Generally, riscv ISA extensions are fixed for any specific hardware
platform, so a hart's features won't change after booting. This
chacteristic makes it straightforward to use a static branch to check
if a specific ISA extension is supported or not to optimize
performance.
However, some ISA extensions such as SVPBMT and ZICBOM are handled
via. the alternative sequences.
Basically, for ease of maintenance, we prefer to use static branches
in C code, but recently, Samuel found that the static branch usage in
cpu_relax() breaks building with CONFIG_CC_OPTIMIZE_FOR_SIZE[1]. As
Samuel pointed out, "Having a static branch in cpu_relax() is
problematic because that function is widely inlined, including in some
quite complex functions like in the VDSO. A quick measurement shows
this static branch is responsible by itself for around 40% of the jump
table."
Samuel's findings pointed out one of a few downsides of static branches
usage in C code to handle ISA extensions detected at boot time:
static branch's metadata in the __jump_table section, which is not
discarded after ISA extensions are finalized, wastes some space.
I want to try to solve the issue for all possible dynamic handling of
ISA extensions at boot time. Inspired by Mark[2], this patch introduces
riscv_has_extension_*() helpers, which work like static branches but
are patched using alternatives, thus the metadata can be freed after
patching.
Link: https://lore.kernel.org/linux-riscv/[email protected]/ [1]
Link: https://lore.kernel.org/linux-arm-kernel/[email protected]/ [2]
Signed-off-by: Jisheng Zhang <[email protected]>
Reviewed-by: Andrew Jones <[email protected]>
---
arch/riscv/include/asm/hwcap.h | 37 ++++++++++++++++++++++++++++++++++
1 file changed, 37 insertions(+)
diff --git a/arch/riscv/include/asm/hwcap.h b/arch/riscv/include/asm/hwcap.h
index 09a7767723f6..1767a9ce1a04 100644
--- a/arch/riscv/include/asm/hwcap.h
+++ b/arch/riscv/include/asm/hwcap.h
@@ -8,6 +8,7 @@
#ifndef _ASM_RISCV_HWCAP_H
#define _ASM_RISCV_HWCAP_H
+#include <asm/alternative-macros.h>
#include <asm/errno.h>
#include <linux/bits.h>
#include <uapi/asm/hwcap.h>
@@ -97,6 +98,42 @@ static __always_inline int riscv_isa_ext2key(int num)
}
}
+static __always_inline bool
+riscv_has_extension_likely(const unsigned long ext)
+{
+ compiletime_assert(ext < RISCV_ISA_EXT_MAX,
+ "ext must be < RISCV_ISA_EXT_MAX");
+
+ asm_volatile_goto(
+ ALTERNATIVE("j %l[l_no]", "nop", 0, %[ext], 1)
+ :
+ : [ext] "i" (ext)
+ :
+ : l_no);
+
+ return true;
+l_no:
+ return false;
+}
+
+static __always_inline bool
+riscv_has_extension_unlikely(const unsigned long ext)
+{
+ compiletime_assert(ext < RISCV_ISA_EXT_MAX,
+ "ext must be < RISCV_ISA_EXT_MAX");
+
+ asm_volatile_goto(
+ ALTERNATIVE("nop", "j %l[l_yes]", 0, %[ext], 1)
+ :
+ : [ext] "i" (ext)
+ :
+ : l_yes);
+
+ return false;
+l_yes:
+ return true;
+}
+
unsigned long riscv_isa_extension_base(const unsigned long *isa_bitmap);
#define riscv_isa_extension_mask(ext) BIT_MASK(RISCV_ISA_EXT_##ext)
--
2.38.1
Hey Jisheng,
On Sun, Jan 15, 2023 at 11:49:46PM +0800, Jisheng Zhang wrote:
> Generally, riscv ISA extensions are fixed for any specific hardware
> platform, so a hart's features won't change after booting. This
> chacteristic makes it straightforward to use a static branch to check
> if a specific ISA extension is supported or not to optimize
> performance.
>
> However, some ISA extensions such as SVPBMT and ZICBOM are handled
> via. the alternative sequences.
>
> Basically, for ease of maintenance, we prefer to use static branches
> in C code, but recently, Samuel found that the static branch usage in
> cpu_relax() breaks building with CONFIG_CC_OPTIMIZE_FOR_SIZE[1]. As
> Samuel pointed out, "Having a static branch in cpu_relax() is
> problematic because that function is widely inlined, including in some
> quite complex functions like in the VDSO. A quick measurement shows
> this static branch is responsible by itself for around 40% of the jump
> table."
>
> Samuel's findings pointed out one of a few downsides of static branches
> usage in C code to handle ISA extensions detected at boot time:
> static branch's metadata in the __jump_table section, which is not
> discarded after ISA extensions are finalized, wastes some space.
>
> I want to try to solve the issue for all possible dynamic handling of
> ISA extensions at boot time. Inspired by Mark[2], this patch introduces
> riscv_has_extension_*() helpers, which work like static branches but
> are patched using alternatives, thus the metadata can be freed after
> patching.
>
> Link: https://lore.kernel.org/linux-riscv/[email protected]/ [1]
> Link: https://lore.kernel.org/linux-arm-kernel/[email protected]/ [2]
> Signed-off-by: Jisheng Zhang <[email protected]>
> Reviewed-by: Andrew Jones <[email protected]>
It'd be great, if, in the future, you would hold off on sending new
versions of patchsets where the previous version is still being
discussed [3].
~3 days between versions is not very much, especially when that includes
a weekend!
I know you replied there earlier today with your opinion, but please
give people a chance to read and respond, before resubmitting, so as not
to split discussion between several threads.
3 - https://lore.kernel.org/linux-riscv/2398293.3Lj2Plt8kZ@diego/
Thanks,
Conor.
On Sun, Jan 15, 2023 at 11:49:46PM +0800, Jisheng Zhang wrote:
> Generally, riscv ISA extensions are fixed for any specific hardware
> platform, so a hart's features won't change after booting. This
> chacteristic makes it straightforward to use a static branch to check
> if a specific ISA extension is supported or not to optimize
> performance.
>
> However, some ISA extensions such as SVPBMT and ZICBOM are handled
> via. the alternative sequences.
>
> Basically, for ease of maintenance, we prefer to use static branches
> in C code, but recently, Samuel found that the static branch usage in
> cpu_relax() breaks building with CONFIG_CC_OPTIMIZE_FOR_SIZE[1]. As
> Samuel pointed out, "Having a static branch in cpu_relax() is
> problematic because that function is widely inlined, including in some
> quite complex functions like in the VDSO. A quick measurement shows
> this static branch is responsible by itself for around 40% of the jump
> table."
>
> Samuel's findings pointed out one of a few downsides of static branches
> usage in C code to handle ISA extensions detected at boot time:
> static branch's metadata in the __jump_table section, which is not
> discarded after ISA extensions are finalized, wastes some space.
>
> I want to try to solve the issue for all possible dynamic handling of
> ISA extensions at boot time. Inspired by Mark[2], this patch introduces
> riscv_has_extension_*() helpers, which work like static branches but
> are patched using alternatives, thus the metadata can be freed after
> patching.
>
> Link: https://lore.kernel.org/linux-riscv/[email protected]/ [1]
> Link: https://lore.kernel.org/linux-arm-kernel/[email protected]/ [2]
> Signed-off-by: Jisheng Zhang <[email protected]>
> Reviewed-by: Andrew Jones <[email protected]>
I think I just didn't leave an R-b last time because of the ongoing
discussion & the functions naming.
Given the conversation I attempted to summarise at [1], I'm satisfied
that we're not just introducing something that'd conflict with trying
to use this mechanism for non-extension capabilities:
Acked-by: Conor Dooley <[email protected]>
1 - https://lore.kernel.org/linux-riscv/Y8hqptFcUgjhns4F@spud/
> ---
> arch/riscv/include/asm/hwcap.h | 37 ++++++++++++++++++++++++++++++++++
> 1 file changed, 37 insertions(+)
>
> diff --git a/arch/riscv/include/asm/hwcap.h b/arch/riscv/include/asm/hwcap.h
> index 09a7767723f6..1767a9ce1a04 100644
> --- a/arch/riscv/include/asm/hwcap.h
> +++ b/arch/riscv/include/asm/hwcap.h
> @@ -8,6 +8,7 @@
> #ifndef _ASM_RISCV_HWCAP_H
> #define _ASM_RISCV_HWCAP_H
>
> +#include <asm/alternative-macros.h>
> #include <asm/errno.h>
> #include <linux/bits.h>
> #include <uapi/asm/hwcap.h>
> @@ -97,6 +98,42 @@ static __always_inline int riscv_isa_ext2key(int num)
> }
> }
>
> +static __always_inline bool
> +riscv_has_extension_likely(const unsigned long ext)
> +{
> + compiletime_assert(ext < RISCV_ISA_EXT_MAX,
> + "ext must be < RISCV_ISA_EXT_MAX");
> +
> + asm_volatile_goto(
> + ALTERNATIVE("j %l[l_no]", "nop", 0, %[ext], 1)
> + :
> + : [ext] "i" (ext)
> + :
> + : l_no);
> +
> + return true;
> +l_no:
> + return false;
> +}
> +
> +static __always_inline bool
> +riscv_has_extension_unlikely(const unsigned long ext)
> +{
> + compiletime_assert(ext < RISCV_ISA_EXT_MAX,
> + "ext must be < RISCV_ISA_EXT_MAX");
> +
> + asm_volatile_goto(
> + ALTERNATIVE("nop", "j %l[l_yes]", 0, %[ext], 1)
> + :
> + : [ext] "i" (ext)
> + :
> + : l_yes);
> +
> + return false;
> +l_yes:
> + return true;
> +}
> +
> unsigned long riscv_isa_extension_base(const unsigned long *isa_bitmap);
>
> #define riscv_isa_extension_mask(ext) BIT_MASK(RISCV_ISA_EXT_##ext)
> --
> 2.38.1
>
>
> _______________________________________________
> linux-riscv mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-riscv