mm, arch: add generic implementation of pfn_valid() for FLATMEM
Subsystems (22)
v2 cover mm, arch: add generic implementation of pfn_valid() for FLATMEM
From: "Mike Rapoport (IBM)" <[email protected]>
Hi,
Every architecture that supports FLATMEM memory model defines its own version of pfn_valid() that essentially compares a pfn to max_mapnr.
Use mips/powerpc version implemented as static inline as a generic implementation of pfn_valid() and drop its per-architecture definitions
v2:
* fix build on ARM and xtensa
* add Acked- and Reviewed-by, thanks everybody
v1: https://lore.kernel.org/all/[email protected]
Mike Rapoport (IBM) (4):
arm: include asm-generic/memory_model.h from page.h rather than
memory.h
m68k: use asm-generic/memory_model.h for both MMU and !MMU
mips: drop definition of pfn_valid() for DISCONTIGMEM mm, arch: add generic implementation of pfn_valid() for FLATMEM
arch/alpha/include/asm/page.h | 4 ---- arch/arc/include/asm/page.h | 1 - arch/arm/include/asm/memory.h | 2 -- arch/arm/include/asm/page.h | 2 ++ arch/csky/include/asm/page.h | 1 - arch/hexagon/include/asm/page.h | 1 - arch/ia64/include/asm/page.h | 4 ---- arch/loongarch/include/asm/page.h | 13 ------------- arch/m68k/include/asm/page.h | 6 +----- arch/m68k/include/asm/page_mm.h | 1 - arch/m68k/include/asm/page_no.h | 4 ---- arch/microblaze/include/asm/page.h | 1 - arch/mips/include/asm/page.h | 28 ---------------------------- arch/nios2/include/asm/page.h | 9 --------- arch/openrisc/include/asm/page.h | 2 -- arch/parisc/include/asm/page.h | 4 ---- arch/powerpc/include/asm/page.h | 9 --------- arch/riscv/include/asm/page.h | 5 ----- arch/sh/include/asm/page.h | 3 --- arch/sparc/include/asm/page_32.h | 1 - arch/um/include/asm/page.h | 1 - arch/x86/include/asm/page_32.h | 4 ---- arch/x86/include/asm/page_64.h | 4 ---- arch/xtensa/include/asm/page.h | 4 ++-- include/asm-generic/memory_model.h | 12 ++++++++++++ include/asm-generic/page.h | 2 -- 26 files changed, 17 insertions(+), 111 deletions(-)
base-commit: 2241ab53cbb5cdb08a6b2d4688feb13971058f65
Signature · 1 line
2.35.1
PATCH v2 1/4 arm: include asm-generic/memory_model.h from page.h rather than memory.h
From: "Mike Rapoport (IBM)" <[email protected]>
Makes it consistent with other architectures and allows for generic definition of pfn_valid() in asm-generic/memory_model.h with clear override in arch/arm/include/asm/page.h
--- arch/arm/include/asm/memory.h | 2 -- arch/arm/include/asm/page.h | 2 ++ 2 files changed, 2 insertions(+), 2 deletions(-)
arch/arm/include/asm/memory.h +0 -2
@@ -386,6 +386,4 @@ static inline unsigned long __virt_to_idmap(unsigned long x)#endif -#include <asm-generic/memory_model.h> -#endif
arch/arm/include/asm/page.h +2 -0
@@ -158,6 +158,7 @@ typedef struct page *pgtable_t;#ifdef CONFIG_HAVE_ARCH_PFN_VALID extern int pfn_valid (unsigned long ); +#define pfn_valid pfn_valid#endif #include <asm/memory.h>
@@ -167,5 +168,6 @@ extern int pfn_valid(unsigned long);#define VM_DATA_DEFAULT_FLAGS VM_DATA_FLAGS_TSK_EXEC#include <asm-generic/getorder.h> +#include <asm-generic/memory_model.h> #endif
Signature · 1 line
2.35.1
Re: [PATCH v2 1/4] arm: include asm-generic/memory_model.h from page.h rather than memory.h
Quoted 8 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Makes it consistent with other architectures and allows for generic definition of
pfn_valid()inasm-generic/memory_model.hwith clear override inarch/arm/include/asm/page.hSigned-off-by Mike Rapoport (IBM) <[email protected]>---
Signature · 3 lines
Thanks, David / dhildenb
PATCH v2 2/4 m68k: use asm-generic/memory_model.h for both MMU and !MMU
In reply to Mike RapoportFrom: "Mike Rapoport (IBM)" <[email protected]>
The MMU variant uses generic definitions of page_to_pfn() and pfn_to_page(), but !MMU defines them in include/asm/page_no.h for no good reason.
Include asm-generic/memory_model.h in the common include/asm/page.h and drop redundant definitions.
--- arch/m68k/include/asm/page.h | 6 +----- arch/m68k/include/asm/page_mm.h | 1 - arch/m68k/include/asm/page_no.h | 2 -- 3 files changed, 1 insertion(+), 8 deletions(-)
arch/m68k/include/asm/page.h +1 -5
@@ -62,11 +62,7 @@ extern unsigned long _ramend;#include <asm/page_no.h> #endif -#ifndef CONFIG_MMU -#define __phys_to_pfn (paddr) ((unsigned long)((paddr) >> PAGE_SHIFT)) -#define __pfn_to_phys (pfn) PFN_PHYS(pfn) -#endif -#include <asm-generic/getorder.h> +#include <asm-generic/memory_model.h> #endif /* _M68K_PAGE_H */
arch/m68k/include/asm/page_mm.h +0 -1
@@ -134,7 +134,6 @@ extern int m68k_virt_to_node_shift; })#define ARCH_PFN_OFFSET (m68k_memory[0].addr >> PAGE_SHIFT) -#include <asm-generic/memory_model.h> #define virt_addr_valid (kaddr) ((unsigned long)(kaddr) >= PAGE_OFFSET && (unsigned long)(kaddr) < (unsigned long)high_memory)#define pfn_valid (pfn) virt_addr_valid(pfn_to_virt(pfn))
arch/m68k/include/asm/page_no.h +0 -2
@@ -26,8 +26,6 @@ extern unsigned long memory_end;#define virt_to_page (addr) (mem_map + (((unsigned long)(addr)-PAGE_OFFSET) >> PAGE_SHIFT))#define page_to_virt (page) __va(((((page) - mem_map) << PAGE_SHIFT) + PAGE_OFFSET)) -#define pfn_to_page (pfn) virt_to_page(pfn_to_virt(pfn)) -#define page_to_pfn (page) virt_to_pfn(page_to_virt(page))#define pfn_valid (pfn) ((pfn) < max_mapnr)#define virt_addr_valid (kaddr) (((unsigned long)(kaddr) >= PAGE_OFFSET) && \
Signature · 1 line
2.35.1
Re: [PATCH v2 2/4] m68k: use asm-generic/memory_model.h for both MMU and !MMU
Quoted 6 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
The MMU variant uses generic definitions of
page_to_pfn()andpfn_to_page(), but !MMU defines them ininclude/asm/page_no.hfor no good reason.
Quoted 7 lines
Include
asm-generic/memory_model.hin the commoninclude/asm/page.hand drop redundant definitions.Signed-off-by Mike Rapoport (IBM) <[email protected]>Reviewed-by Geert Uytterhoeven <[email protected]>---
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH v2 2/4] m68k: use asm-generic/memory_model.h for both MMU and !MMU
In reply to Mike RapoportHi,
Quoted 6 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
The MMU variant uses generic definitions of
page_to_pfn()andpfn_to_page(), but !MMU defines them ininclude/asm/page_no.hfor no good reason.
Quoted 6 lines
Include
asm-generic/memory_model.hin the commoninclude/asm/page.hand drop redundant definitions.Signed-off-by Mike Rapoport (IBM) <[email protected]>Reviewed-by Geert Uytterhoeven <[email protected]>
This patch results in a boot failure when trying to boot the mcf5208evb qemu emulation. Reverting it together with "mm, arch: add generic implementation of pfn_valid() for FLATMEM" fixes the problem. There is no error log - the emulation hangs silently until aborted.
Guenter
---
bisect log:
# bad: [6ba8a227fd19d19779005fb66ad7562608e1df83] Add linux-next specific files for 20230210
# good: [4ec5183ec48656cec489c49f989c508b68b518e3] Linux 6.2-rc7
git bisect start 'HEAD' 'v6.2-rc7'
# good: [94613f0efc69ed41f9229ef5c294db3ec37145da] Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good 94613f0efc69ed41f9229ef5c294db3ec37145da
# good: [19e62c715fe70dae4582c2874ed3e66715d09af6] Merge branch 'rcu/next' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git
git bisect good 19e62c715fe70dae4582c2874ed3e66715d09af6
# good: [5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6] Merge branch 'for-next' of https://gitlab.com/peda-linux/mux.git
git bisect good 5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6
# good: [c349bf6ec83903b20fe570c5609b9a864a64e09c] Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode.git
git bisect good c349bf6ec83903b20fe570c5609b9a864a64e09c
# good: [5a06a9f17454df38f35672be522ff5eb9b4277d2] selftest: add testing unsharing and counting ksm zero page
git bisect good 5a06a9f17454df38f35672be522ff5eb9b4277d2
# bad: [f5d115a7b06e5661ed5218ffa9a2644c4ff1c135] Merge branch 'mm-nonmm-unstable' into mm-everything
git bisect bad f5d115a7b06e5661ed5218ffa9a2644c4ff1c135
# bad: [acb018d6ea0c055381fba7dddaef386ee28f8075] mm/vmalloc.c: allow vread() to read out vm_map_ram areas
git bisect bad acb018d6ea0c055381fba7dddaef386ee28f8075
# good: [1a5d9782ac969dc6e61c6786500b5160603188ea] mm/mmap: remove __vma_adjust()
git bisect good 1a5d9782ac969dc6e61c6786500b5160603188ea
# good: [4b32363697de957dcc890b6245bec3f58903639a] arm: include asm-generic/memory_model.h from page.h rather than memory.h git bisect good 4b32363697de957dcc890b6245bec3f58903639a
# bad: [328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8] mm/migrate: convert isolate_movable_page() to use folios
git bisect bad 328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8
# bad: [b704c765b08cabe82adf76a4d1a74f3688eee410] mm/mempolicy: convert queue_pages_pmd() to queue_folios_pmd()
git bisect bad b704c765b08cabe82adf76a4d1a74f3688eee410
# bad: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation of pfn_valid() for FLATMEM
git bisect bad e5734c8b0edfd2a053a5c256189586a3b1e9f63d
# bad: [ad8aecea034c591b9754bc5908da9719853aa7fa] mips: drop definition of pfn_valid() for DISCONTIGMEM
git bisect bad ad8aecea034c591b9754bc5908da9719853aa7fa
# bad: [1f6271a0dfdf952c2e3981f424784d48f243a2be] m68k: use asm-generic/memory_model.h for both MMU and !MMU
git bisect bad 1f6271a0dfdf952c2e3981f424784d48f243a2be
# first bad commit: [1f6271a0dfdf952c2e3981f424784d48f243a2be] m68k: use asm-generic/memory_model.h for both MMU and !MMU
Re: [PATCH v2 2/4] m68k: use asm-generic/memory_model.h for both MMU and !MMU
Hi,
Quoted 12 lines · Guenter Roeck
Hi,
Quoted 9 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
The MMU variant uses generic definitions of
page_to_pfn()andpfn_to_page(), but !MMU defines them ininclude/asm/page_no.hfor no good reason.Include
asm-generic/memory_model.hin the commoninclude/asm/page.hand drop redundant definitions.
Quoted 8 lines
Quoted 3 lines
Signed-off-by Mike Rapoport (IBM) <[email protected]>Reviewed-by Geert Uytterhoeven <[email protected]>This patch results in a boot failure when trying to boot the mcf5208evb qemu emulation. Reverting it together with "mm, arch: add generic implementation of
pfn_valid()for FLATMEM" fixes the problem. There is no error log - the emulation hangs silently until aborted.
With the patch below I was able to get to the mount of the root file system, but I don't have one, so I couldn't test the boot properly.
arch/m68k/include/asm/page_no.h +2 -0
@@ -28,6 +28,8 @@ extern unsigned long memory_end;#define virt_addr_valid (kaddr) (((unsigned long)(kaddr) >= PAGE_OFFSET) && \ ((unsigned long)(kaddr) < memory_end)) +#define ARCH_PFN_OFFSET PHYS_PFN(PAGE_OFFSET_RAW) +#endif /* __ASSEMBLY__ */ #endif /* _M68K_PAGE_NO_H */
Quoted 37 lines
Guenter
---
bisect log:# bad: [
6ba8a227fd19d19779005fb66ad7562608e1df83] Add linux-next specific files for 20230210
# good: [4ec5183ec48656cec489c49f989c508b68b518e3] Linux 6.2-rc7
git bisect start 'HEAD' 'v6.2-rc7'
# good: [94613f0efc69ed41f9229ef5c294db3ec37145da] Merge branch 'master' ofgit://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good94613f0efc69ed41f9229ef5c294db3ec37145da
# good: [19e62c715fe70dae4582c2874ed3e66715d09af6] Merge branch 'rcu/next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git
git bisect good19e62c715fe70dae4582c2874ed3e66715d09af6
# good: [5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6] Merge branch 'for-next' of https://gitlab.com/peda-linux/mux.git
git bisect good5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6
# good: [c349bf6ec83903b20fe570c5609b9a864a64e09c] Merge branch 'for-next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode.git
git bisect goodc349bf6ec83903b20fe570c5609b9a864a64e09c
# good: [5a06a9f17454df38f35672be522ff5eb9b4277d2] selftest: add testing unsharing and counting ksm zero page
git bisect good5a06a9f17454df38f35672be522ff5eb9b4277d2
# bad: [f5d115a7b06e5661ed5218ffa9a2644c4ff1c135] Merge branch 'mm-nonmm-unstable' into mm-everything
git bisect badf5d115a7b06e5661ed5218ffa9a2644c4ff1c135
# bad: [acb018d6ea0c055381fba7dddaef386ee28f8075]mm/vmalloc.c: allowvread()to read outvm_map_ramareas
git bisect badacb018d6ea0c055381fba7dddaef386ee28f8075
# good: [1a5d9782ac969dc6e61c6786500b5160603188ea] mm/mmap: remove__vma_adjust()
git bisect good1a5d9782ac969dc6e61c6786500b5160603188ea
# good: [4b32363697de957dcc890b6245bec3f58903639a] arm: includeasm-generic/memory_model.hfrompage.hrather thanmemory.hgit bisect good4b32363697de957dcc890b6245bec3f58903639a
# bad: [328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8] mm/migrate: convertisolate_movable_page()to use folios
git bisect bad328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8
# bad: [b704c765b08cabe82adf76a4d1a74f3688eee410] mm/mempolicy: convertqueue_pages_pmd()toqueue_folios_pmd()
git bisect badb704c765b08cabe82adf76a4d1a74f3688eee410
# bad: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation ofpfn_valid()for FLATMEM
git bisect bade5734c8b0edfd2a053a5c256189586a3b1e9f63d
# bad: [ad8aecea034c591b9754bc5908da9719853aa7fa] mips: drop definition ofpfn_valid()for DISCONTIGMEM
git bisect badad8aecea034c591b9754bc5908da9719853aa7fa
# bad: [1f6271a0dfdf952c2e3981f424784d48f243a2be] m68k: useasm-generic/memory_model.hfor both MMU and !MMU
git bisect bad1f6271a0dfdf952c2e3981f424784d48f243a2be# first bad commit: [1f6271a0dfdf952c2e3981f424784d48f243a2be] m68k: use asm-generic/memory_model.h for both MMU and !MMU
Signature · 2 lines
Sincerely yours, Mike.
Re: [PATCH v2 2/4] m68k: use asm-generic/memory_model.h for both MMU and !MMU
Quoted 33 lines · Mike Rapoport
Hi,
Quoted 12 lines · Guenter Roeck
Hi,
Quoted 9 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
The MMU variant uses generic definitions of
page_to_pfn()andpfn_to_page(), but !MMU defines them ininclude/asm/page_no.hfor no good reason.Include
asm-generic/memory_model.hin the commoninclude/asm/page.hand drop redundant definitions.Quoted 8 lines
Quoted 3 lines
Signed-off-by Mike Rapoport (IBM) <[email protected]>Reviewed-by Geert Uytterhoeven <[email protected]>This patch results in a boot failure when trying to boot the mcf5208evb qemu emulation. Reverting it together with "mm, arch: add generic implementation of
pfn_valid()for FLATMEM" fixes the problem. There is no error log - the emulation hangs silently until aborted.With the patch below I was able to get to the mount of the root file system, but I don't have one, so I couldn't test the boot properly.
arch/m68k/include/asm/page_no.h +0
-0
arch/m68k/include/asm/page_no.h extern unsigned long memory_end; · line 35
Quoted 8 lines
arch/m68k/include/asm/page_no.h +2
-0@@ -28,6 +28,8 @@ extern unsigned long memory_end;#define virt_addr_valid (kaddr) (((unsigned long)(kaddr) >= PAGE_OFFSET) && \ ((unsigned long)(kaddr) < memory_end)) +#define ARCH_PFN_OFFSET PHYS_PFN(PAGE_OFFSET_RAW) +#endif /* __ASSEMBLY__ */ #endif /* _M68K_PAGE_NO_H */
Yes, that works.
Thanks,
Guenter
Quoted 38 lines
Quoted 37 lines
Guenter
---
bisect log:# bad: [
6ba8a227fd19d19779005fb66ad7562608e1df83] Add linux-next specific files for 20230210
# good: [4ec5183ec48656cec489c49f989c508b68b518e3] Linux 6.2-rc7
git bisect start 'HEAD' 'v6.2-rc7'
# good: [94613f0efc69ed41f9229ef5c294db3ec37145da] Merge branch 'master' ofgit://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good94613f0efc69ed41f9229ef5c294db3ec37145da
# good: [19e62c715fe70dae4582c2874ed3e66715d09af6] Merge branch 'rcu/next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git
git bisect good19e62c715fe70dae4582c2874ed3e66715d09af6
# good: [5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6] Merge branch 'for-next' of https://gitlab.com/peda-linux/mux.git
git bisect good5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6
# good: [c349bf6ec83903b20fe570c5609b9a864a64e09c] Merge branch 'for-next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode.git
git bisect goodc349bf6ec83903b20fe570c5609b9a864a64e09c
# good: [5a06a9f17454df38f35672be522ff5eb9b4277d2] selftest: add testing unsharing and counting ksm zero page
git bisect good5a06a9f17454df38f35672be522ff5eb9b4277d2
# bad: [f5d115a7b06e5661ed5218ffa9a2644c4ff1c135] Merge branch 'mm-nonmm-unstable' into mm-everything
git bisect badf5d115a7b06e5661ed5218ffa9a2644c4ff1c135
# bad: [acb018d6ea0c055381fba7dddaef386ee28f8075]mm/vmalloc.c: allowvread()to read outvm_map_ramareas
git bisect badacb018d6ea0c055381fba7dddaef386ee28f8075
# good: [1a5d9782ac969dc6e61c6786500b5160603188ea] mm/mmap: remove__vma_adjust()
git bisect good1a5d9782ac969dc6e61c6786500b5160603188ea
# good: [4b32363697de957dcc890b6245bec3f58903639a] arm: includeasm-generic/memory_model.hfrompage.hrather thanmemory.hgit bisect good4b32363697de957dcc890b6245bec3f58903639a
# bad: [328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8] mm/migrate: convertisolate_movable_page()to use folios
git bisect bad328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8
# bad: [b704c765b08cabe82adf76a4d1a74f3688eee410] mm/mempolicy: convertqueue_pages_pmd()toqueue_folios_pmd()
git bisect badb704c765b08cabe82adf76a4d1a74f3688eee410
# bad: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation ofpfn_valid()for FLATMEM
git bisect bade5734c8b0edfd2a053a5c256189586a3b1e9f63d
# bad: [ad8aecea034c591b9754bc5908da9719853aa7fa] mips: drop definition ofpfn_valid()for DISCONTIGMEM
git bisect badad8aecea034c591b9754bc5908da9719853aa7fa
# bad: [1f6271a0dfdf952c2e3981f424784d48f243a2be] m68k: useasm-generic/memory_model.hfor both MMU and !MMU
git bisect bad1f6271a0dfdf952c2e3981f424784d48f243a2be# first bad commit: [1f6271a0dfdf952c2e3981f424784d48f243a2be] m68k: use asm-generic/memory_model.h for both MMU and !MMU
PATCH v2 3/4 mips: drop definition of pfn_valid() for DISCONTIGMEM
In reply to Mike RapoportFrom: "Mike Rapoport (IBM)" <[email protected]>
There is stale definition of pfn_valid() for DISCONTINGMEM memory model guarded !FLATMEM && !SPARSEMEM && NUMA ifdefery.
Remove everything but definition of pfn_valid() for FLATMEM.
--- arch/mips/include/asm/page.h | 15 --------------- 1 file changed, 15 deletions(-)
arch/mips/include/asm/page.h +0 -15
@@ -235,21 +235,6 @@ static inline int pfn_valid(unsigned long pfn) return pfn >= pfn_offset && pfn < max_mapnr; } -#elif defined(CONFIG_SPARSEMEM) - -/* pfn_valid is defined in linux/mmzone.h */ - -#elif defined(CONFIG_NUMA) - -#define pfn_valid (pfn) \ -({ \ - unsigned long __pfn = (pfn); \ - int __n = pfn_to_nid(__pfn); \ - ((__n >= 0) ? (__pfn < NODE_DATA(__n)->node_start_pfn + \ - NODE_DATA(__n)->node_spanned_pages) \ - : 0); \ -}) -#endif #define virt_to_pfn (kaddr) PFN_DOWN(virt_to_phys((void *)(kaddr)))
Signature · 1 line
2.35.1
Re: [PATCH v2 3/4] mips: drop definition of pfn_valid() for DISCONTIGMEM
Quoted 8 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
There is stale definition of
pfn_valid()for DISCONTINGMEM memory model guarded !FLATMEM && !SPARSEMEM && NUMA ifdefery.Remove everything but definition of
pfn_valid()for FLATMEM.Signed-off-by Mike Rapoport (IBM) <[email protected]>
Signature · 3 lines
Thanks, David / dhildenb
PATCH v2 4/4 mm, arch: add generic implementation of pfn_valid() for FLATMEM
In reply to Mike RapoportFrom: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of pfn_valid() that essentially compares a pfn to max_mapnr.
Use mips/powerpc version implemented as static inline as a generic implementation of pfn_valid() and drop its per-architecture definitions.
--- arch/alpha/include/asm/page.h | 4 ---- arch/arc/include/asm/page.h | 1 - arch/csky/include/asm/page.h | 1 - arch/hexagon/include/asm/page.h | 1 - arch/ia64/include/asm/page.h | 4 ---- arch/loongarch/include/asm/page.h | 13 ------------- arch/m68k/include/asm/page_no.h | 2 -- arch/microblaze/include/asm/page.h | 1 - arch/mips/include/asm/page.h | 13 ------------- arch/nios2/include/asm/page.h | 9 --------- arch/openrisc/include/asm/page.h | 2 -- arch/parisc/include/asm/page.h | 4 ---- arch/powerpc/include/asm/page.h | 9 --------- arch/riscv/include/asm/page.h | 5 ----- arch/sh/include/asm/page.h | 3 --- arch/sparc/include/asm/page_32.h | 1 - arch/um/include/asm/page.h | 1 - arch/x86/include/asm/page_32.h | 4 ---- arch/x86/include/asm/page_64.h | 4 ---- arch/xtensa/include/asm/page.h | 4 ++-- include/asm-generic/memory_model.h | 12 ++++++++++++ include/asm-generic/page.h | 2 -- 22 files changed, 14 insertions(+), 86 deletions(-)
arch/alpha/include/asm/page.h +0 -4
@@ -87,10 +87,6 @@ typedef struct page *pgtable_t;#define virt_to_page (kaddr) pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)#define virt_addr_valid (kaddr) pfn_valid((__pa(kaddr) >> PAGE_SHIFT)) -#ifdef CONFIG_FLATMEM -#define pfn_valid (pfn) ((pfn) < max_mapnr) -#endif /* CONFIG_FLATMEM */ -#include <asm-generic/memory_model.h> #include <asm-generic/getorder.h>
arch/arc/include/asm/page.h +0 -1
@@ -109,7 +109,6 @@ extern int pfn_valid(unsigned long pfn);#else /* CONFIG_HIGHMEM */ #define ARCH_PFN_OFFSET virt_to_pfn(CONFIG_LINUX_RAM_BASE) -#define pfn_valid (pfn) (((pfn) - ARCH_PFN_OFFSET) < max_mapnr)#endif /* CONFIG_HIGHMEM */
arch/csky/include/asm/page.h +0 -1
@@ -39,7 +39,6 @@#define virt_addr_valid (kaddr) ((void *)(kaddr) >= (void *)PAGE_OFFSET && \ (void *)(kaddr) < high_memory) -#define pfn_valid (pfn) ((pfn) >= ARCH_PFN_OFFSET && ((pfn) - ARCH_PFN_OFFSET) < max_mapnr)extern void *memset (void *dest,int c,size_t l);extern void *memcpy (void *to,const void *from,size_t l);
arch/hexagon/include/asm/page.h +0 -1
@@ -95,7 +95,6 @@ struct page;/* Default vm area behavior is non-executable. */ #define VM_DATA_DEFAULT_FLAGS VM_DATA_FLAGS_NON_EXEC -#define pfn_valid (pfn) ((pfn) < max_mapnr)#define virt_addr_valid (kaddr) pfn_valid(__pa(kaddr) >> PAGE_SHIFT)/* Need to not use a define for linesize; may move this to another file. */
arch/ia64/include/asm/page.h +0 -4
@@ -97,10 +97,6 @@ do { \#include <asm-generic/memory_model.h> -#ifdef CONFIG_FLATMEM -# define pfn_valid (pfn) ((pfn) < max_mapnr) -#endif -#define page_to_phys (page) (page_to_pfn(page) << PAGE_SHIFT)#define virt_to_page (kaddr) pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)#define pfn_to_kaddr (pfn) __va((pfn) << PAGE_SHIFT)
arch/loongarch/include/asm/page.h +0 -13
@@ -82,19 +82,6 @@ typedef struct { unsigned long pgprot; } pgprot_t;#define pfn_to_kaddr (pfn) __va((pfn) << PAGE_SHIFT) -#ifdef CONFIG_FLATMEM - -static inline int pfn_valid (unsigned long pfn) -{ -/* avoid <linux/mm.h> include hell */ -extern unsigned long max_mapnr; -unsigned long pfn_offset =ARCH_PFN_OFFSET ; - -return pfn >= pfn_offset && pfn < max_mapnr; -} - -#endif -#define virt_to_pfn (kaddr) PFN_DOWN(PHYSADDR(kaddr))#define virt_to_page (kaddr) pfn_to_page(virt_to_pfn(kaddr))
arch/m68k/include/asm/page_no.h +0 -2
@@ -26,8 +26,6 @@ extern unsigned long memory_end;#define virt_to_page (addr) (mem_map + (((unsigned long)(addr)-PAGE_OFFSET) >> PAGE_SHIFT))#define page_to_virt (page) __va(((((page) - mem_map) << PAGE_SHIFT) + PAGE_OFFSET)) -#define pfn_valid (pfn) ((pfn) < max_mapnr) -#define virt_addr_valid (kaddr) (((unsigned long)(kaddr) >= PAGE_OFFSET) && \ ((unsigned long)(kaddr) < memory_end))
arch/microblaze/include/asm/page.h +0 -1
@@ -112,7 +112,6 @@ extern int page_is_ram(unsigned long pfn);# define page_to_phys (page) (page_to_pfn(page) << PAGE_SHIFT)# define ARCH_PFN_OFFSET (memory_start >> PAGE_SHIFT) -# define pfn_valid (pfn) ((pfn) >= ARCH_PFN_OFFSET && (pfn) < (max_mapnr + ARCH_PFN_OFFSET))# endif /* __ASSEMBLY__ */ #define virt_addr_valid (vaddr) (pfn_valid(virt_to_pfn(vaddr)))
arch/mips/include/asm/page.h +0 -13
@@ -224,19 +224,6 @@ extern phys_addr_t __phys_addr_symbol(unsigned long x);#define pfn_to_kaddr (pfn) __va((pfn) << PAGE_SHIFT) -#ifdef CONFIG_FLATMEM - -static inline int pfn_valid (unsigned long pfn) -{ -/* avoid <linux/mm.h> include hell */ -extern unsigned long max_mapnr; -unsigned long pfn_offset =ARCH_PFN_OFFSET ; - -return pfn >= pfn_offset && pfn < max_mapnr; -} - -#endif -#define virt_to_pfn (kaddr) PFN_DOWN(virt_to_phys((void *)(kaddr)))#define virt_to_page (kaddr) pfn_to_page(virt_to_pfn(kaddr))
arch/nios2/include/asm/page.h +0 -9
@@ -86,15 +86,6 @@ extern struct page *mem_map;# define pfn_to_kaddr (pfn) __va((pfn) << PAGE_SHIFT) -static inline bool pfn_valid (unsigned long pfn) -{ -/* avoid <linux/mm.h> include hell */ -extern unsigned long max_mapnr; -unsigned long pfn_offset =ARCH_PFN_OFFSET ; - -return pfn >= pfn_offset && pfn < max_mapnr; -} -# define virt_to_page (vaddr) pfn_to_page(PFN_DOWN(virt_to_phys(vaddr)))# define virt_addr_valid (vaddr) pfn_valid(PFN_DOWN(virt_to_phys(vaddr)))
arch/openrisc/include/asm/page.h +0 -2
@@ -80,8 +80,6 @@ typedef struct page *pgtable_t;#define page_to_phys (page) ((dma_addr_t)page_to_pfn(page) << PAGE_SHIFT) -#define pfn_valid (pfn) ((pfn) < max_mapnr) -#define virt_addr_valid (kaddr) (pfn_valid(virt_to_pfn(kaddr)))#endif /* __ASSEMBLY__ */
arch/parisc/include/asm/page.h +0 -4
@@ -155,10 +155,6 @@ extern int npmem_ranges;#define __pa (x) ((unsigned long)(x)-PAGE_OFFSET)#define __va (x) ((void *)((unsigned long)(x)+PAGE_OFFSET)) -#ifndef CONFIG_SPARSEMEM -#define pfn_valid (pfn) ((pfn) < max_mapnr) -#endif -#ifdef CONFIG_HUGETLB_PAGE #define HPAGE_SHIFT PMD_SHIFT/* fixed for transparent huge pages */ #define HPAGE_SIZE ((1UL) << HPAGE_SHIFT)
arch/powerpc/include/asm/page.h +0 -9
@@ -117,15 +117,6 @@ extern long long virt_phys_offset;#ifdef CONFIG_FLATMEM #define ARCH_PFN_OFFSET ((unsigned long)(MEMORY_START >> PAGE_SHIFT)) -#ifndef __ASSEMBLY__ -extern unsigned long max_mapnr; -static inline bool pfn_valid (unsigned long pfn) -{ -unsigned long min_pfn =ARCH_PFN_OFFSET ; - -return pfn >= min_pfn && pfn < max_mapnr; -} -#endif #endif #define virt_to_pfn (kaddr) (__pa(kaddr) >> PAGE_SHIFT)
arch/riscv/include/asm/page.h +0 -5
@@ -171,11 +171,6 @@ extern phys_addr_t __phys_addr_symbol(unsigned long x);#define sym_to_pfn (x) __phys_to_pfn(__pa_symbol(x)) -#ifdef CONFIG_FLATMEM -#define pfn_valid (pfn) \ - (((pfn) >= ARCH_PFN_OFFSET) && (((pfn) - ARCH_PFN_OFFSET) < max_mapnr)) -#endif -#endif /* __ASSEMBLY__ */ #define virt_addr_valid (vaddr) ({ \
arch/sh/include/asm/page.h +0 -3
@@ -169,9 +169,6 @@ typedef struct page *pgtable_t;#define PFN_START (__MEMORY_START >> PAGE_SHIFT)#define ARCH_PFN_OFFSET (PFN_START)#define virt_to_page (kaddr) pfn_to_page(__pa(kaddr) >> PAGE_SHIFT) -#ifdef CONFIG_FLATMEM -#define pfn_valid (pfn) ((pfn) >= min_low_pfn && (pfn) < max_low_pfn) -#endif #define virt_addr_valid (kaddr) pfn_valid(__pa(kaddr) >> PAGE_SHIFT)#include <asm-generic/memory_model.h>
arch/sparc/include/asm/page_32.h +0 -1
@@ -130,7 +130,6 @@ extern unsigned long pfn_base;#define ARCH_PFN_OFFSET (pfn_base)#define virt_to_page (kaddr) pfn_to_page(__pa(kaddr) >> PAGE_SHIFT) -#define pfn_valid (pfn) (((pfn) >= (pfn_base)) && (((pfn)-(pfn_base)) < max_mapnr))#define virt_addr_valid (kaddr) ((((unsigned long)(kaddr)-PAGE_OFFSET)>>PAGE_SHIFT) < max_mapnr)#include <asm-generic/memory_model.h>
arch/um/include/asm/page.h +0 -1
@@ -108,7 +108,6 @@ extern unsigned long uml_physmem;#define phys_to_pfn (p) ((p) >> PAGE_SHIFT)#define pfn_to_phys (pfn) PFN_PHYS(pfn) -#define pfn_valid (pfn) ((pfn) < max_mapnr)#define virt_addr_valid (v) pfn_valid(phys_to_pfn(__pa(v)))#include <asm-generic/memory_model.h>
arch/x86/include/asm/page_32.h +0 -4
@@ -15,10 +15,6 @@ extern unsigned long __phys_addr(unsigned long);#define __phys_addr_symbol (x) __phys_addr(x)#define __phys_reloc_hide (x) RELOC_HIDE((x), 0) -#ifdef CONFIG_FLATMEM -#define pfn_valid (pfn) ((pfn) < max_mapnr) -#endif /* CONFIG_FLATMEM */ -#include <linux/string.h> static inline void clear_page (void *page)
arch/x86/include/asm/page_64.h +0 -4
@@ -39,10 +39,6 @@ extern unsigned long __phys_addr_symbol(unsigned long);#define __phys_reloc_hide (x) (x) -#ifdef CONFIG_FLATMEM -#define pfn_valid (pfn) ((pfn) < max_pfn) -#endif -void clear_page_orig (void *page);void clear_page_rep (void *page);void clear_page_erms (void *page);
arch/xtensa/include/asm/page.h +2 -2
@@ -11,6 +11,8 @@#ifndef _XTENSA_PAGE_H#define _XTENSA_PAGE_H +#include <linux/const.h> +#include <asm/processor.h> #include <asm/types.h> #include <asm/cache.h>
@@ -189,8 +191,6 @@ static inline unsigned long ___pa(unsigned long va)#endif #define __va (x) \ ((void *)((unsigned long) (x) - PHYS_OFFSET + PAGE_OFFSET)) -#define pfn_valid (pfn) \ - ((pfn) >= ARCH_PFN_OFFSET && ((pfn) - ARCH_PFN_OFFSET) < max_mapnr)#define virt_to_page (kaddr) pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)#define page_to_virt (page) __va(page_to_pfn(page) << PAGE_SHIFT)
include/asm-generic/memory_model.h +12 -0
@@ -19,6 +19,18 @@#define __page_to_pfn (page) ((unsigned long)((page) - mem_map) + \ ARCH_PFN_OFFSET) +#ifndef pfn_valid +static inline int pfn_valid (unsigned long pfn) +{ +/* avoid <linux/mm.h> include hell */ +extern unsigned long max_mapnr; +unsigned long pfn_offset =ARCH_PFN_OFFSET ; + +return pfn >= pfn_offset && pfn < max_mapnr; +} +#define pfn_valid pfn_valid +#endif +#elif defined(CONFIG_SPARSEMEM_VMEMMAP )/* memmap is virtually contiguous. */
include/asm-generic/page.h +0 -2
@@ -84,8 +84,6 @@ extern unsigned long memory_end;#define page_to_phys (page) ((dma_addr_t)page_to_pfn(page) << PAGE_SHIFT)#endif -#define pfn_valid (pfn) ((pfn) >= ARCH_PFN_OFFSET && ((pfn) - ARCH_PFN_OFFSET) < max_mapnr) -#define virt_addr_valid (kaddr) (((void *)(kaddr) >= (void *)PAGE_OFFSET) && \ ((void *)(kaddr) < (void *)memory_end))
Signature · 1 line
2.35.1
Quoted 13 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.Signed-off-by Mike Rapoport (IBM) <[email protected]>Acked-by Arnd Bergmann <[email protected]>Acked-by Guo Ren <[email protected]> # csky
Quoted 14 lines
--- arch/alpha/include/asm/page.h | 4 ---- arch/arc/include/asm/page.h | 1 - arch/csky/include/asm/page.h | 1 - arch/hexagon/include/asm/page.h | 1 - arch/ia64/include/asm/page.h | 4 ---- arch/loongarch/include/asm/page.h | 13 ------------- arch/m68k/include/asm/page_no.h | 2 -- arch/microblaze/include/asm/page.h | 1 - arch/mips/include/asm/page.h | 13 ------------- arch/nios2/include/asm/page.h | 9 --------- arch/openrisc/include/asm/page.h | 2 -- arch/parisc/include/asm/page.h | 4 ---- arch/powerpc/include/asm/page.h | 9 ---------
cheers
Hey Mike,
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.
Quoted 8 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.Signed-off-by Mike Rapoport (IBM) <[email protected]>Acked-by Arnd Bergmann <[email protected]>Acked-by Guo Ren <[email protected]> # csky
Hmm, so this landed in linux-next today and I bisected a boot failure in my CI to it. However, I am not really sure if it is a real issue worth worrying about as the platform it triggered on is supposed to be using SPARSEMEM, but isn't.
I had thought that my CI was using a config with SPARSEMEM since that became required for riscv defconfig builds to boot in v6.1-rc1, but I must have just forgotten to add it to my $platform_defconfig builds too. However, those $platform_defconfig builds continued booting without SPARSEMEM enabled until today.
for $platform_defconfig builds I now see the following, which is in turn followed by an opps. The full output is at the bottom of this mail.
WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:479 __vmap_pages_range_noflush+0x2d4/0x4c8 Modules linked in: CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.0-rc4-00538-g6e265f6be0c8 #1 Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) epc : __vmap_pages_range_noflush+0x2d4/0x4c8 ra : __vmap_pages_range_noflush+0x3e0/0x4c8 epc : ffffffff801461e4 ra : ffffffff801462f0 sp : ffffffff81203b90 gp : ffffffff812e9fe0 tp : ffffffff8120fa80 t0 : ffffffe7bfe57000 t1 : ffffffffffff8000 t2 : 00000000000001a7 s0 : ffffffff81203c50 s1 : ffffffe7bfe56040 a0 : 0000000000000000 a1 : ffffffe7c7df2840 a2 : ffffffff812ebae8 a3 : 0000000000001000 a4 : 0000000000080200 a5 : 0000000000fffe00 a6 : 0000000001040056 a7 : ffffffffffffffff s2 : 0000000001040024 s3 : ffffffff80de1118 s4 : ffffffc80400c000 s5 : ffffffc80400c000 s6 : 00000000000000e7 s7 : 0000000000000000 s8 : ffffffc80400bfff s9 : 00fffffffffff000 s10: ffffffc804008000 s11: ffffffe7bfe0d100 t3 : 0000000000000000 t4 : fffffffffff00000 t5 : 00000000000f0000 t6 : ffffffe7bfe22ac4 status: 0000000200000120 badaddr: 0000000000000000 cause: 0000000000000003 [<ffffffff8014762a>] __vmalloc_node_range+0x392/0x52e [<ffffffff8000dfc4>] copy_process+0x636/0x1196 [<ffffffff8000ec42>] kernel_clone+0x4a/0x2f4 [<ffffffff8000f15e>] user_mode_thread+0x7c/0x9a [<ffffffff8077056e>] rest_init+0x28/0xea [<ffffffff80800670>] arch_post_acpi_subsys_init+0x0/0x18 [<ffffffff80800de4>] start_kernel+0x72c/0x75c
In my mind, there's a kernel misconfiguration issue, but it's not the same splat as I used to get when using FLATMEM in this configuration:
OF: fdt: Ignoring memory range 0x80000000 - 0x80200000 Machine model: Microchip PolarFire-SoC Icicle Kit earlycon: ns16550a0 at MMIO32 0x0000000020100000 (options '115200n8') printk: bootconsole [ns16550a0] enabled printk: debug: skip boot console de-registration. efi: UEFI not found. Zone ranges: DMA32 [mem 0x0000000080200000-0x00000000ffffffff] Normal [mem 0x0000000100000000-0x000000107fffffff] Movable zone start for each node Early memory node ranges node 0: [mem 0x0000000080200000-0x00000000bfbfffff] node 0: [mem 0x00000000bfc00000-0x00000000bfffffff] node 0: [mem 0x0000001040000000-0x000000107fffffff] Initmem setup node 0 [mem 0x0000000080200000-0x000000107fffffff] Kernel panic - not syncing: Failed to allocate 1073741824 bytes for node 0 memory map CPU: 0 PID: 0 Comm: swapper Not tainted 5.19.0-dirty #1 Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) Call Trace: [<ffffffff800057f0>] show_stack+0x30/0x3c [<ffffffff807d5802>] dump_stack_lvl+0x4a/0x66 [<ffffffff807d5836>] dump_stack+0x18/0x20 [<ffffffff807d1ae8>] panic+0x124/0x2c6 [<ffffffff80814064>] free_area_init_core+0x0/0x11e [<ffffffff80813720>] free_area_init_node+0xc2/0xf6 [<ffffffff8081331e>] free_area_init+0x222/0x260 [<ffffffff808064d6>] misc_mem_init+0x62/0x9a [<ffffffff80803cb2>] setup_arch+0xb0/0xea [<ffffffff8080039a>] start_kernel+0x88/0x4ee
Turning on SPARSEMEM fixes both problems, but I'm just not sure if there is some underlying issue here that I don't know enough about the area to understand.
Thanks,
Conor.
[ 0.000000] Linux version 6.2.0-rc4-00538-g6e265f6be0c8 (conor@wendy) (riscv64-unknown-linux-gnu-gcc (g5964b5cd727) 11.1.0, GNU ld (GNU Binutils) 2.37) #1 SMP PREEMPT @7 [ 0.000000] OF: fdt: Ignoring memory range 0x80000000 - 0x80200000 [ 0.000000] Machine model: Microchip PolarFire-SoC Icicle Kit [ 0.000000] earlycon: ns16550a0 at MMIO32 0x0000000020100000 (options '115200n8') [ 0.000000] printk: bootconsole [ns16550a0] enabled [ 0.000000] printk: debug: skip boot console de-registration. [ 0.000000] efi: UEFI not found. [ 0.000000] Zone ranges: [ 0.000000] DMA32 [mem 0x0000000080200000-0x00000000ffffffff] [ 0.000000] Normal [mem 0x0000000100000000-0x000000107fffffff] [ 0.000000] Movable zone start for each node [ 0.000000] Early memory node ranges [ 0.000000] node 0: [mem 0x0000000080200000-0x00000000bfbfffff] [ 0.000000] node 0: [mem 0x00000000bfc00000-0x00000000bfffffff] [ 0.000000] node 0: [mem 0x0000001040000000-0x000000107fffffff] [ 0.000000] Initmem setup node 0 [mem 0x0000000080200000-0x000000107fffffff] [ 0.000000] On node 0, zone Normal: 15990272 pages in unavailable ranges [ 0.000000] SBI specification v0.3 detected [ 0.000000] SBI implementation ID=0x1 Version=0x10000 [ 0.000000] SBI TIME extension detected [ 0.000000] SBI IPI extension detected [ 0.000000] SBI RFENCE extension detected [ 0.000000] SBI SRST extension detected [ 0.000000] SBI HSM extension detected [ 0.000000] CPU with hartid=0 is not available [ 0.000000] CPU with hartid=0 is not available [ 0.000000] CPU with hartid=0 is not available [ 0.000000] riscv: base ISA extensions acdfim [ 0.000000] riscv: ELF capabilities acdfim [ 0.000000] percpu: Embedded 19 pages/cpu s37048 r8192 d32584 u77824 [ 0.000000] CPU node for /cpus/cpu@0 exist but the possible cpu range is :0-3 [ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 294407 [ 0.000000] Kernel command line: earlycon keep_bootcon reboot=cold [ 0.000000] Dentry cache hash table entries: 262144 (order: 9, 2097152 bytes, linear) [ 0.000000] Inode-cache hash table entries: 131072 (order: 8, 1048576 bytes, linear) [ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off [ 0.000000] software IO TLB: area num 4. [ 0.000000] software IO TLB: mapped [mem 0x00000000bbc00000-0x00000000bfc00000] (64MB) [ 0.000000] Virtual kernel memory layout: [ 0.000000] fixmap : 0xffffffc6fee00000 - 0xffffffc6ff000000 (2048 kB) [ 0.000000] pci io : 0xffffffc6ff000000 - 0xffffffc700000000 ( 16 MB) [ 0.000000] vmemmap : 0xffffffc700000000 - 0xffffffc800000000 (4096 MB) [ 0.000000] vmalloc : 0xffffffc800000000 - 0xffffffd800000000 ( 64 GB) [ 0.000000] modules : 0xffffffff01352000 - 0xffffffff80000000 (2028 MB) [ 0.000000] lowmem : 0xffffffd800000000 - 0xffffffe7ffe00000 ( 63 GB) [ 0.000000] kernel : 0xffffffff80000000 - 0xffffffffffffffff (2047 MB) [ 0.000000] Memory: 1067600K/2095104K available (7639K kernel code, 4898K rwdata, 4096K rodata, 2183K init, 410K bss, 1027504K reserved, 0K cma-reserved) [ 0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=4, Nodes=1 [ 0.000000] rcu: Preemptible hierarchical RCU implementation. [ 0.000000] rcu: RCU restricting CPUs from NR_CPUS=64 to nr_cpu_ids=4. [ 0.000000] rcu: RCU debug extended QS entry/exit. [ 0.000000] Trampoline variant of Tasks RCU enabled. [ 0.000000] Tracing variant of Tasks RCU enabled. [ 0.000000] rcu: RCU calculated value of scheduler-enlistment delay is 25 jiffies. [ 0.000000] rcu: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=4 [ 0.000000] NR_IRQS: 64, nr_irqs: 64, preallocated irqs: 0 [ 0.000000] CPU with hartid=0 is not available [ 0.000000] riscv-intc: unable to find hart id for /cpus/cpu@0/interrupt-controller [ 0.000000] riscv-intc: 64 local interrupts mapped [ 0.000000] plic: interrupt-controller@c000000: mapped 186 interrupts with 4 handlers for 9 contexts. [ 0.000000] rcu: srcu_init: Setting srcu_struct sizes based on contention. [ 0.000000] riscv-timer: riscv_timer_init_dt: Registering clocksource cpuid [0] hartid [1] [ 0.000000] clocksource: riscv_clocksource: mask: 0xffffffffffffffff max_cycles: 0x1d854df40, max_idle_ns: 3526361616960 ns [ 0.000004] sched_clock: 64 bits at 1000kHz, resolution 1000ns, wraps every 2199023255500ns [ 0.009687] Console: colour dummy device 80x25 [ 0.014654] printk: console [tty0] enabled [ 0.019313] Calibrating delay loop (skipped), value calculated using timer frequency.. 2.00 BogoMIPS (lpj=4000) [ 0.030490] pid_max: default: 32768 minimum: 301 [ 0.036094] Mount-cache hash table entries: 4096 (order: 3, 32768 bytes, linear) [ 0.044458] Mountpoint-cache hash table entries: 4096 (order: 3, 32768 bytes, linear) [ 0.055493] ------------[ cut here ]------------ [ 0.060606] WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:479 __vmap_pages_range_noflush+0x2d4/0x4c8 [ 0.070040] Modules linked in: [ 0.073421] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.0-rc4-00538-g6e265f6be0c8 #1 [ 0.082157] Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) [ 0.088972] epc : __vmap_pages_range_noflush+0x2d4/0x4c8 [ 0.094838] ra : __vmap_pages_range_noflush+0x3e0/0x4c8 [ 0.100707] epc : ffffffff801461e4 ra : ffffffff801462f0 sp : ffffffff81203b90 [ 0.108674] gp : ffffffff812e9fe0 tp : ffffffff8120fa80 t0 : ffffffe7bfe57000 [ 0.116643] t1 : ffffffffffff8000 t2 : 00000000000001a7 s0 : ffffffff81203c50 [ 0.124611] s1 : ffffffe7bfe56040 a0 : 0000000000000000 a1 : ffffffe7c7df2840 [ 0.132588] a2 : ffffffff812ebae8 a3 : 0000000000001000 a4 : 0000000000080200 [ 0.140564] a5 : 0000000000fffe00 a6 : 0000000001040056 a7 : ffffffffffffffff [ 0.148541] s2 : 0000000001040024 s3 : ffffffff80de1118 s4 : ffffffc80400c000 [ 0.156518] s5 : ffffffc80400c000 s6 : 00000000000000e7 s7 : 0000000000000000 [ 0.164495] s8 : ffffffc80400bfff s9 : 00fffffffffff000 s10: ffffffc804008000 [ 0.172472] s11: ffffffe7bfe0d100 t3 : 0000000000000000 t4 : fffffffffff00000 [ 0.180449] t5 : 00000000000f0000 t6 : ffffffe7bfe22ac4 [ 0.186317] status: 0000000200000120 badaddr: 0000000000000000 cause: 0000000000000003 [ 0.195060] [<ffffffff8014762a>] __vmalloc_node_range+0x392/0x52e [ 0.201802] [<ffffffff8000dfc4>] copy_process+0x636/0x1196 [ 0.207878] [<ffffffff8000ec42>] kernel_clone+0x4a/0x2f4 [ 0.213755] [<ffffffff8000f15e>] user_mode_thread+0x7c/0x9a [ 0.219918] [<ffffffff8077056e>] rest_init+0x28/0xea [ 0.225429] [<ffffffff80800670>] arch_post_acpi_subsys_init+0x0/0x18 [ 0.232460] [<ffffffff80800de4>] start_kernel+0x72c/0x75c [ 0.238434] ---[ end trace 0000000000000000 ]--- n):0kB isolated(file):0kB mapped:0kB dirty:0kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 0kB anon_thp: 0kB writeback_tmp:0kB kernel_stack:0kB pagetables:4kB sec_pagetables:0kB all_unreclaimable? no [ 0.423801] DMA32 free:940780kB boost:0kB min:0kB low:0kB high:0kB reserved_highatomic:0KB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB writepending:0kB present:1046528kB managed:940780kB mlocked:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB [ 0.452034] lowmem_reserve[]: 0 0 0 [ 0.455919] Normal free:126248kB boost:0kB min:0kB low:0kB high:0kB reserved_highatomic:0KB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB writepending:0kB present:1048576kB managed:126820kB mlocked:0kB bounce:0kB free_pcp:156kB local_pcp:156kB free_cma:0kB [ 0.484652] lowmem_reserve[]: 0 0 0 [ 0.488538] DMA32: 1*4kB (M) 1*8kB (M) 0*16kB 1*32kB (M) 1*64kB (M) 1*128kB (M) 0*256kB 1*512kB (M) 0*1024kB 1*2048kB (M) 229*4096kB (M) = 940780kB [ 0.503367] Normal: 2*4kB (ME) 0*8kB 2*16kB (UE) 2*32kB (ME) 3*64kB (UME) 0*128kB 2*256kB (ME) 3*512kB (UME) 3*1024kB (UME) 3*2048kB (UME) 28*4096kB (M) = 126248kB [ 0.519722] 0 total pagecache pages [ 0.523579] 0 pages in swap cache [ 0.527264] Free swap = 0kB [ 0.530445] Total swap = 0kB [ 0.533652] 523776 pages RAM [ 0.536833] 0 pages HighMem/MovableOnly [ 0.541090] 256876 pages reserved [ 0.544833] Unable to handle kernel NULL pointer dereference at virtual address 000000000000003c [ 0.554576] Oops [#1] [ 0.557092] Modules linked in: [ 0.560470] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G W 6.2.0-rc4-00538-g6e265f6be0c8 #1 [ 0.570841] Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) [ 0.577664] epc : rest_init+0x44/0xea [ 0.581726] ra : rest_init+0x44/0xea [ 0.585788] epc : ffffffff8077058a ra : ffffffff8077058a sp : ffffffff81203f70 [ 0.593764] gp : ffffffff812e9fe0 tp : ffffffff8120fa80 t0 : 0000000000000000 [ 0.601741] t1 : 0000000000000000 t2 : 0000000000000040 s0 : ffffffff81203f90 [ 0.609718] s1 : fffffffffffffff4 a0 : 0000000000000000 a1 : fffffffffffffff4 [ 0.617695] a2 : 0000000000000000 a3 : 0000000000000000 a4 : 0000000000000000 [ 0.625672] a5 : 0000000000000000 a6 : 0000000001212dec a7 : ffffffffd8bc86ae [ 0.633649] s2 : ffffffff80de0bf0 s3 : ffffffe7c7de7700 s4 : 0000000000000000 [ 0.641626] s5 : ffffffff812eb018 s6 : ffffffff80a00008 s7 : 0000000000000000 [ 0.649603] s8 : 00000000bfb85036 s9 : 0000000000000000 s10: 0000000000000000 [ 0.657580] s11: 0000000000000000 t3 : 0000000000000002 t4 : 0000000000000402 [ 0.665557] t5 : ffffffff812158a8 t6 : 0000000000000000 [ 0.671416] status: 0000000200000120 badaddr: 000000000000003c cause: 000000000000000d [ 0.680157] [<ffffffff80800670>] arch_post_acpi_subsys_init+0x0/0x18 [ 0.687188] [<ffffffff80800de4>] start_kernel+0x72c/0x75c [ 0.693195] ---[ end trace 0000000000000000 ]--- [ 0.698325] Kernel panic - not syncing: Attempted to kill the idle task! [ 0.705729] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---
Hi Conor,
Quoted 17 lines · Conor Dooley
Hey Mike,
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Quoted 8 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.Signed-off-by Mike Rapoport (IBM) <[email protected]>Acked-by Arnd Bergmann <[email protected]>Acked-by Guo Ren <[email protected]> # csky
Quoted 9 lines
Hmm, so this landed in linux-next today and I bisected a boot failure in my CI to it. However, I am not really sure if it is a real issue worth worrying about as the platform it triggered on is supposed to be using SPARSEMEM, but isn't.
I had thought that my CI was using a config with SPARSEMEM since that became required for riscv defconfig builds to boot in v6.1-rc1, but I must have just forgotten to add it to my $platform_defconfig builds too. However, those $platform_defconfig builds continued booting without SPARSEMEM enabled until today.
The issue seems to be that the generic pfn_valid() does not take into account pfn_offset when it compares it with max_mapnr.
Can you please test with the patch below?
include/asm-generic/memory_model.h +1 -1
@@ -26,7 +26,7 @@ static inline int pfn_valid(unsigned long pfn)extern unsigned long max_mapnr;unsigned long pfn_offset =ARCH_PFN_OFFSET ; - return pfn >= pfn_offset && pfn < max_mapnr; + return pfn >= pfn_offset && (pfn - pfn_offset) < max_mapnr; }#define pfn_valid pfn_valid#endif
Quoted 204 lines
for $platform_defconfig builds I now see the following, which is in turn followed by an opps. The full output is at the bottom of this mail.
WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:479 __vmap_pages_range_noflush+0x2d4/0x4c8 Modules linked in: CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.0-rc4-00538-g6e265f6be0c8 #1 Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) epc : __vmap_pages_range_noflush+0x2d4/0x4c8 ra : __vmap_pages_range_noflush+0x3e0/0x4c8 epc : ffffffff801461e4 ra : ffffffff801462f0 sp : ffffffff81203b90 gp : ffffffff812e9fe0 tp : ffffffff8120fa80 t0 : ffffffe7bfe57000 t1 : ffffffffffff8000 t2 : 00000000000001a7 s0 : ffffffff81203c50 s1 : ffffffe7bfe56040 a0 : 0000000000000000 a1 : ffffffe7c7df2840 a2 : ffffffff812ebae8 a3 : 0000000000001000 a4 : 0000000000080200 a5 : 0000000000fffe00 a6 : 0000000001040056 a7 : ffffffffffffffff s2 : 0000000001040024 s3 : ffffffff80de1118 s4 : ffffffc80400c000 s5 : ffffffc80400c000 s6 : 00000000000000e7 s7 : 0000000000000000 s8 : ffffffc80400bfff s9 : 00fffffffffff000 s10: ffffffc804008000 s11: ffffffe7bfe0d100 t3 : 0000000000000000 t4 : fffffffffff00000 t5 : 00000000000f0000 t6 : ffffffe7bfe22ac4 status: 0000000200000120 badaddr: 0000000000000000 cause: 0000000000000003 [<ffffffff8014762a>] __vmalloc_node_range+0x392/0x52e [<ffffffff8000dfc4>] copy_process+0x636/0x1196 [<ffffffff8000ec42>] kernel_clone+0x4a/0x2f4 [<ffffffff8000f15e>] user_mode_thread+0x7c/0x9a [<ffffffff8077056e>] rest_init+0x28/0xea [<ffffffff80800670>] arch_post_acpi_subsys_init+0x0/0x18 [<ffffffff80800de4>] start_kernel+0x72c/0x75cIn my mind, there's a kernel misconfiguration issue, but it's not the same splat as I used to get when using FLATMEM in this configuration:
OF: fdt: Ignoring memory range 0x80000000 - 0x80200000 Machine model: Microchip PolarFire-SoC Icicle Kit earlycon: ns16550a0 at MMIO32 0x0000000020100000 (options '115200n8') printk: bootconsole [ns16550a0] enabled printk: debug: skip boot console de-registration. efi: UEFI not found. Zone ranges: DMA32 [mem 0x0000000080200000-0x00000000ffffffff] Normal [mem 0x0000000100000000-0x000000107fffffff] Movable zone start for each node Early memory node ranges node 0: [mem 0x0000000080200000-0x00000000bfbfffff] node 0: [mem 0x00000000bfc00000-0x00000000bfffffff] node 0: [mem 0x0000001040000000-0x000000107fffffff] Initmem setup node 0 [mem 0x0000000080200000-0x000000107fffffff] Kernel panic - not syncing: Failed to allocate 1073741824 bytes for node 0 memory map CPU: 0 PID: 0 Comm: swapper Not tainted 5.19.0-dirty #1 Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) Call Trace: [<ffffffff800057f0>] show_stack+0x30/0x3c [<ffffffff807d5802>] dump_stack_lvl+0x4a/0x66 [<ffffffff807d5836>] dump_stack+0x18/0x20 [<ffffffff807d1ae8>] panic+0x124/0x2c6 [<ffffffff80814064>] free_area_init_core+0x0/0x11e [<ffffffff80813720>] free_area_init_node+0xc2/0xf6 [<ffffffff8081331e>] free_area_init+0x222/0x260 [<ffffffff808064d6>] misc_mem_init+0x62/0x9a [<ffffffff80803cb2>] setup_arch+0xb0/0xea [<ffffffff8080039a>] start_kernel+0x88/0x4eeTurning on SPARSEMEM fixes both problems, but I'm just not sure if there is some underlying issue here that I don't know enough about the area to understand.
Thanks,
Conor.[ 0.000000] Linux version 6.2.0-rc4-00538-g6e265f6be0c8 (conor@wendy) (riscv64-unknown-linux-gnu-gcc (g5964b5cd727) 11.1.0, GNU ld (GNU Binutils) 2.37) #1 SMP PREEMPT @7 [ 0.000000] OF: fdt: Ignoring memory range 0x80000000 - 0x80200000 [ 0.000000] Machine model: Microchip PolarFire-SoC Icicle Kit [ 0.000000] earlycon: ns16550a0 at MMIO32 0x0000000020100000 (options '115200n8') [ 0.000000] printk: bootconsole [ns16550a0] enabled [ 0.000000] printk: debug: skip boot console de-registration. [ 0.000000] efi: UEFI not found. [ 0.000000] Zone ranges: [ 0.000000] DMA32 [mem 0x0000000080200000-0x00000000ffffffff] [ 0.000000] Normal [mem 0x0000000100000000-0x000000107fffffff] [ 0.000000] Movable zone start for each node [ 0.000000] Early memory node ranges [ 0.000000] node 0: [mem 0x0000000080200000-0x00000000bfbfffff] [ 0.000000] node 0: [mem 0x00000000bfc00000-0x00000000bfffffff] [ 0.000000] node 0: [mem 0x0000001040000000-0x000000107fffffff] [ 0.000000] Initmem setup node 0 [mem 0x0000000080200000-0x000000107fffffff] [ 0.000000] On node 0, zone Normal: 15990272 pages in unavailable ranges [ 0.000000] SBI specification v0.3 detected [ 0.000000] SBI implementation ID=0x1 Version=0x10000 [ 0.000000] SBI TIME extension detected [ 0.000000] SBI IPI extension detected [ 0.000000] SBI RFENCE extension detected [ 0.000000] SBI SRST extension detected [ 0.000000] SBI HSM extension detected [ 0.000000] CPU with hartid=0 is not available [ 0.000000] CPU with hartid=0 is not available [ 0.000000] CPU with hartid=0 is not available [ 0.000000] riscv: base ISA extensions acdfim [ 0.000000] riscv: ELF capabilities acdfim [ 0.000000] percpu: Embedded 19 pages/cpu s37048 r8192 d32584 u77824 [ 0.000000] CPU node for /cpus/cpu@0 exist but the possible cpu range is :0-3 [ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 294407 [ 0.000000] Kernel command line: earlycon keep_bootcon reboot=cold [ 0.000000] Dentry cache hash table entries: 262144 (order: 9, 2097152 bytes, linear) [ 0.000000] Inode-cache hash table entries: 131072 (order: 8, 1048576 bytes, linear) [ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off [ 0.000000] software IO TLB: area num 4. [ 0.000000] software IO TLB: mapped [mem 0x00000000bbc00000-0x00000000bfc00000] (64MB) [ 0.000000] Virtual kernel memory layout: [ 0.000000] fixmap : 0xffffffc6fee00000 - 0xffffffc6ff000000 (2048 kB) [ 0.000000] pci io : 0xffffffc6ff000000 - 0xffffffc700000000 ( 16 MB) [ 0.000000] vmemmap : 0xffffffc700000000 - 0xffffffc800000000 (4096 MB) [ 0.000000] vmalloc : 0xffffffc800000000 - 0xffffffd800000000 ( 64 GB) [ 0.000000] modules : 0xffffffff01352000 - 0xffffffff80000000 (2028 MB) [ 0.000000] lowmem : 0xffffffd800000000 - 0xffffffe7ffe00000 ( 63 GB) [ 0.000000] kernel : 0xffffffff80000000 - 0xffffffffffffffff (2047 MB) [ 0.000000] Memory: 1067600K/2095104K available (7639K kernel code, 4898K rwdata, 4096K rodata, 2183K init, 410K bss, 1027504K reserved, 0K cma-reserved) [ 0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=4, Nodes=1 [ 0.000000] rcu: Preemptible hierarchical RCU implementation. [ 0.000000] rcu: RCU restricting CPUs from NR_CPUS=64 to nr_cpu_ids=4. [ 0.000000] rcu: RCU debug extended QS entry/exit. [ 0.000000] Trampoline variant of Tasks RCU enabled. [ 0.000000] Tracing variant of Tasks RCU enabled. [ 0.000000] rcu: RCU calculated value of scheduler-enlistment delay is 25 jiffies. [ 0.000000] rcu: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=4 [ 0.000000] NR_IRQS: 64, nr_irqs: 64, preallocated irqs: 0 [ 0.000000] CPU with hartid=0 is not available [ 0.000000] riscv-intc: unable to find hart id for /cpus/cpu@0/interrupt-controller [ 0.000000] riscv-intc: 64 local interrupts mapped [ 0.000000] plic: interrupt-controller@c000000: mapped 186 interrupts with 4 handlers for 9 contexts. [ 0.000000] rcu: srcu_init: Setting srcu_struct sizes based on contention. [ 0.000000] riscv-timer: riscv_timer_init_dt: Registering clocksource cpuid [0] hartid [1] [ 0.000000] clocksource: riscv_clocksource: mask: 0xffffffffffffffff max_cycles: 0x1d854df40, max_idle_ns: 3526361616960 ns [ 0.000004] sched_clock: 64 bits at 1000kHz, resolution 1000ns, wraps every 2199023255500ns [ 0.009687] Console: colour dummy device 80x25 [ 0.014654] printk: console [tty0] enabled [ 0.019313] Calibrating delay loop (skipped), value calculated using timer frequency.. 2.00 BogoMIPS (lpj=4000) [ 0.030490] pid_max: default: 32768 minimum: 301 [ 0.036094] Mount-cache hash table entries: 4096 (order: 3, 32768 bytes, linear) [ 0.044458] Mountpoint-cache hash table entries: 4096 (order: 3, 32768 bytes, linear) [ 0.055493] ------------[ cut here ]------------ [ 0.060606] WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:479 __vmap_pages_range_noflush+0x2d4/0x4c8 [ 0.070040] Modules linked in: [ 0.073421] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.0-rc4-00538-g6e265f6be0c8 #1 [ 0.082157] Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) [ 0.088972] epc : __vmap_pages_range_noflush+0x2d4/0x4c8 [ 0.094838] ra : __vmap_pages_range_noflush+0x3e0/0x4c8 [ 0.100707] epc : ffffffff801461e4 ra : ffffffff801462f0 sp : ffffffff81203b90 [ 0.108674] gp : ffffffff812e9fe0 tp : ffffffff8120fa80 t0 : ffffffe7bfe57000 [ 0.116643] t1 : ffffffffffff8000 t2 : 00000000000001a7 s0 : ffffffff81203c50 [ 0.124611] s1 : ffffffe7bfe56040 a0 : 0000000000000000 a1 : ffffffe7c7df2840 [ 0.132588] a2 : ffffffff812ebae8 a3 : 0000000000001000 a4 : 0000000000080200 [ 0.140564] a5 : 0000000000fffe00 a6 : 0000000001040056 a7 : ffffffffffffffff [ 0.148541] s2 : 0000000001040024 s3 : ffffffff80de1118 s4 : ffffffc80400c000 [ 0.156518] s5 : ffffffc80400c000 s6 : 00000000000000e7 s7 : 0000000000000000 [ 0.164495] s8 : ffffffc80400bfff s9 : 00fffffffffff000 s10: ffffffc804008000 [ 0.172472] s11: ffffffe7bfe0d100 t3 : 0000000000000000 t4 : fffffffffff00000 [ 0.180449] t5 : 00000000000f0000 t6 : ffffffe7bfe22ac4 [ 0.186317] status: 0000000200000120 badaddr: 0000000000000000 cause: 0000000000000003 [ 0.195060] [<ffffffff8014762a>] __vmalloc_node_range+0x392/0x52e [ 0.201802] [<ffffffff8000dfc4>] copy_process+0x636/0x1196 [ 0.207878] [<ffffffff8000ec42>] kernel_clone+0x4a/0x2f4 [ 0.213755] [<ffffffff8000f15e>] user_mode_thread+0x7c/0x9a [ 0.219918] [<ffffffff8077056e>] rest_init+0x28/0xea [ 0.225429] [<ffffffff80800670>] arch_post_acpi_subsys_init+0x0/0x18 [ 0.232460] [<ffffffff80800de4>] start_kernel+0x72c/0x75c [ 0.238434] ---[ end trace 0000000000000000 ]--- n):0kB isolated(file):0kB mapped:0kB dirty:0kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 0kB anon_thp: 0kB writeback_tmp:0kB kernel_stack:0kB pagetables:4kB sec_pagetables:0kB all_unreclaimable? no [ 0.423801] DMA32 free:940780kB boost:0kB min:0kB low:0kB high:0kB reserved_highatomic:0KB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB writepending:0kB present:1046528kB managed:940780kB mlocked:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB [ 0.452034] lowmem_reserve[]: 0 0 0 [ 0.455919] Normal free:126248kB boost:0kB min:0kB low:0kB high:0kB reserved_highatomic:0KB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB writepending:0kB present:1048576kB managed:126820kB mlocked:0kB bounce:0kB free_pcp:156kB local_pcp:156kB free_cma:0kB [ 0.484652] lowmem_reserve[]: 0 0 0 [ 0.488538] DMA32: 1*4kB (M) 1*8kB (M) 0*16kB 1*32kB (M) 1*64kB (M) 1*128kB (M) 0*256kB 1*512kB (M) 0*1024kB 1*2048kB (M) 229*4096kB (M) = 940780kB [ 0.503367] Normal: 2*4kB (ME) 0*8kB 2*16kB (UE) 2*32kB (ME) 3*64kB (UME) 0*128kB 2*256kB (ME) 3*512kB (UME) 3*1024kB (UME) 3*2048kB (UME) 28*4096kB (M) = 126248kB [ 0.519722] 0 total pagecache pages [ 0.523579] 0 pages in swap cache [ 0.527264] Free swap = 0kB [ 0.530445] Total swap = 0kB [ 0.533652] 523776 pages RAM [ 0.536833] 0 pages HighMem/MovableOnly [ 0.541090] 256876 pages reserved [ 0.544833] Unable to handle kernel NULL pointer dereference at virtual address 000000000000003c [ 0.554576] Oops [#1] [ 0.557092] Modules linked in: [ 0.560470] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G W 6.2.0-rc4-00538-g6e265f6be0c8 #1 [ 0.570841] Hardware name: Microchip PolarFire-SoC Icicle Kit (DT) [ 0.577664] epc : rest_init+0x44/0xea [ 0.581726] ra : rest_init+0x44/0xea [ 0.585788] epc : ffffffff8077058a ra : ffffffff8077058a sp : ffffffff81203f70 [ 0.593764] gp : ffffffff812e9fe0 tp : ffffffff8120fa80 t0 : 0000000000000000 [ 0.601741] t1 : 0000000000000000 t2 : 0000000000000040 s0 : ffffffff81203f90 [ 0.609718] s1 : fffffffffffffff4 a0 : 0000000000000000 a1 : fffffffffffffff4 [ 0.617695] a2 : 0000000000000000 a3 : 0000000000000000 a4 : 0000000000000000 [ 0.625672] a5 : 0000000000000000 a6 : 0000000001212dec a7 : ffffffffd8bc86ae [ 0.633649] s2 : ffffffff80de0bf0 s3 : ffffffe7c7de7700 s4 : 0000000000000000 [ 0.641626] s5 : ffffffff812eb018 s6 : ffffffff80a00008 s7 : 0000000000000000 [ 0.649603] s8 : 00000000bfb85036 s9 : 0000000000000000 s10: 0000000000000000 [ 0.657580] s11: 0000000000000000 t3 : 0000000000000002 t4 : 0000000000000402 [ 0.665557] t5 : ffffffff812158a8 t6 : 0000000000000000 [ 0.671416] status: 0000000200000120 badaddr: 000000000000003c cause: 000000000000000d [ 0.680157] [<ffffffff80800670>] arch_post_acpi_subsys_init+0x0/0x18 [ 0.687188] [<ffffffff80800de4>] start_kernel+0x72c/0x75c [ 0.693195] ---[ end trace 0000000000000000 ]--- [ 0.698325] Kernel panic - not syncing: Attempted to kill the idle task! [ 0.705729] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---
Signature · 2 lines
Sincerely yours, Mike.
Conor Dooley
Tested-byHi Mike,
Quoted 35 lines · Mike Rapoport
Quoted 15 lines · Conor Dooley
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Quoted 8 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.Signed-off-by Mike Rapoport (IBM) <[email protected]>Acked-by Arnd Bergmann <[email protected]>Acked-by Guo Ren <[email protected]> # cskyQuoted 9 lines
Hmm, so this landed in linux-next today and I bisected a boot failure in my CI to it. However, I am not really sure if it is a real issue worth worrying about as the platform it triggered on is supposed to be using SPARSEMEM, but isn't.
I had thought that my CI was using a config with SPARSEMEM since that became required for riscv defconfig builds to boot in v6.1-rc1, but I must have just forgotten to add it to my $platform_defconfig builds too. However, those $platform_defconfig builds continued booting without SPARSEMEM enabled until today.The issue seems to be that the generic
pfn_valid()does not take into accountpfn_offsetwhen it compares it withmax_mapnr.
Can you please test with the patch below?include/asm-generic/memory_model.h +0
-0
include/asm-generic/memory_model.h static inline int pfn_valid(unsigned long pfn) · line 32
Quoted 8 lines
include/asm-generic/memory_model.h +1
-1@@ -26,7 +26,7 @@ static inline int pfn_valid(unsigned long pfn)extern unsigned long max_mapnr;unsigned long pfn_offset =ARCH_PFN_OFFSET ; - return pfn >= pfn_offset && pfn < max_mapnr; + return pfn >= pfn_offset && (pfn - pfn_offset) < max_mapnr; }#define pfn_valid pfn_valid#endif
Gave that a go, board is booting properly again! Feel free to add a:
Thanks for the prompt fix!
Quoted 8 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.
Quoted 6 lines
Signed-off-by Mike Rapoport (IBM) <[email protected]>Acked-by Arnd Bergmann <[email protected]>Acked-by Guo Ren <[email protected]> # csky---
LGTM with the fixup
Signature · 3 lines
Thanks, David / dhildenb
Quoted 8 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.
With this patch in the tree, sh4 and sh4eb qemu emulations no longer boot. Reverting this patch fixes the problem.
Guenter
---
# bad: [6ba8a227fd19d19779005fb66ad7562608e1df83] Add linux-next specific files for 20230210
# good: [4ec5183ec48656cec489c49f989c508b68b518e3] Linux 6.2-rc7
git bisect start 'HEAD' 'v6.2-rc7'
# good: [94613f0efc69ed41f9229ef5c294db3ec37145da] Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good 94613f0efc69ed41f9229ef5c294db3ec37145da
# good: [19e62c715fe70dae4582c2874ed3e66715d09af6] Merge branch 'rcu/next' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git
git bisect good 19e62c715fe70dae4582c2874ed3e66715d09af6
# good: [5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6] Merge branch 'for-next' of https://gitlab.com/peda-linux/mux.git
git bisect good 5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6
# good: [c349bf6ec83903b20fe570c5609b9a864a64e09c] Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode.git
git bisect good c349bf6ec83903b20fe570c5609b9a864a64e09c
# good: [5a06a9f17454df38f35672be522ff5eb9b4277d2] selftest: add testing unsharing and counting ksm zero page
git bisect good 5a06a9f17454df38f35672be522ff5eb9b4277d2
# bad: [f5d115a7b06e5661ed5218ffa9a2644c4ff1c135] Merge branch 'mm-nonmm-unstable' into mm-everything
git bisect bad f5d115a7b06e5661ed5218ffa9a2644c4ff1c135
# bad: [acb018d6ea0c055381fba7dddaef386ee28f8075] mm/vmalloc.c: allow vread() to read out vm_map_ram areas
git bisect bad acb018d6ea0c055381fba7dddaef386ee28f8075
# good: [1a5d9782ac969dc6e61c6786500b5160603188ea] mm/mmap: remove __vma_adjust()
git bisect good 1a5d9782ac969dc6e61c6786500b5160603188ea
# good: [4b32363697de957dcc890b6245bec3f58903639a] arm: include asm-generic/memory_model.h from page.h rather than memory.h git bisect good 4b32363697de957dcc890b6245bec3f58903639a
# bad: [328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8] mm/migrate: convert isolate_movable_page() to use folios
git bisect bad 328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8
# bad: [b704c765b08cabe82adf76a4d1a74f3688eee410] mm/mempolicy: convert queue_pages_pmd() to queue_folios_pmd()
git bisect bad b704c765b08cabe82adf76a4d1a74f3688eee410
# bad: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation of pfn_valid() for FLATMEM
git bisect bad e5734c8b0edfd2a053a5c256189586a3b1e9f63d
# good: [ad8aecea034c591b9754bc5908da9719853aa7fa] mips: drop definition of pfn_valid() for DISCONTIGMEM
git bisect good ad8aecea034c591b9754bc5908da9719853aa7fa
# first bad commit: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation of pfn_valid() for FLATMEM
Hi Guenter,
Quoted 6 lines · Guenter Roeck
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.
Quoted 6 lines
Quoted 3 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.With this patch in the tree, sh4 and sh4eb qemu emulations no longer boot. Reverting this patch fixes the problem.
Can you please test with only partial revert for arch/sh?
Quoted 1 line
Guenter
Signature · 2 lines
Sincerely yours, Mike.
Quoted 9 lines · Mike Rapoport
Hi Guenter,
Quoted 6 lines · Guenter Roeck
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.
Quoted 9 lines
Quoted 6 lines
Quoted 3 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.With this patch in the tree, sh4 and sh4eb qemu emulations no longer boot. Reverting this patch fixes the problem.
Can you please test with only partial revert for
arch/sh?
Sure, that works as well.
Guenter
Quoted 6 lines · Guenter Roeck
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.
Quoted 6 lines
Quoted 3 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.With this patch in the tree, sh4 and sh4eb qemu emulations no longer boot. Reverting this patch fixes the problem.
This should be a better fix than a revert:
arch/sh/mm/init.c +1 -0
@@ -301,6 +301,7 @@ void __init paging_init(void)*/ max_low_pfn = max_pfn =memblock_end_of_DRAM () >>PAGE_SHIFT ; min_low_pfn = __MEMORY_START >>PAGE_SHIFT ; +set_max_mapnr (max_low_pfn - min_low_pfn);nodes_clear (node_online_map);
Quoted 33 lines
Guenter
---
# bad: [6ba8a227fd19d19779005fb66ad7562608e1df83] Add linux-next specific files for 20230210
# good: [4ec5183ec48656cec489c49f989c508b68b518e3] Linux 6.2-rc7
git bisect start 'HEAD' 'v6.2-rc7'
# good: [94613f0efc69ed41f9229ef5c294db3ec37145da] Merge branch 'master' ofgit://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good94613f0efc69ed41f9229ef5c294db3ec37145da
# good: [19e62c715fe70dae4582c2874ed3e66715d09af6] Merge branch 'rcu/next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git
git bisect good19e62c715fe70dae4582c2874ed3e66715d09af6
# good: [5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6] Merge branch 'for-next' of https://gitlab.com/peda-linux/mux.git
git bisect good5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6
# good: [c349bf6ec83903b20fe570c5609b9a864a64e09c] Merge branch 'for-next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode.git
git bisect goodc349bf6ec83903b20fe570c5609b9a864a64e09c
# good: [5a06a9f17454df38f35672be522ff5eb9b4277d2] selftest: add testing unsharing and counting ksm zero page
git bisect good5a06a9f17454df38f35672be522ff5eb9b4277d2
# bad: [f5d115a7b06e5661ed5218ffa9a2644c4ff1c135] Merge branch 'mm-nonmm-unstable' into mm-everything
git bisect badf5d115a7b06e5661ed5218ffa9a2644c4ff1c135
# bad: [acb018d6ea0c055381fba7dddaef386ee28f8075]mm/vmalloc.c: allowvread()to read outvm_map_ramareas
git bisect badacb018d6ea0c055381fba7dddaef386ee28f8075
# good: [1a5d9782ac969dc6e61c6786500b5160603188ea] mm/mmap: remove__vma_adjust()
git bisect good1a5d9782ac969dc6e61c6786500b5160603188ea
# good: [4b32363697de957dcc890b6245bec3f58903639a] arm: includeasm-generic/memory_model.hfrompage.hrather thanmemory.hgit bisect good4b32363697de957dcc890b6245bec3f58903639a
# bad: [328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8] mm/migrate: convertisolate_movable_page()to use folios
git bisect bad328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8
# bad: [b704c765b08cabe82adf76a4d1a74f3688eee410] mm/mempolicy: convertqueue_pages_pmd()toqueue_folios_pmd()
git bisect badb704c765b08cabe82adf76a4d1a74f3688eee410
# bad: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation ofpfn_valid()for FLATMEM
git bisect bade5734c8b0edfd2a053a5c256189586a3b1e9f63d
# good: [ad8aecea034c591b9754bc5908da9719853aa7fa] mips: drop definition ofpfn_valid()for DISCONTIGMEM
git bisect goodad8aecea034c591b9754bc5908da9719853aa7fa# first bad commit: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation of pfn_valid() for FLATMEM
Signature · 2 lines
Sincerely yours, Mike.
Quoted 21 lines · Mike Rapoport
Quoted 6 lines · Guenter Roeck
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Quoted 6 lines
Quoted 3 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.With this patch in the tree, sh4 and sh4eb qemu emulations no longer boot. Reverting this patch fixes the problem.
This should be a better fix than a revert:
arch/sh/mm/init.c +0
-0
arch/sh/mm/init.c void __init paging_init(void) · line 307
Quoted 7 lines
arch/sh/mm/init.c +1
-0@@ -301,6 +301,7 @@ void __init paging_init(void)*/ max_low_pfn = max_pfn =memblock_end_of_DRAM () >>PAGE_SHIFT ; min_low_pfn = __MEMORY_START >>PAGE_SHIFT ; +set_max_mapnr (max_low_pfn - min_low_pfn);nodes_clear (node_online_map);
Confirmed, this fixes the problem for me.
Thanks,
Guenter
Quoted 35 lines
Quoted 34 lines
Guenter
---
# bad: [6ba8a227fd19d19779005fb66ad7562608e1df83] Add linux-next specific files for 20230210
# good: [4ec5183ec48656cec489c49f989c508b68b518e3] Linux 6.2-rc7
git bisect start 'HEAD' 'v6.2-rc7'
# good: [94613f0efc69ed41f9229ef5c294db3ec37145da] Merge branch 'master' ofgit://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good94613f0efc69ed41f9229ef5c294db3ec37145da
# good: [19e62c715fe70dae4582c2874ed3e66715d09af6] Merge branch 'rcu/next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git
git bisect good19e62c715fe70dae4582c2874ed3e66715d09af6
# good: [5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6] Merge branch 'for-next' of https://gitlab.com/peda-linux/mux.git
git bisect good5d8b7ecef7f4a681b6e5538db59ff26c389c0ab6
# good: [c349bf6ec83903b20fe570c5609b9a864a64e09c] Merge branch 'for-next' ofgit://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode.git
git bisect goodc349bf6ec83903b20fe570c5609b9a864a64e09c
# good: [5a06a9f17454df38f35672be522ff5eb9b4277d2] selftest: add testing unsharing and counting ksm zero page
git bisect good5a06a9f17454df38f35672be522ff5eb9b4277d2
# bad: [f5d115a7b06e5661ed5218ffa9a2644c4ff1c135] Merge branch 'mm-nonmm-unstable' into mm-everything
git bisect badf5d115a7b06e5661ed5218ffa9a2644c4ff1c135
# bad: [acb018d6ea0c055381fba7dddaef386ee28f8075]mm/vmalloc.c: allowvread()to read outvm_map_ramareas
git bisect badacb018d6ea0c055381fba7dddaef386ee28f8075
# good: [1a5d9782ac969dc6e61c6786500b5160603188ea] mm/mmap: remove__vma_adjust()
git bisect good1a5d9782ac969dc6e61c6786500b5160603188ea
# good: [4b32363697de957dcc890b6245bec3f58903639a] arm: includeasm-generic/memory_model.hfrompage.hrather thanmemory.hgit bisect good4b32363697de957dcc890b6245bec3f58903639a
# bad: [328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8] mm/migrate: convertisolate_movable_page()to use folios
git bisect bad328cf3fa6682ce6a4de6f8bb8009c833dc33f3c8
# bad: [b704c765b08cabe82adf76a4d1a74f3688eee410] mm/mempolicy: convertqueue_pages_pmd()toqueue_folios_pmd()
git bisect badb704c765b08cabe82adf76a4d1a74f3688eee410
# bad: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation ofpfn_valid()for FLATMEM
git bisect bade5734c8b0edfd2a053a5c256189586a3b1e9f63d
# good: [ad8aecea034c591b9754bc5908da9719853aa7fa] mips: drop definition ofpfn_valid()for DISCONTIGMEM
git bisect goodad8aecea034c591b9754bc5908da9719853aa7fa# first bad commit: [e5734c8b0edfd2a053a5c256189586a3b1e9f63d] mm, arch: add generic implementation of pfn_valid() for FLATMEM
Andrew,
Quoted 17 lines · Guenter Roeck
Quoted 16 lines · Mike Rapoport
Quoted 6 lines · Guenter Roeck
Quoted 5 lines · Mike Rapoport
From: "Mike Rapoport (IBM)" <[email protected]>
Every architecture that supports FLATMEM memory model defines its own version of
pfn_valid()that essentially compares a pfn tomax_mapnr.Quoted 6 lines
Quoted 3 lines
Use mips/powerpc version implemented as static inline as a generic implementation of
pfn_valid()and drop its per-architecture definitions.With this patch in the tree, sh4 and sh4eb qemu emulations no longer boot. Reverting this patch fixes the problem.
This should be a better fix than a revert:
Quoted 12 lines
arch/sh/mm/init.c void __init paging_init(void) · line 305
Quoted 10 lines
arch/sh/mm/init.c +1
-0@@ -301,6 +301,7 @@ void __init paging_init(void)*/ max_low_pfn = max_pfn =memblock_end_of_DRAM () >>PAGE_SHIFT ; min_low_pfn = __MEMORY_START >>PAGE_SHIFT ; +set_max_mapnr (max_low_pfn - min_low_pfn);nodes_clear (node_online_map);Confirmed, this fixes the problem for me.
What is your preference for this and m68k fix? Fixups on top of mm-stable or v3 of the entire series?
Quoted 2 lines
Thanks,
Guenter
Signature · 2 lines
Sincerely yours, Mike.