* [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready
@ 2023-06-14 13:19 Woody Zhang
2023-06-21 1:26 ` Song Shuai
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Woody Zhang @ 2023-06-14 13:19 UTC (permalink / raw)
To: Paul Walmsley, Palmer Dabbelt, Albert Ou, Alexandre Ghiti, Conor Dooley
Cc: linux-riscv, linux-kernel, Woody Zhang
The initial memblock metadata is accessed from kernel image mapping. The
regions arrays need to "reallocated" from memblock and accessed through
linear mapping to cover more memblock regions. So the resizing should
not be allowed until linear mapping is ready. Note that there are
memblock allocations when building linear mapping.
This patch is similar to 24cc61d8cb5a ("arm64: memblock: don't permit
memblock resizing until linear mapping is up").
In following log, many memblock regions are reserved before
create_linear_mapping_page_table(). And then it triggered reallocation
of memblock.reserved.regions and memcpy the old array in kernel image
mapping to the new array in linear mapping which caused a page fault.
[ 0.000000] memblock_reserve: [0x00000000bf01f000-0x00000000bf01ffff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf021000-0x00000000bf021fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf023000-0x00000000bf023fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf025000-0x00000000bf025fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf027000-0x00000000bf027fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf029000-0x00000000bf029fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf02b000-0x00000000bf02bfff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf02d000-0x00000000bf02dfff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf02f000-0x00000000bf02ffff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] memblock_reserve: [0x00000000bf030000-0x00000000bf030fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
[ 0.000000] OF: reserved mem: 0x0000000080000000..0x000000008007ffff (512 KiB) map non-reusable mmode_resv0@80000000
[ 0.000000] memblock_reserve: [0x00000000bf000000-0x00000000bf001fed] paging_init+0x19a/0x5ae
[ 0.000000] memblock_phys_alloc_range: 4096 bytes align=0x1000 from=0x0000000000000000 max_addr=0x0000000000000000 alloc_pmd_fixmap+0x14/0x1c
[ 0.000000] memblock_reserve: [0x000000017ffff000-0x000000017fffffff] memblock_alloc_range_nid+0xb8/0x128
[ 0.000000] memblock: reserved is doubled to 256 at [0x000000017fffd000-0x000000017fffe7ff]
[ 0.000000] Unable to handle kernel paging request at virtual address ff600000ffffd000
[ 0.000000] Oops [#1]
[ 0.000000] Modules linked in:
[ 0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 6.4.0-rc1-00011-g99a670b2069c #66
[ 0.000000] Hardware name: riscv-virtio,qemu (DT)
[ 0.000000] epc : __memcpy+0x60/0xf8
[ 0.000000] ra : memblock_double_array+0x192/0x248
[ 0.000000] epc : ffffffff8081d214 ra : ffffffff80a3dfc0 sp : ffffffff81403bd0
[ 0.000000] gp : ffffffff814fbb38 tp : ffffffff8140dac0 t0 : 0000000001600000
[ 0.000000] t1 : 0000000000000000 t2 : 000000008f001000 s0 : ffffffff81403c60
[ 0.000000] s1 : ffffffff80c0bc98 a0 : ff600000ffffd000 a1 : ffffffff80c0bcd8
[ 0.000000] a2 : 0000000000000c00 a3 : ffffffff80c0c8d8 a4 : 0000000080000000
[ 0.000000] a5 : 0000000000080000 a6 : 0000000000000000 a7 : 0000000080200000
[ 0.000000] s2 : ff600000ffffd000 s3 : 0000000000002000 s4 : 0000000000000c00
[ 0.000000] s5 : ffffffff80c0bc60 s6 : ffffffff80c0bcc8 s7 : 0000000000000000
[ 0.000000] s8 : ffffffff814fd0a8 s9 : 000000017fffe7ff s10: 0000000000000000
[ 0.000000] s11: 0000000000001000 t3 : 0000000000001000 t4 : 0000000000000000
[ 0.000000] t5 : 000000008f003000 t6 : ff600000ffffd000
[ 0.000000] status: 0000000200000100 badaddr: ff600000ffffd000 cause: 000000000000000f
[ 0.000000] [<ffffffff8081d214>] __memcpy+0x60/0xf8
[ 0.000000] [<ffffffff80a3e1a2>] memblock_add_range.isra.14+0x12c/0x162
[ 0.000000] [<ffffffff80a3e36a>] memblock_reserve+0x6e/0x8c
[ 0.000000] [<ffffffff80a123fc>] memblock_alloc_range_nid+0xb8/0x128
[ 0.000000] [<ffffffff80a1256a>] memblock_phys_alloc_range+0x5e/0x6a
[ 0.000000] [<ffffffff80a04732>] alloc_pmd_fixmap+0x14/0x1c
[ 0.000000] [<ffffffff80a0475a>] alloc_p4d_fixmap+0xc/0x14
[ 0.000000] [<ffffffff80a04a36>] create_pgd_mapping+0x98/0x17c
[ 0.000000] [<ffffffff80a04e9e>] create_linear_mapping_range.constprop.10+0xe4/0x112
[ 0.000000] [<ffffffff80a05bb8>] paging_init+0x3ec/0x5ae
[ 0.000000] [<ffffffff80a03354>] setup_arch+0xb2/0x576
[ 0.000000] [<ffffffff80a00726>] start_kernel+0x72/0x57e
[ 0.000000] Code: b303 0285 b383 0305 be03 0385 be83 0405 bf03 0485 (b023) 00ef
[ 0.000000] ---[ end trace 0000000000000000 ]---
[ 0.000000] Kernel panic - not syncing: Attempted to kill the idle task!
[ 0.000000] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---
Fixes: 671f9a3e2e24 ("RISC-V: Setup initial page tables in two stages")
Signed-off-by: Woody Zhang <woodylab@foxmail.com>
---
Changes in v3:
- Add panic log for the bug case.
- Move memblock_allow_resize to paging_init per suggestions from Alex and
Song Shuai.
Changes in v2:
- Refine commit log
---
arch/riscv/mm/init.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
index 4fa420faa780..1306149aad57 100644
--- a/arch/riscv/mm/init.c
+++ b/arch/riscv/mm/init.c
@@ -267,7 +267,6 @@ static void __init setup_bootmem(void)
dma_contiguous_reserve(dma32_phys_limit);
if (IS_ENABLED(CONFIG_64BIT))
hugetlb_cma_reserve(PUD_SHIFT - PAGE_SHIFT);
- memblock_allow_resize();
}
#ifdef CONFIG_MMU
@@ -1370,6 +1369,9 @@ void __init paging_init(void)
{
setup_bootmem();
setup_vm_final();
+
+ /* Depend on that Linear Mapping is ready */
+ memblock_allow_resize();
}
void __init misc_mem_init(void)
--
2.39.2
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready
2023-06-14 13:19 [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready Woody Zhang
@ 2023-06-21 1:26 ` Song Shuai
2023-07-05 23:38 ` Palmer Dabbelt
2023-07-05 23:50 ` patchwork-bot+linux-riscv
2 siblings, 0 replies; 4+ messages in thread
From: Song Shuai @ 2023-06-21 1:26 UTC (permalink / raw)
To: Woody Zhang, Paul Walmsley, Palmer Dabbelt, Albert Ou,
Alexandre Ghiti, Conor Dooley
Cc: linux-riscv, linux-kernel, songshuaishuai
在 2023/6/14 21:19, Woody Zhang 写道:
> The initial memblock metadata is accessed from kernel image mapping. The
> regions arrays need to "reallocated" from memblock and accessed through
> linear mapping to cover more memblock regions. So the resizing should
> not be allowed until linear mapping is ready. Note that there are
> memblock allocations when building linear mapping.
>
> This patch is similar to 24cc61d8cb5a ("arm64: memblock: don't permit
> memblock resizing until linear mapping is up").
>
> In following log, many memblock regions are reserved before
> create_linear_mapping_page_table(). And then it triggered reallocation
> of memblock.reserved.regions and memcpy the old array in kernel image
> mapping to the new array in linear mapping which caused a page fault.
>
> [ 0.000000] memblock_reserve: [0x00000000bf01f000-0x00000000bf01ffff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf021000-0x00000000bf021fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf023000-0x00000000bf023fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf025000-0x00000000bf025fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf027000-0x00000000bf027fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf029000-0x00000000bf029fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf02b000-0x00000000bf02bfff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf02d000-0x00000000bf02dfff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf02f000-0x00000000bf02ffff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] memblock_reserve: [0x00000000bf030000-0x00000000bf030fff] early_init_fdt_scan_reserved_mem+0x28c/0x2c6
> [ 0.000000] OF: reserved mem: 0x0000000080000000..0x000000008007ffff (512 KiB) map non-reusable mmode_resv0@80000000
> [ 0.000000] memblock_reserve: [0x00000000bf000000-0x00000000bf001fed] paging_init+0x19a/0x5ae
> [ 0.000000] memblock_phys_alloc_range: 4096 bytes align=0x1000 from=0x0000000000000000 max_addr=0x0000000000000000 alloc_pmd_fixmap+0x14/0x1c
> [ 0.000000] memblock_reserve: [0x000000017ffff000-0x000000017fffffff] memblock_alloc_range_nid+0xb8/0x128
> [ 0.000000] memblock: reserved is doubled to 256 at [0x000000017fffd000-0x000000017fffe7ff]
> [ 0.000000] Unable to handle kernel paging request at virtual address ff600000ffffd000
> [ 0.000000] Oops [#1]
> [ 0.000000] Modules linked in:
> [ 0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 6.4.0-rc1-00011-g99a670b2069c #66
> [ 0.000000] Hardware name: riscv-virtio,qemu (DT)
> [ 0.000000] epc : __memcpy+0x60/0xf8
> [ 0.000000] ra : memblock_double_array+0x192/0x248
> [ 0.000000] epc : ffffffff8081d214 ra : ffffffff80a3dfc0 sp : ffffffff81403bd0
> [ 0.000000] gp : ffffffff814fbb38 tp : ffffffff8140dac0 t0 : 0000000001600000
> [ 0.000000] t1 : 0000000000000000 t2 : 000000008f001000 s0 : ffffffff81403c60
> [ 0.000000] s1 : ffffffff80c0bc98 a0 : ff600000ffffd000 a1 : ffffffff80c0bcd8
> [ 0.000000] a2 : 0000000000000c00 a3 : ffffffff80c0c8d8 a4 : 0000000080000000
> [ 0.000000] a5 : 0000000000080000 a6 : 0000000000000000 a7 : 0000000080200000
> [ 0.000000] s2 : ff600000ffffd000 s3 : 0000000000002000 s4 : 0000000000000c00
> [ 0.000000] s5 : ffffffff80c0bc60 s6 : ffffffff80c0bcc8 s7 : 0000000000000000
> [ 0.000000] s8 : ffffffff814fd0a8 s9 : 000000017fffe7ff s10: 0000000000000000
> [ 0.000000] s11: 0000000000001000 t3 : 0000000000001000 t4 : 0000000000000000
> [ 0.000000] t5 : 000000008f003000 t6 : ff600000ffffd000
> [ 0.000000] status: 0000000200000100 badaddr: ff600000ffffd000 cause: 000000000000000f
> [ 0.000000] [<ffffffff8081d214>] __memcpy+0x60/0xf8
> [ 0.000000] [<ffffffff80a3e1a2>] memblock_add_range.isra.14+0x12c/0x162
> [ 0.000000] [<ffffffff80a3e36a>] memblock_reserve+0x6e/0x8c
> [ 0.000000] [<ffffffff80a123fc>] memblock_alloc_range_nid+0xb8/0x128
> [ 0.000000] [<ffffffff80a1256a>] memblock_phys_alloc_range+0x5e/0x6a
> [ 0.000000] [<ffffffff80a04732>] alloc_pmd_fixmap+0x14/0x1c
> [ 0.000000] [<ffffffff80a0475a>] alloc_p4d_fixmap+0xc/0x14
> [ 0.000000] [<ffffffff80a04a36>] create_pgd_mapping+0x98/0x17c
> [ 0.000000] [<ffffffff80a04e9e>] create_linear_mapping_range.constprop.10+0xe4/0x112
> [ 0.000000] [<ffffffff80a05bb8>] paging_init+0x3ec/0x5ae
> [ 0.000000] [<ffffffff80a03354>] setup_arch+0xb2/0x576
> [ 0.000000] [<ffffffff80a00726>] start_kernel+0x72/0x57e
> [ 0.000000] Code: b303 0285 b383 0305 be03 0385 be83 0405 bf03 0485 (b023) 00ef
> [ 0.000000] ---[ end trace 0000000000000000 ]---
> [ 0.000000] Kernel panic - not syncing: Attempted to kill the idle task!
> [ 0.000000] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---
>
> Fixes: 671f9a3e2e24 ("RISC-V: Setup initial page tables in two stages")
> Signed-off-by: Woody Zhang <woodylab@foxmail.com>
> ---
> Changes in v3:
> - Add panic log for the bug case.
> - Move memblock_allow_resize to paging_init per suggestions from Alex and
> Song Shuai.
>
> Changes in v2:
> - Refine commit log
> ---
> arch/riscv/mm/init.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
> index 4fa420faa780..1306149aad57 100644
> --- a/arch/riscv/mm/init.c
> +++ b/arch/riscv/mm/init.c
> @@ -267,7 +267,6 @@ static void __init setup_bootmem(void)
> dma_contiguous_reserve(dma32_phys_limit);
> if (IS_ENABLED(CONFIG_64BIT))
> hugetlb_cma_reserve(PUD_SHIFT - PAGE_SHIFT);
> - memblock_allow_resize();
> }
>
> #ifdef CONFIG_MMU
> @@ -1370,6 +1369,9 @@ void __init paging_init(void)
> {
> setup_bootmem();
> setup_vm_final();
> +
> + /* Depend on that Linear Mapping is ready */
> + memblock_allow_resize();
> }
>
> void __init misc_mem_init(void)
Tested-by: Song Shuai <songshuaishuai@tinylab.org>
--
Thanks
Song Shuai
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready
2023-06-14 13:19 [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready Woody Zhang
2023-06-21 1:26 ` Song Shuai
@ 2023-07-05 23:38 ` Palmer Dabbelt
2023-07-05 23:50 ` patchwork-bot+linux-riscv
2 siblings, 0 replies; 4+ messages in thread
From: Palmer Dabbelt @ 2023-07-05 23:38 UTC (permalink / raw)
To: Paul Walmsley, Palmer Dabbelt, Albert Ou, Alexandre Ghiti,
Conor Dooley, Woody Zhang
Cc: linux-riscv, linux-kernel
On Wed, 14 Jun 2023 21:19:07 +0800, Woody Zhang wrote:
> The initial memblock metadata is accessed from kernel image mapping. The
> regions arrays need to "reallocated" from memblock and accessed through
> linear mapping to cover more memblock regions. So the resizing should
> not be allowed until linear mapping is ready. Note that there are
> memblock allocations when building linear mapping.
>
> This patch is similar to 24cc61d8cb5a ("arm64: memblock: don't permit
> memblock resizing until linear mapping is up").
>
> [...]
Applied, thanks!
[1/1] riscv: move memblock_allow_resize() after linear mapping is ready
https://git.kernel.org/palmer/c/85fadc0d0411
Best regards,
--
Palmer Dabbelt <palmer@rivosinc.com>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready
2023-06-14 13:19 [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready Woody Zhang
2023-06-21 1:26 ` Song Shuai
2023-07-05 23:38 ` Palmer Dabbelt
@ 2023-07-05 23:50 ` patchwork-bot+linux-riscv
2 siblings, 0 replies; 4+ messages in thread
From: patchwork-bot+linux-riscv @ 2023-07-05 23:50 UTC (permalink / raw)
To: Woody Zhang
Cc: linux-riscv, paul.walmsley, palmer, aou, alexghiti, conor.dooley,
linux-kernel
Hello:
This patch was applied to riscv/linux.git (for-next)
by Palmer Dabbelt <palmer@rivosinc.com>:
On Wed, 14 Jun 2023 21:19:07 +0800 you wrote:
> The initial memblock metadata is accessed from kernel image mapping. The
> regions arrays need to "reallocated" from memblock and accessed through
> linear mapping to cover more memblock regions. So the resizing should
> not be allowed until linear mapping is ready. Note that there are
> memblock allocations when building linear mapping.
>
> This patch is similar to 24cc61d8cb5a ("arm64: memblock: don't permit
> memblock resizing until linear mapping is up").
>
> [...]
Here is the summary with links:
- [v3] riscv: move memblock_allow_resize() after linear mapping is ready
https://git.kernel.org/riscv/c/85fadc0d0411
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2023-07-05 23:50 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-06-14 13:19 [PATCH v3] riscv: move memblock_allow_resize() after linear mapping is ready Woody Zhang
2023-06-21 1:26 ` Song Shuai
2023-07-05 23:38 ` Palmer Dabbelt
2023-07-05 23:50 ` patchwork-bot+linux-riscv
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).