 +98,6 @@ config KEXEC_HANDOVER bool "kexec handover" depends on ARCH_SUPPORTS_KEXEC_HANDOVER && ARCH_SUPPORTS_KEXEC_FILE depends on !DEFERRED_STRUCT_PAGE_INIT - select MEMBLOCK_KHO_SCRATCH select KEXEC_FILE select DEBUG_FS select LIBFDT
mm/Kconfig +0 -4
@@ -442,10 +442,6 @@ config HAVE_GUP_FAST depends on MMU bool -# Enable memblock support for scratch memory which is needed for kexec handover -config MEMBLOCK_KHO_SCRATCH - bool - # Don't discard allocated memory used to track "memory" and "reserved" memblocks # after early boot, so it can still be used to test for validity of memory. # Also, memblocks are updated with memory hot(un)plug.
mm/memblock.c +2 -2
@@ -112,7 +112,7 @@ unsigned long min_low_pfn;unsigned long max_pfn;unsigned long long max_possible_pfn; -#ifdef CONFIG_MEMBLOCK_KHO_SCRATCH +#ifdef CONFIG_KEXEC_HANDOVER /* When set to true, only allocate from MEMBLOCK_KHO_SCRATCH ranges */ static bool kho_scratch_only;#else
@@ -948,7 +948,7 @@ int __init_memblock memblock_physmem_add(phys_addr_t base, phys_addr_t size) }#endif -#ifdef CONFIG_MEMBLOCK_KHO_SCRATCH +#ifdef CONFIG_KEXEC_HANDOVER __init void memblock_set_kho_scratch_only (void ) { kho_scratch_only = true;
Signature · 1 line
2.47.3
PATCH v3 2/2 mm/memblock: only mark/clear KHO scratch memory when needed
In reply to Usama ArifThe scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when CONFIG_KEXEC_HANDOVER is enabled and only if it is a KHO boot (both checked by is_kho_boot). Add check to prevent marking a KHO scratch region unless needed.
--- mm/memblock.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)
mm/memblock.c +9 -5
@@ -20,8 +20,8 @@#ifdef CONFIG_KEXEC_HANDOVER #include <linux/libfdt.h> -#include <linux/kexec_handover.h> #endif /* CONFIG_KEXEC_HANDOVER */ +#include <linux/kexec_handover.h> #include <asm/sections.h> #include <linux/io.h>
@@ -1126,8 +1126,10 @@ int __init_memblock memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t*/ __init int memblock_mark_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,1 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ()) +return memblock_setclr_flag (&memblock.memory, base, size,1 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; }/**
@@ -1140,8 +1142,10 @@ __init int memblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size)*/ __init int memblock_clear_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,0 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ()) +return memblock_setclr_flag (&memblock.memory, base, size,0 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; } static bool should_skip_region(struct memblock_type *type,
Signature · 1 line
2.47.3
Mike Rapoport
maintainerRe: [PATCH v3 2/2] mm/memblock: only mark/clear KHO scratch memory when needed
Quoted 6 lines · Usama Arif
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when
CONFIG_KEXEC_HANDOVERis enabled and only if it is a KHO boot (both checked byis_kho_boot). Add check to prevent marking a KHO scratch region unless needed.
I'm going to rewrite the changelog and queue this for upstream:
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel and it is only needed when it is a KHO boot, i.e. a kexec boot with handover data passed from the previous kernel.
Currently x86 marks the first megabyte of memory as KHO scratch even for non-KHO boots if CONFIG_KEXEC_HANDOVER is enabled.
Add check to prevent marking a KHO scratch regions unless they are actually needed.
Quoted 51 lines
Fixes a2daf83e10378 ("x86/e820: temporarily enable KHO scratch for memory below 1M")Reported-by Vlad Poenaru <[email protected]>Signed-off-by Usama Arif <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>--- mm/memblock.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)mm/memblock.c +9
-5@@ -20,8 +20,8 @@#ifdef CONFIG_KEXEC_HANDOVER #include <linux/libfdt.h> -#include <linux/kexec_handover.h> #endif /* CONFIG_KEXEC_HANDOVER */ +#include <linux/kexec_handover.h> #include <asm/sections.h> #include <linux/io.h> @@ -1126,8 +1126,10 @@ int __init_memblock memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t*/ __init int memblock_mark_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,1 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ()) +return memblock_setclr_flag (&memblock.memory, base, size,1 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; }/** @@ -1140,8 +1142,10 @@ __init int memblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size)*/ __init int memblock_clear_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,0 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ()) +return memblock_setclr_flag (&memblock.memory, base, size,0 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; } static bool should_skip_region(struct memblock_type *type,Signature · 1 line
2.47.3
Signature · 2 lines
Sincerely yours, Mike.
Pasha Tatashin
maintainerRe: [PATCH v3 2/2] mm/memblock: only mark/clear KHO scratch memory when needed
Quoted 15 lines · Mike Rapoport
Quoted 6 lines · Usama Arif
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when
CONFIG_KEXEC_HANDOVERis enabled and only if it is a KHO boot (both checked byis_kho_boot). Add check to prevent marking a KHO scratch region unless needed.I'm going to rewrite the changelog and queue this for upstream:
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel and it is only needed when it is a KHO boot, i.e. a kexec boot with handover data passed from the previous kernel.
Quoted 10 lines
Currently x86 marks the first megabyte of memory as KHO scratch even for non-KHO boots if
CONFIG_KEXEC_HANDOVERis enabled.Add check to prevent marking a KHO scratch regions unless they are actually needed.
Quoted 4 lines
Fixes a2daf83e10378 ("x86/e820: temporarily enable KHO scratch for memory below 1M")Reported-by Vlad Poenaru <[email protected]>Signed-off-by Usama Arif <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>
This patch causes panic with my tests in linux-next.
[ 0.000000] Kernel panic - not syncing: Cannot allocate 17280 bytes for node 0 data [ 0.000000] CPU: 0 UID: 0 PID: 0 Comm: swapper Not tainted 6.18.0-next-20251203 #2 PREEMPT(undef) [ 0.000000] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.1 11/11/2019 [ 0.000000] Call Trace: [ 0.000000] <TASK> [ 0.000000] ? dump_stack_lvl+0x4e/0x70 [ 0.000000] ? vpanic+0xcf/0x2b0 [ 0.000000] ? panic+0x66/0x66 [ 0.000000] ? alloc_node_data+0x32/0x90 [ 0.000000] ? numa_register_nodes+0x82/0x100 [ 0.000000] ? numa_init+0x36/0x120 [ 0.000000] ? setup_arch+0x667/0x7f0 [ 0.000000] ? start_kernel+0x58/0x640 [ 0.000000] ? x86_64_start_reservations+0x24/0x30 [ 0.000000] ? x86_64_start_kernel+0xc5/0xd0 [ 0.000000] ? common_startup_64+0x13e/0x148 [ 0.000000] </TASK> [ 0.000000] ---[ end Kernel panic - not syncing: Cannot allocate
17280 bytes for node 0 data ]---
PANIC: early exception 0x0d IP 10:ffffffff89007a13 error 763 cr2 0xffff991090a01000
Quoted 16 lines
Quoted 16 lines
--- mm/memblock.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)mm/memblock.c +1
-1@@ -20,8 +20,8 @@#ifdef CONFIG_KEXEC_HANDOVER #include <linux/libfdt.h> -#include <linux/kexec_handover.h> #endif /* CONFIG_KEXEC_HANDOVER */ +#include <linux/kexec_handover.h>
Quoted 9 lines
mm/memblock.c int __init_memblock memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t · line 1129
Quoted 9 lines
mm/memblock.c +1
-2#include <asm/sections.h> #include <linux/io.h> @@ -1126,8 +1126,10 @@ int __init_memblock memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t*/ __init int memblock_mark_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,1 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ())
Looks like memblock_mark_kho_scratch() is called before is_kho_boot() is working:
[ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 0 [ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 0 [ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 0 [ 0.000000] KHO: kho_populate: is_kho_boot: 1 [ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 1 [ 0.000000] memblock_clear_kho_scratch: is_kho_boot: 1
...
kernel/liveupdate/kexec_handover.c +0 -0
@@ -1514,6 +1514,7 @@ void __init kho_populate(phys_addr_t fdt_phys,
u64 fdt_len,
memblock_set_kho_scratch_only();
kho_in.fdt_phys = fdt_phys;
+ pr_err("%s: is_kho_boot: %d\n", __func__, is_kho_boot());
kho_in.scratch_phys = scratch_phys;
kho_scratch_cnt = scratch_cnt;
pr_info("found kexec handover data.\n");mm/memblock.c +0 -0
@@ -1126,6 +1126,7 @@ int __init_memblock
memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t
*/
__init int memblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size) {
+ pr_err("%s: is_kho_boot: %d\n", __func__, is_kho_boot());
if (is_kho_boot())
return memblock_setclr_flag(&memblock.memory, base, size, 1,
MEMBLOCK_KHO_SCRATCH);@@ -1142,6 +1143,7 @@ __init int memblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size)
*/
__init int memblock_clear_kho_scratch(phys_addr_t base, phys_addr_t size) {
+ pr_err("%s: is_kho_boot: %d\n", __func__, is_kho_boot());
if (is_kho_boot())
return memblock_setclr_flag(&memblock.memory, base, size, 0,
MEMBLOCK_KHO_SCRATCH);Quoted 27 lines
Quoted 20 lines
mm/memblock.c +7
-3+return memblock_setclr_flag (&memblock.memory, base, size,1 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; }/** @@ -1140,8 +1142,10 @@ __init int memblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size)*/ __init int memblock_clear_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,0 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ()) +return memblock_setclr_flag (&memblock.memory, base, size,0 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; } static bool should_skip_region(struct memblock_type *type, --Quoted 2 lines
2.47.3
--
Sincerely yours,
Mike.
Re: [PATCH v3 2/2] mm/memblock: only mark/clear KHO scratch memory when needed
Quoted 50 lines · Pasha Tatashin
Quoted 15 lines · Mike Rapoport
Quoted 6 lines · Usama Arif
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when
CONFIG_KEXEC_HANDOVERis enabled and only if it is a KHO boot (both checked byis_kho_boot). Add check to prevent marking a KHO scratch region unless needed.I'm going to rewrite the changelog and queue this for upstream:
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel and it is only needed when it is a KHO boot, i.e. a kexec boot with handover data passed from the previous kernel.
Quoted 10 lines
Currently x86 marks the first megabyte of memory as KHO scratch even for non-KHO boots if
CONFIG_KEXEC_HANDOVERis enabled.Add check to prevent marking a KHO scratch regions unless they are actually needed.
Quoted 4 lines
Fixes a2daf83e10378 ("x86/e820: temporarily enable KHO scratch for memory below 1M")Reported-by Vlad Poenaru <[email protected]>Signed-off-by Usama Arif <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>This patch causes panic with my tests in linux-next.
[ 0.000000] Kernel panic - not syncing: Cannot allocate 17280 bytes for node 0 data [ 0.000000] CPU: 0 UID: 0 PID: 0 Comm: swapper Not tainted 6.18.0-next-20251203 #2 PREEMPT(undef) [ 0.000000] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.1 11/11/2019 [ 0.000000] Call Trace: [ 0.000000] <TASK> [ 0.000000] ? dump_stack_lvl+0x4e/0x70 [ 0.000000] ? vpanic+0xcf/0x2b0 [ 0.000000] ? panic+0x66/0x66 [ 0.000000] ? alloc_node_data+0x32/0x90 [ 0.000000] ? numa_register_nodes+0x82/0x100 [ 0.000000] ? numa_init+0x36/0x120 [ 0.000000] ? setup_arch+0x667/0x7f0 [ 0.000000] ? start_kernel+0x58/0x640 [ 0.000000] ? x86_64_start_reservations+0x24/0x30 [ 0.000000] ? x86_64_start_kernel+0xc5/0xd0 [ 0.000000] ? common_startup_64+0x13e/0x148 [ 0.000000] </TASK> [ 0.000000] ---[ end Kernel panic - not syncing: Cannot allocate
Quoted 4 lines
17280 bytes for node 0 data ]---
PANIC: early exception0x0dIP 10:ffffffff89007a13 error 763 cr20xffff991090a01000
Thanks for reporting this and sorry for the bug!
So the patch was designed to remove the memblock_mark_kho_scratch in e820__memblock_setup if not in KHO boot. But it broke memblock_mark_kho_scratch in kho_populate.
Moving kho_in.fdt_phys = fdt_phys to before the memblock_mark_scratch
should fix it. I dont have a setup where I can easily test KHO, but I think below should fix it?
TBH using fdt_phys to check if the boot is KHO might be a bit hacky? Is it possible to have a better check for this?
kernel/liveupdate/kexec_handover.c +4 -1
@@ -1483,6 +1483,7 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len, goto out; } + kho_in.fdt_phys = fdt_phys;/* * We pass a safe contiguous blocks of memory to use for early boot * purporses from the previous kernel so that we can resize the
@@ -1513,7 +1514,6 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,*/ memblock_set_kho_scratch_only (); - kho_in.fdt_phys = fdt_phys; kho_in.scratch_phys = scratch_phys; kho_scratch_cnt = scratch_cnt;pr_info ("found kexec handover data.\n" );
@@ -1524,7 +1524,10 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,if (scratch)early_memunmap (scratch, scratch_len);if (err) + { + kho_in.fdt_phys =0 ;pr_warn ("disabling KHO revival: %d\n" , err); + } } >
Quoted 98 lines
Quoted 16 lines
Quoted 16 lines
--- mm/memblock.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)mm/memblock.c +1
-1@@ -20,8 +20,8 @@#ifdef CONFIG_KEXEC_HANDOVER #include <linux/libfdt.h> -#include <linux/kexec_handover.h> #endif /* CONFIG_KEXEC_HANDOVER */ +#include <linux/kexec_handover.h> Quoted 9 lines
mm/memblock.c int __init_memblock memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t · line 1129
Quoted 9 lines
mm/memblock.c +1
-2#include <asm/sections.h> #include <linux/io.h> @@ -1126,8 +1126,10 @@ int __init_memblock memblock_reserved_mark_noinit(phys_addr_t base, phys_addr_t*/ __init int memblock_mark_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,1 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ())Looks like
memblock_mark_kho_scratch()is called beforeis_kho_boot()is working:[ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 0 [ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 0 [ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 0 [ 0.000000] KHO: kho_populate: is_kho_boot: 1 [ 0.000000] memblock_mark_kho_scratch: is_kho_boot: 1 [ 0.000000] memblock_clear_kho_scratch: is_kho_boot: 1...
kernel/liveupdate/kexec_handover.c +1
-0@@ -1514,6 +1514,7 @@ void __init kho_populate(phys_addr_t fdt_phys,u64 fdt_len,memblock_set_kho_scratch_only (); kho_in.fdt_phys = fdt_phys; +pr_err ("%s: is_kho_boot: %d\n" , __func__,is_kho_boot ()); kho_in.scratch_phys = scratch_phys; kho_scratch_cnt = scratch_cnt;pr_info("found kexec handover data.\n");mm/memblock.c +0
-0@@ -1126,6 +1126,7 @@ int __init_memblockmemblock_reserved_mark_noinit(phys_addr_t base,
phys_addr_t
*/
__initintmemblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size){mm/memblock.c +2
-0+pr_err ("%s: is_kho_boot: %d\n" , __func__,is_kho_boot ());if (is_kho_boot ())return memblock_setclr_flag (&memblock.memory, base, size,1 ,MEMBLOCK_KHO_SCRATCH );@@ -1142,6 +1143,7 @@ __init int memblock_mark_kho_scratch(phys_addr_tbase, phys_addr_t size) */ __init int memblock_clear_kho_scratch (phys_addr_t base,phys_addr_t size) { +pr_err ("%s: is_kho_boot: %d\n" , __func__,is_kho_boot ());if (is_kho_boot ()) return memblock_setclr_flag(&memblock.memory, base, size, 0,MEMBLOCK_KHO_SCRATCH);Quoted 27 lines
Quoted 20 lines
mm/memblock.c +7
-3+return memblock_setclr_flag (&memblock.memory, base, size,1 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; }/** @@ -1140,8 +1142,10 @@ __init int memblock_mark_kho_scratch(phys_addr_t base, phys_addr_t size)*/ __init int memblock_clear_kho_scratch (phys_addr_t base,phys_addr_t size) { -return memblock_setclr_flag (&memblock.memory, base, size,0 , -MEMBLOCK_KHO_SCRATCH ); +if (is_kho_boot ()) +return memblock_setclr_flag (&memblock.memory, base, size,0 , +MEMBLOCK_KHO_SCRATCH ); +return 0 ; } static bool should_skip_region(struct memblock_type *type, --Quoted 2 lines
2.47.3
--
Sincerely yours,
Mike.
Mike Rapoport
maintainerRe: [PATCH v3 2/2] mm/memblock: only mark/clear KHO scratch memory when needed
Hi Usama,
Quoted 55 lines · Usama Arif
Quoted 50 lines · Mike Rapoport
Quoted 15 lines · Usama Arif
Quoted 6 lines
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when
CONFIG_KEXEC_HANDOVERis enabled and only if it is a KHO boot (both checked byis_kho_boot). Add check to prevent marking a KHO scratch region unless needed.I'm going to rewrite the changelog and queue this for upstream:
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel and it is only needed when it is a KHO boot, i.e. a kexec boot with handover data passed from the previous kernel.
Quoted 10 lines
Currently x86 marks the first megabyte of memory as KHO scratch even for non-KHO boots if
CONFIG_KEXEC_HANDOVERis enabled.Add check to prevent marking a KHO scratch regions unless they are actually needed.
Quoted 4 lines
Fixes a2daf83e10378 ("x86/e820: temporarily enable KHO scratch for memory below 1M")Reported-by Vlad Poenaru <[email protected]>Signed-off-by Usama Arif <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>This patch causes panic with my tests in linux-next.
[ 0.000000] Kernel panic - not syncing: Cannot allocate 17280 bytes for node 0 data [ 0.000000] CPU: 0 UID: 0 PID: 0 Comm: swapper Not tainted 6.18.0-next-20251203 #2 PREEMPT(undef) [ 0.000000] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.1 11/11/2019 [ 0.000000] Call Trace: [ 0.000000] <TASK> [ 0.000000] ? dump_stack_lvl+0x4e/0x70 [ 0.000000] ? vpanic+0xcf/0x2b0 [ 0.000000] ? panic+0x66/0x66 [ 0.000000] ? alloc_node_data+0x32/0x90 [ 0.000000] ? numa_register_nodes+0x82/0x100 [ 0.000000] ? numa_init+0x36/0x120 [ 0.000000] ? setup_arch+0x667/0x7f0 [ 0.000000] ? start_kernel+0x58/0x640 [ 0.000000] ? x86_64_start_reservations+0x24/0x30 [ 0.000000] ? x86_64_start_kernel+0xc5/0xd0 [ 0.000000] ? common_startup_64+0x13e/0x148 [ 0.000000] </TASK> [ 0.000000] ---[ end Kernel panic - not syncing: Cannot allocateQuoted 4 lines
17280 bytes for node 0 data ]---
PANIC: early exception0x0dIP 10:ffffffff89007a13 error 763 cr20xffff991090a01000
Quoted 7 lines
Thanks for reporting this and sorry for the bug!
So the patch was designed to remove the
memblock_mark_kho_scratchine820__memblock_setupif not in KHO boot. But it brokememblock_mark_kho_scratchinkho_populate.
Moving kho_in.fdt_phys =fdt_physto before thememblock_mark_scratch
should fix it. I dont have a setup where I can easily test KHO, but I think below should fix it?
This might, but this is too late for v6.19-rc1. For now I'm dropping this series from memblock/for-next. We can resume working on this after merge window closes.
Quoted 2 lines
TBH using
fdt_physto check if the boot is KHO might be a bit hacky? Is it possible to have a better check for this?
Presence of KHO FDT is a clear indication that it is a KHO boot. The issue is that during early boot ordering is hard and it's not always clear in which order features and configuration are detected and used.
Quoted 29 lines
kernel/liveupdate/kexec_handover.c +4
-1@@ -1483,6 +1483,7 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len, goto out; } + kho_in.fdt_phys = fdt_phys;/* * We pass a safe contiguous blocks of memory to use for early boot * purporses from the previous kernel so that we can resize the @@ -1513,7 +1514,6 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,*/ memblock_set_kho_scratch_only (); - kho_in.fdt_phys = fdt_phys; kho_in.scratch_phys = scratch_phys; kho_scratch_cnt = scratch_cnt;pr_info ("found kexec handover data.\n" );@@ -1524,7 +1524,10 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,if (scratch)early_memunmap (scratch, scratch_len);if (err) + { + kho_in.fdt_phys =0 ;pr_warn ("disabling KHO revival: %d\n" , err); + } }
Signature · 2 lines
Sincerely yours, Mike.
Re: [PATCH v3 2/2] mm/memblock: only mark/clear KHO scratch memory when needed
Quoted 66 lines · Mike Rapoport
Hi Usama,
Quoted 55 lines · Usama Arif
Quoted 50 lines · Mike Rapoport
Quoted 15 lines · Usama Arif
Quoted 6 lines
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel. Only the 1st 1MB is used as scratch, and its a hack to get around limitations with KHO. It is only needed when
CONFIG_KEXEC_HANDOVERis enabled and only if it is a KHO boot (both checked byis_kho_boot). Add check to prevent marking a KHO scratch region unless needed.I'm going to rewrite the changelog and queue this for upstream:
The scratch memory for kexec handover is used to bootstrap the kexec'ed kernel and it is only needed when it is a KHO boot, i.e. a kexec boot with handover data passed from the previous kernel.
Quoted 10 lines
Currently x86 marks the first megabyte of memory as KHO scratch even for non-KHO boots if
CONFIG_KEXEC_HANDOVERis enabled.Add check to prevent marking a KHO scratch regions unless they are actually needed.
Quoted 4 lines
Fixes a2daf83e10378 ("x86/e820: temporarily enable KHO scratch for memory below 1M")Reported-by Vlad Poenaru <[email protected]>Signed-off-by Usama Arif <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>This patch causes panic with my tests in linux-next.
[ 0.000000] Kernel panic - not syncing: Cannot allocate 17280 bytes for node 0 data [ 0.000000] CPU: 0 UID: 0 PID: 0 Comm: swapper Not tainted 6.18.0-next-20251203 #2 PREEMPT(undef) [ 0.000000] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.1 11/11/2019 [ 0.000000] Call Trace: [ 0.000000] <TASK> [ 0.000000] ? dump_stack_lvl+0x4e/0x70 [ 0.000000] ? vpanic+0xcf/0x2b0 [ 0.000000] ? panic+0x66/0x66 [ 0.000000] ? alloc_node_data+0x32/0x90 [ 0.000000] ? numa_register_nodes+0x82/0x100 [ 0.000000] ? numa_init+0x36/0x120 [ 0.000000] ? setup_arch+0x667/0x7f0 [ 0.000000] ? start_kernel+0x58/0x640 [ 0.000000] ? x86_64_start_reservations+0x24/0x30 [ 0.000000] ? x86_64_start_kernel+0xc5/0xd0 [ 0.000000] ? common_startup_64+0x13e/0x148 [ 0.000000] </TASK> [ 0.000000] ---[ end Kernel panic - not syncing: Cannot allocateQuoted 4 lines
17280 bytes for node 0 data ]---
PANIC: early exception0x0dIP 10:ffffffff89007a13 error 763 cr20xffff991090a01000Quoted 7 lines
Thanks for reporting this and sorry for the bug!
So the patch was designed to remove the
memblock_mark_kho_scratchine820__memblock_setupif not in KHO boot. But it brokememblock_mark_kho_scratchinkho_populate.
Moving kho_in.fdt_phys =fdt_physto before thememblock_mark_scratch
should fix it. I dont have a setup where I can easily test KHO, but I think below should fix it?
Quoted 4 lines
This might, but this is too late for v6.19-rc1. For now I'm dropping this series from memblock/for-next. We can resume working on this after merge window closes.
Yes makes sense.
How would you like me to proceed with the fix? Should I send just the fix now, or these 2 patches plus the fix after the merge window closes?
Thanks!
Quoted 7 lines
Quoted 2 lines
TBH using
fdt_physto check if the boot is KHO might be a bit hacky? Is it possible to have a better check for this?Presence of KHO FDT is a clear indication that it is a KHO boot. The issue is that during early boot ordering is hard and it's not always clear in which order features and configuration are detected and used.
ack
Quoted 30 lines
Quoted 29 lines
kernel/liveupdate/kexec_handover.c +4
-1@@ -1483,6 +1483,7 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len, goto out; } + kho_in.fdt_phys = fdt_phys;/* * We pass a safe contiguous blocks of memory to use for early boot * purporses from the previous kernel so that we can resize the @@ -1513,7 +1514,6 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,*/ memblock_set_kho_scratch_only (); - kho_in.fdt_phys = fdt_phys; kho_in.scratch_phys = scratch_phys; kho_scratch_cnt = scratch_cnt;pr_info ("found kexec handover data.\n" );@@ -1524,7 +1524,10 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,if (scratch)early_memunmap (scratch, scratch_len);if (err) + { + kho_in.fdt_phys =0 ;pr_warn ("disabling KHO revival: %d\n" , err); + } }
Mike Rapoport
maintainerRe: [PATCH v3 2/2] mm/memblock: only mark/clear KHO scratch memory when needed
Quoted 7 lines · Usama Arif
Quoted 6 lines · Mike Rapoport
Quoted 6 lines
So the patch was designed to remove the
memblock_mark_kho_scratchine820__memblock_setupif not in KHO boot. But it brokememblock_mark_kho_scratchinkho_populate.
Moving kho_in.fdt_phys =fdt_physto before thememblock_mark_scratch
should fix it. I dont have a setup where I can easily test KHO, but I think below should fix it?
There's a simple KHO sefltest in tools/testing/selftest/kho
Quoted 9 lines
Quoted 4 lines
This might, but this is too late for v6.19-rc1. For now I'm dropping this series from memblock/for-next. We can resume working on this after merge window closes.
Yes makes sense.
How would you like me to proceed with the fix? Should I send just the fix now, or these 2 patches plus the fix after the merge window closes?
The fix should come before the changes in memblock_mark_kho_scratch(), so please resend the whole series.
Quoted 9 lines
Quoted 9 lines
kernel/liveupdate/kexec_handover.c void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len, · line 1486
Quoted 9 lines
kernel/liveupdate/kexec_handover.c +1
-0@@ -1483,6 +1483,7 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len, goto out; } + kho_in.fdt_phys = fdt_phys;
This should happen before the calls to memblock_mark_kho_scratch().
Quoted 22 lines
Quoted 21 lines
Quoted 20 lines
kernel/liveupdate/kexec_handover.c +3
-1/* * We pass a safe contiguous blocks of memory to use for early boot * purporses from the previous kernel so that we can resize the @@ -1513,7 +1514,6 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,*/ memblock_set_kho_scratch_only (); - kho_in.fdt_phys = fdt_phys; kho_in.scratch_phys = scratch_phys; kho_scratch_cnt = scratch_cnt;pr_info ("found kexec handover data.\n" );@@ -1524,7 +1524,10 @@ void __init kho_populate(phys_addr_t fdt_phys, u64 fdt_len,if (scratch)early_memunmap (scratch, scratch_len);if (err) + { + kho_in.fdt_phys =0 ;pr_warn ("disabling KHO revival: %d\n" , err); + } }
Signature · 2 lines
Sincerely yours, Mike.