* [PATCH] riscv: fix isa validation for virtual [ms]envcfg extension
@ 2024-07-30 7:38 Andy Chiu
2024-07-30 13:14 ` Andrew Jones
0 siblings, 1 reply; 2+ messages in thread
From: Andy Chiu @ 2024-07-30 7:38 UTC (permalink / raw)
To: linux-riscv, palmer
Cc: greentime.hu, guoren, bjorn, debug, Andy Chiu, Paul Walmsley,
Albert Ou, Conor Dooley, Clément Léger, Evan Green,
Andrew Jones, Charlie Jenkins
The exntension RISCV_ISA_EXT_XLINUXENVCFG was accidentally turned off
after introducing ISA validation. The reason is that this extension is
not in the riscv_isa_ext[] list, causing riscv_get_isa_ext_data() fail
to find a corresponding riscv_isa_ext_data pointer. As a result, the
kernel skipped setting resolved_isa bitmask. To fix this, we proceed
with the bit information obtained from source_isa bitmask. This means
that the bit in resolved_isa is set unconditionally, regardless of
whether we find an entry from riscv_isa_ext[] list. This should be safe
as the kernel already parse isa string before writing into source_isa.
Fixes: 625034abd52a ("riscv: add ISA extensions validation callback")
Signed-off-by: Andy Chiu <andy.chiu@sifive.com>
---
arch/riscv/kernel/cpufeature.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/cpufeature.c
index 8f20607adb40..3f91a1f39a50 100644
--- a/arch/riscv/kernel/cpufeature.c
+++ b/arch/riscv/kernel/cpufeature.c
@@ -432,10 +432,7 @@ static void __init riscv_resolve_isa(unsigned long *source_isa,
bitmap_copy(prev_resolved_isa, resolved_isa, RISCV_ISA_EXT_MAX);
for_each_set_bit(bit, source_isa, RISCV_ISA_EXT_MAX) {
ext = riscv_get_isa_ext_data(bit);
- if (!ext)
- continue;
-
- if (ext->validate) {
+ if (ext && ext->validate) {
ret = ext->validate(ext, resolved_isa);
if (ret == -EPROBE_DEFER) {
loop = true;
@@ -447,13 +444,13 @@ static void __init riscv_resolve_isa(unsigned long *source_isa,
}
}
- set_bit(ext->id, resolved_isa);
+ set_bit(bit, resolved_isa);
/* No need to keep it in source isa now that it is enabled */
- clear_bit(ext->id, source_isa);
+ clear_bit(bit, source_isa);
/* Single letter extensions get set in hwcap */
- if (ext->id < RISCV_ISA_EXT_BASE)
- *this_hwcap |= isa2hwcap[ext->id];
+ if (bit < RISCV_ISA_EXT_BASE)
+ *this_hwcap |= isa2hwcap[bit];
}
} while (loop && memcmp(prev_resolved_isa, resolved_isa, sizeof(prev_resolved_isa)));
}
--
2.43.0
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] riscv: fix isa validation for virtual [ms]envcfg extension
2024-07-30 7:38 [PATCH] riscv: fix isa validation for virtual [ms]envcfg extension Andy Chiu
@ 2024-07-30 13:14 ` Andrew Jones
0 siblings, 0 replies; 2+ messages in thread
From: Andrew Jones @ 2024-07-30 13:14 UTC (permalink / raw)
To: Andy Chiu
Cc: linux-riscv, palmer, greentime.hu, guoren, bjorn, debug,
Paul Walmsley, Albert Ou, Conor Dooley, Clément Léger,
Evan Green, Charlie Jenkins, samuel.holland
On Tue, Jul 30, 2024 at 03:38:45PM GMT, Andy Chiu wrote:
> The exntension RISCV_ISA_EXT_XLINUXENVCFG was accidentally turned off
> after introducing ISA validation. The reason is that this extension is
> not in the riscv_isa_ext[] list, causing riscv_get_isa_ext_data() fail
> to find a corresponding riscv_isa_ext_data pointer. As a result, the
> kernel skipped setting resolved_isa bitmask. To fix this, we proceed
> with the bit information obtained from source_isa bitmask. This means
> that the bit in resolved_isa is set unconditionally, regardless of
> whether we find an entry from riscv_isa_ext[] list. This should be safe
> as the kernel already parse isa string before writing into source_isa.
>
> Fixes: 625034abd52a ("riscv: add ISA extensions validation callback")
> Signed-off-by: Andy Chiu <andy.chiu@sifive.com>
> ---
> arch/riscv/kernel/cpufeature.c | 13 +++++--------
> 1 file changed, 5 insertions(+), 8 deletions(-)
>
> diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/cpufeature.c
> index 8f20607adb40..3f91a1f39a50 100644
> --- a/arch/riscv/kernel/cpufeature.c
> +++ b/arch/riscv/kernel/cpufeature.c
> @@ -432,10 +432,7 @@ static void __init riscv_resolve_isa(unsigned long *source_isa,
> bitmap_copy(prev_resolved_isa, resolved_isa, RISCV_ISA_EXT_MAX);
> for_each_set_bit(bit, source_isa, RISCV_ISA_EXT_MAX) {
> ext = riscv_get_isa_ext_data(bit);
> - if (!ext)
> - continue;
> -
> - if (ext->validate) {
> + if (ext && ext->validate) {
> ret = ext->validate(ext, resolved_isa);
> if (ret == -EPROBE_DEFER) {
> loop = true;
> @@ -447,13 +444,13 @@ static void __init riscv_resolve_isa(unsigned long *source_isa,
> }
> }
>
> - set_bit(ext->id, resolved_isa);
> + set_bit(bit, resolved_isa);
> /* No need to keep it in source isa now that it is enabled */
> - clear_bit(ext->id, source_isa);
> + clear_bit(bit, source_isa);
>
> /* Single letter extensions get set in hwcap */
> - if (ext->id < RISCV_ISA_EXT_BASE)
> - *this_hwcap |= isa2hwcap[ext->id];
> + if (bit < RISCV_ISA_EXT_BASE)
> + *this_hwcap |= isa2hwcap[bit];
> }
> } while (loop && memcmp(prev_resolved_isa, resolved_isa, sizeof(prev_resolved_isa)));
> }
> --
> 2.43.0
>
Hi Andy,
This looks like the same solution that Samuel sent which is currently
under discussion[1].
[1] https://lore.kernel.org/all/20240718213011.2600150-1-samuel.holland@sifive.com/
Thanks,
drew
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2024-07-30 13:14 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-07-30 7:38 [PATCH] riscv: fix isa validation for virtual [ms]envcfg extension Andy Chiu
2024-07-30 13:14 ` Andrew Jones
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox