Impveovements for random.h/archrandom.h
Subsystems (7)
v2 cover Impveovements for random.h/archrandom.h
This is a resend of a series from Richard Henderson last posted back in November:
https://lore.kernel.org/linux-arm-kernel/[email protected]/
Back then Borislav said they looked good and asked if he should take them through the tip tree but things seem to have got lost since then.
Original cover letter:
During patch review for an addition of archrandom.h for arm64, it was suggeted that the arch_random_get_* functions should be marked __must_check. Which does sound like a good idea, since the by-reference integer output may be uninitialized when the boolean result is false.
In addition, it turns out that arch_has_random() and arch_has_random_seed() are not used, and not easy to support for arm64. Rather than cobble something together that would not be testable, remove the interfaces against some future accidental use.
In addition, I noticed a few other minor inconsistencies between the different architectures, e.g. powerpc isn't using bool.
Change since v1:
* Remove arch_has_random, arch_has_random_seed.
Richard Henderson (10):
x86: Remove arch_has_random, arch_has_random_seed
powerpc: Remove arch_has_random, arch_has_random_seed
s390: Remove arch_has_random, arch_has_random_seed linux/random.h: Remove arch_has_random, arch_has_random_seed linux/random.h: Use false with bool
linux/random.h: Mark CONFIG_ARCH_RANDOM functions __must_check
x86: Mark archrandom.h functions __must_check
powerpc: Use bool in archrandom.h
powerpc: Mark archrandom.h functions __must_check
s390x: Mark archrandom.h functions __must_check
arch/powerpc/include/asm/archrandom.h | 27 +++++++++----------------- arch/s390/include/asm/archrandom.h | 20 ++++--------------- arch/x86/include/asm/archrandom.h | 28 ++++++++++++--------------- include/linux/random.h | 24 ++++++++--------------- 4 files changed, 33 insertions(+), 66 deletions(-)
Signature · 1 line
2.20.1
PATCH v2 1/10 x86: Remove arch_has_random, arch_has_random_seed
From: Richard Henderson <[email protected]>
Use the expansion of these macros directly in arch_get_random_*.
These symbols are currently part of the generic archrandom.h interface, but are currently unused and can be removed.
--- arch/x86/include/asm/archrandom.h | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-)
arch/x86/include/asm/archrandom.h +4 -8
@@ -73,10 +73,6 @@ static inline bool rdseed_int(unsigned int *v)return ok; } -/* Conditional execution based on CPU type */ -#define arch_has_random () static_cpu_has(X86_FEATURE_RDRAND) -#define arch_has_random_seed () static_cpu_has(X86_FEATURE_RDSEED) -/* * These are the generic interfaces; they must not be declared if the * stubs in <linux/random.h> are to be invoked,
@@ -86,22 +82,22 @@ static inline bool rdseed_int(unsigned int *v)static inline bool arch_get_random_long (unsigned long *v) { -return arch_has_random () ?rdrand_long (v) : false; +return static_cpu_has (X86_FEATURE_RDRAND ) ?rdrand_long (v) : false; }static inline bool arch_get_random_int (unsigned int *v) { -return arch_has_random () ?rdrand_int (v) : false; +return static_cpu_has (X86_FEATURE_RDRAND ) ?rdrand_int (v) : false; }static inline bool arch_get_random_seed_long (unsigned long *v) { -return arch_has_random_seed () ?rdseed_long (v) : false; +return static_cpu_has (X86_FEATURE_RDSEED ) ?rdseed_long (v) : false; }static inline bool arch_get_random_seed_int (unsigned int *v) { -return arch_has_random_seed () ?rdseed_int (v) : false; +return static_cpu_has (X86_FEATURE_RDSEED ) ?rdseed_int (v) : false; }extern void x86_init_rdrand (struct cpuinfo_x86 *c);
Signature · 1 line
2.20.1
PATCH v2 2/10 powerpc: Remove arch_has_random, arch_has_random_seed
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
These symbols are currently part of the generic archrandom.h interface, but are currently unused and can be removed.
--- arch/powerpc/include/asm/archrandom.h | 10 ---------- 1 file changed, 10 deletions(-)
arch/powerpc/include/asm/archrandom.h +0 -10
@@ -34,16 +34,6 @@ static inline int arch_get_random_seed_int(unsigned int *v) return rc; } - -static inline int arch_has_random (void ) -{ -return 0 ; -} - -static inline int arch_has_random_seed (void ) -{ -return !!ppc_md.get_random_seed; -}#endif /* CONFIG_ARCH_RANDOM */ #ifdef CONFIG_PPC_POWERNV
Signature · 1 line
2.20.1
PATCH v2 3/10 s390: Remove arch_has_random, arch_has_random_seed
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
These symbols are currently part of the generic archrandom.h interface, but are currently unused and can be removed.
--- arch/s390/include/asm/archrandom.h | 12 ------------ 1 file changed, 12 deletions(-)
arch/s390/include/asm/archrandom.h +0 -12
@@ -21,18 +21,6 @@ extern atomic64_t s390_arch_random_counter;bool s390_arch_random_generate (u8 *buf,unsigned int nbytes); -static inline bool arch_has_random (void ) -{ -return false; -} - -static inline bool arch_has_random_seed (void ) -{ -if (static_branch_likely (&s390_arch_random_available)) -return true; -return false; -} -static inline bool arch_get_random_long (unsigned long *v) {return false;
Signature · 1 line
2.20.1
PATCH v2 4/10 linux/random.h: Remove arch_has_random, arch_has_random_seed
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
The arm64 version of archrandom.h will need to be able to test for support and read the random number without preemption, so a separate query predicate is not practical.
Since this part of the generic interface is unused, remove it.
--- include/linux/random.h | 8 -------- 1 file changed, 8 deletions(-)
include/linux/random.h +0 -8
@@ -175,10 +175,6 @@ static inline bool arch_get_random_int(unsigned int *v) {return 0 ; } -static inline bool arch_has_random (void ) -{ -return 0 ; -}static inline bool arch_get_random_seed_long (unsigned long *v) {return 0 ;
@@ -187,10 +183,6 @@ static inline bool arch_get_random_seed_int(unsigned int *v) {return 0 ; } -static inline bool arch_has_random_seed (void ) -{ -return 0 ; -}#endif /* Pseudo random number generator from numerical recipes. */
Signature · 1 line
2.20.1
PATCH v2 5/10 linux/random.h: Use false with bool
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
Keep the generic fallback versions in sync with the other architecture specific implementations and use the proper name for false.
--- include/linux/random.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
include/linux/random.h +4 -4
@@ -169,19 +169,19 @@ static inline void prandom_seed_state(struct rnd_state *state, u64 seed)#else static inline bool arch_get_random_long (unsigned long *v) { -return 0 ; +return false; }static inline bool arch_get_random_int (unsigned int *v) { -return 0 ; +return false; }static inline bool arch_get_random_seed_long (unsigned long *v) { -return 0 ; +return false; }static inline bool arch_get_random_seed_int (unsigned int *v) { -return 0 ; +return false; }#endif
Signature · 1 line
2.20.1
PATCH v2 6/10 linux/random.h: Mark CONFIG_ARCH_RANDOM functions __must_check
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
We must not use the pointer output without validating the success of the random read.
--- include/linux/random.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
include/linux/random.h +4 -4
@@ -167,19 +167,19 @@ static inline void prandom_seed_state(struct rnd_state *state, u64 seed)#ifdef CONFIG_ARCH_RANDOM # include <asm/archrandom.h> #else -static inline bool arch_get_random_long (unsigned long *v) +static inline bool __must_check arch_get_random_long (unsigned long *v) {return false; } -static inline bool arch_get_random_int (unsigned int *v) +static inline bool __must_check arch_get_random_int (unsigned int *v) {return false; } -static inline bool arch_get_random_seed_long (unsigned long *v) +static inline bool __must_check arch_get_random_seed_long (unsigned long *v) {return false; } -static inline bool arch_get_random_seed_int (unsigned int *v) +static inline bool __must_check arch_get_random_seed_int (unsigned int *v) {return false; }
Signature · 1 line
2.20.1
PATCH v2 7/10 x86: Mark archrandom.h functions __must_check
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
We must not use the pointer output without validating the success of the random read.
--- arch/x86/include/asm/archrandom.h | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)
arch/x86/include/asm/archrandom.h +8 -8
@@ -27,7 +27,7 @@/* Unconditional execution of RDRAND and RDSEED */ -static inline bool rdrand_long (unsigned long *v) +static inline bool __must_check rdrand_long (unsigned long *v) {bool ok;unsigned int retry =RDRAND_RETRY_LOOPS ;
@@ -41,7 +41,7 @@ static inline bool rdrand_long(unsigned long *v)return false; } -static inline bool rdrand_int (unsigned int *v) +static inline bool __must_check rdrand_int (unsigned int *v) {bool ok;unsigned int retry =RDRAND_RETRY_LOOPS ;
@@ -55,7 +55,7 @@ static inline bool rdrand_int(unsigned int *v)return false; } -static inline bool rdseed_long (unsigned long *v) +static inline bool __must_check rdseed_long (unsigned long *v) {bool ok; asm volatile(RDSEED_LONG
@@ -64,7 +64,7 @@ static inline bool rdseed_long(unsigned long *v)return ok; } -static inline bool rdseed_int (unsigned int *v) +static inline bool __must_check rdseed_int (unsigned int *v) {bool ok; asm volatile(RDSEED_INT
@@ -80,22 +80,22 @@ static inline bool rdseed_int(unsigned int *v)*/ #ifdef CONFIG_ARCH_RANDOM -static inline bool arch_get_random_long (unsigned long *v) +static inline bool __must_check arch_get_random_long (unsigned long *v) {return static_cpu_has (X86_FEATURE_RDRAND ) ?rdrand_long (v) : false; } -static inline bool arch_get_random_int (unsigned int *v) +static inline bool __must_check arch_get_random_int (unsigned int *v) {return static_cpu_has (X86_FEATURE_RDRAND ) ?rdrand_int (v) : false; } -static inline bool arch_get_random_seed_long (unsigned long *v) +static inline bool __must_check arch_get_random_seed_long (unsigned long *v) {return static_cpu_has (X86_FEATURE_RDSEED ) ?rdseed_long (v) : false; } -static inline bool arch_get_random_seed_int (unsigned int *v) +static inline bool __must_check arch_get_random_seed_int (unsigned int *v) {return static_cpu_has (X86_FEATURE_RDSEED ) ?rdseed_int (v) : false; }
Signature · 1 line
2.20.1
PATCH v2 8/10 powerpc: Use bool in archrandom.h
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
The generic interface uses bool not int; match that.
--- arch/powerpc/include/asm/archrandom.h | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-)
arch/powerpc/include/asm/archrandom.h +9 -8
@@ -6,27 +6,28 @@#include <asm/machdep.h> -static inline int arch_get_random_long (unsigned long *v) +static inline bool arch_get_random_long (unsigned long *v) { -return 0 ; +return false; } -static inline int arch_get_random_int (unsigned int *v) +static inline bool arch_get_random_int (unsigned int *v) { -return 0 ; +return false; } -static inline int arch_get_random_seed_long (unsigned long *v) +static inline bool arch_get_random_seed_long (unsigned long *v) {if (ppc_md.get_random_seed)return ppc_md.get_random_seed (v); -return 0 ; +return false; } -static inline int arch_get_random_seed_int (unsigned int *v) + +static inline bool arch_get_random_seed_int (unsigned int *v) {unsigned long val; -int rc; +bool rc; rc =arch_get_random_seed_long (&val);if (rc)
Signature · 1 line
2.20.1
PATCH v2 9/10 powerpc: Mark archrandom.h functions __must_check
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
We must not use the pointer output without validating the success of the random read.
--- arch/powerpc/include/asm/archrandom.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
arch/powerpc/include/asm/archrandom.h +4 -4
@@ -6,17 +6,17 @@#include <asm/machdep.h> -static inline bool arch_get_random_long (unsigned long *v) +static inline bool __must_check arch_get_random_long (unsigned long *v) {return false; } -static inline bool arch_get_random_int (unsigned int *v) +static inline bool __must_check arch_get_random_int (unsigned int *v) {return false; } -static inline bool arch_get_random_seed_long (unsigned long *v) +static inline bool __must_check arch_get_random_seed_long (unsigned long *v) {if (ppc_md.get_random_seed)return ppc_md.get_random_seed (v);
@@ -24,7 +24,7 @@ static inline bool arch_get_random_seed_long(unsigned long *v)return false; } -static inline bool arch_get_random_seed_int (unsigned int *v) +static inline bool __must_check arch_get_random_seed_int (unsigned int *v) {unsigned long val;bool rc;
Signature · 1 line
2.20.1
PATCH v2 10/10 s390x: Mark archrandom.h functions __must_check
In reply to Mark BrownFrom: Richard Henderson <[email protected]>
We must not use the pointer output without validating the success of the random read.
--- arch/s390/include/asm/archrandom.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
arch/s390/include/asm/archrandom.h +4 -4
@@ -21,17 +21,17 @@ extern atomic64_t s390_arch_random_counter;bool s390_arch_random_generate (u8 *buf,unsigned int nbytes); -static inline bool arch_get_random_long (unsigned long *v) +static inline bool __must_check arch_get_random_long (unsigned long *v) {return false; } -static inline bool arch_get_random_int (unsigned int *v) +static inline bool __must_check arch_get_random_int (unsigned int *v) {return false; } -static inline bool arch_get_random_seed_long (unsigned long *v) +static inline bool __must_check arch_get_random_seed_long (unsigned long *v) {if (static_branch_likely (&s390_arch_random_available)) {return s390_arch_random_generate ((u8 *)v,sizeof (*v));
@@ -39,7 +39,7 @@ static inline bool arch_get_random_seed_long(unsigned long *v)return false; } -static inline bool arch_get_random_seed_int (unsigned int *v) +static inline bool __must_check arch_get_random_seed_int (unsigned int *v) {if (static_branch_likely (&s390_arch_random_available)) {return s390_arch_random_generate ((u8 *)v,sizeof (*v));
Signature · 1 line
2.20.1
Quoted 7 lines · Mark Brown
This is a resend of a series from Richard Henderson last posted back in November:
https://lore.kernel.org/linux-arm-kernel/[email protected]/Back then Borislav said they looked good and asked if he should take them through the tip tree but things seem to have got lost since then.
Or, alternatively, akpm could take them. In any case, if someone else ends up doing that, for the x86 bits:
Thx.
Signature · 4 lines
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquetteTheodore Y. Ts'o
maintainerQuoted 6 lines · Borislav Petkov
Quoted 5 lines · Mark Brown
This is a resend of a series from Richard Henderson last posted back in November:
https://lore.kernel.org/linux-arm-kernel/[email protected]/
Quoted 7 lines
Quoted 2 lines
Back then Borislav said they looked good and asked if he should take them through the tip tree but things seem to have got lost since then.
Or, alternatively, akpm could take them. In any case, if someone else ends up doing that, for the x86 bits:
Reviewed-by Borislav Petkov <[email protected]>
Or I can take them through the random.git tree, since we have a lot of changes this cycle going to Linus anyway. Any objections?
- Ted
Quoted 8 lines · Theodore Y. Ts'o
Quoted 6 lines · Borislav Petkov
Quoted 4 lines · Mark Brown
This is a resend of a series from Richard Henderson last posted back in November:
https://lore.kernel.org/linux-arm-kernel/[email protected]/
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
Back then Borislav said they looked good and asked if he should take them through the tip tree but things seem to have got lost since then.
Or, alternatively, akpm could take them. In any case, if someone else ends up doing that, for the x86 bits:
Or I can take them through the random.git tree, since we have a lot of changes this cycle going to Linus anyway. Any objections?
I think the important thing here is that *someone* takes the patches. We've now got Ted and Borislav both saying they're OK applying the patches, an additional proposal that Andrew takes the patches, nobody saying anything negative about applying the patches and yet the patches are not applied. The random tree sounds like a sensible enough tree to take this so if Ted picks them up perhaps that's most sensible?
Borislav Petkov
maintainerQuoted 6 lines · Mark Brown
I think the important thing here is that *someone* takes the patches. We've now got Ted and Borislav both saying they're OK applying the patches, an additional proposal that Andrew takes the patches, nobody saying anything negative about applying the patches and yet the patches are not applied. The random tree sounds like a sensible enough tree to take this so if Ted picks them up perhaps that's most sensible?
Yes, Ted, pls pick them up so that we're done with this.
Thx.
Signature · 4 lines
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquetteQuoted 9 lines · Borislav Petkov
Quoted 6 lines · Mark Brown
I think the important thing here is that *someone* takes the patches. We've now got Ted and Borislav both saying they're OK applying the patches, an additional proposal that Andrew takes the patches, nobody saying anything negative about applying the patches and yet the patches are not applied. The random tree sounds like a sensible enough tree to take this so if Ted picks them up perhaps that's most sensible?
Yes, Ted, pls pick them up so that we're done with this.
I've picked them up and pushed them to the random tree.
- Ted