[PATCH] Documentation: riscv: Document the sv57 VM layout
Alexandre Ghiti
alex at ghiti.fr
Fri Nov 18 08:07:40 PST 2022
Hi Björn,
On 10/31/22 19:02, Björn Töpel wrote:
> From: Björn Töpel <bjorn at rivosinc.com>
>
> RISC-V has been supporting the "sv57" address translation mode for a
> while, but is has not been added to the VM layout documentation. Let
> us fix that.
>
> Signed-off-by: Björn Töpel <bjorn at rivosinc.com>
> ---
> Documentation/riscv/vm-layout.rst | 36 +++++++++++++++++++++++++++++++
> 1 file changed, 36 insertions(+)
>
> diff --git a/Documentation/riscv/vm-layout.rst b/Documentation/riscv/vm-layout.rst
> index 5b36e45fef60..35f76798b6e4 100644
> --- a/Documentation/riscv/vm-layout.rst
> +++ b/Documentation/riscv/vm-layout.rst
> @@ -97,3 +97,39 @@ RISC-V Linux Kernel SV48
> ffffffff00000000 | -4 GB | ffffffff7fffffff | 2 GB | modules, BPF
> ffffffff80000000 | -2 GB | ffffffffffffffff | 2 GB | kernel
> __________________|____________|__________________|_________|____________________________________________________________
> +
> +
> +RISC-V Linux Kernel SV57
> +------------------------
> +
> +::
> +
> + ========================================================================================================================
> + Start addr | Offset | End addr | Size | VM area description
> + ========================================================================================================================
> + | | | |
> + 0000000000000000 | 0 | 00ffffffffffffff | 64 PB | user-space virtual memory, different per mm
> + __________________|____________|__________________|_________|___________________________________________________________
> + | | | |
> + 0100000000000000 | +64 PB | feffffffffffffff | ~16K PB | ... huge, almost 64 bits wide hole of non-canonical
Very very nit but "+64" is not correctly aligned :)
> + | | | | virtual memory addresses up to the -64 PB
> + | | | | starting offset of kernel mappings.
> + __________________|____________|__________________|_________|___________________________________________________________
> + |
> + | Kernel-space virtual memory, shared between all processes:
> + ____________________________________________________________|___________________________________________________________
> + | | | |
> + ff1bfffffee00000 | -57 PB | ff1bfffffeffffff | 2 MB | fixmap
> + ff1bffffff000000 | -57 PB | ff1bffffffffffff | 16 MB | PCI io
> + ff1c000000000000 | -57 PB | ff1fffffffffffff | 1 PB | vmemmap
> + ff20000000000000 | -56 PB | ff5fffffffffffff | 16 PB | vmalloc/ioremap space
> + ff60000000000000 | -40 PB | ffdffffeffffffff | 32 PB | direct mapping of all physical memory
> + ffdfffff00000000 | - 8 PB | fffffffeffffffff | 8 PB | kasan
To me, the kasan start address is wrong, an sv57 kernel produces the
following output for me:
[ 0.000000] fixmap : 0xff1bfffffee00000 - 0xff1bffffff000000
(2048 kB)
[ 0.000000] pci io : 0xff1bffffff000000 - 0xff1c000000000000
( 16 MB)
[ 0.000000] vmemmap : 0xff1c000000000000 - 0xff20000000000000
(1024 TB)
[ 0.000000] vmalloc : 0xff20000000000000 - 0xff60000000000000
(16384 TB)
[ 0.000000] modules : 0xffffffff01e2a000 - 0xffffffff80000000
(2017 MB)
[ 0.000000] lowmem : 0xff60000000000000 - 0xff60000100000000
(4096 MB)
[ 0.000000] kasan : 0xffdf000000000000 - 0xffffffff00000000
(8447 TB)
[ 0.000000] kernel : 0xffffffff80000000 - 0xffffffffffffffff
(2047 MB)
Because by definition, KASAN_SHADOW_START is aligned on PGDIR_SIZE
(because it is easier to map this way).
And again very very nit: there is a space between the '-' and '8' :)
> + __________________|____________|__________________|_________|____________________________________________________________
> + |
> + | Identical layout to the 39-bit one from here on:
> + ____________________________________________________________|____________________________________________________________
> + | | | |
> + ffffffff00000000 | -4 GB | ffffffff7fffffff | 2 GB | modules, BPF
> + ffffffff80000000 | -2 GB | ffffffffffffffff | 2 GB | kernel
> + __________________|____________|__________________|_________|____________________________________________________________
Thanks,
Alex
More information about the linux-riscv
mailing list