Skip to contentSkip to conversation

sun4i_ss_prng fixes

PATCH In review Artem Savkov 4 participants 7 messages Feb 6, 2018 – Feb 9, 2018

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover sun4i_ss_prng fixes

IPSec hasn't been working on my a10 board since 4.14 and it turned out to be caused by sun4i_ss_rng driver.

Artem Savkov (2):
sun4i_ss_prng: fix return value of sun4i_ss_prng_generate
sun4i_ss_prng: convert lock to _bh in sun4i_ss_prng_generate

 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
Signature · 1 line
2.15.1
#1 lore raw

PATCH 1/2 sun4i_ss_prng: fix return value of sun4i_ss_prng_generate

According to crypto/rng.h generate function should return 0 on success and < 0 on error.

Fixes b8ae5c7387ad ("crypto: sun4i-ss - support the Security System PRNG")
---
 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +1 -1
index 0d01d1624252..5754e0b92fb0 100644
@@ -52,5 +52,5 @@ int sun4i_ss_prng_generate(struct crypto_rng *tfm, const u8 *src,
 
 	writel(0, ss->base + SS_CTL);
 	spin_unlock(&ss->slock);
-	return dlen;
+	return 0;
 }
Signature · 1 line
2.15.1
#2 lore raw

Re: [PATCH 1/2] sun4i_ss_prng: fix return value of sun4i_ss_prng_generate

Quoted 23 lines · Artem Savkov

According to crypto/rng.h generate function should return 0 on success and < 0 on error.

Fixes b8ae5c7387ad ("crypto: sun4i-ss - support the Security System PRNG")
---
 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +1 -1
index 0d01d1624252..5754e0b92fb0 100644
@@ -52,5 +52,5 @@ int sun4i_ss_prng_generate(struct crypto_rng *tfm, const u8 *src,
 
 	writel(0, ss->base + SS_CTL);
 	spin_unlock(&ss->slock);
-	return dlen;
+	return 0;
 }
Signature · 1 line
2.15.1

According to Documentation/crypto/api-samples.rst ("Code Example For Random Number Generator Usage") you must return the length of data generated.

So crypto_rng_generate/crypto_rng_get_bytes documentation in crypto/rng.h must be fixed.

Herbert could you confirm ?

Regards

#3 lore raw

Re: [PATCH 1/2] sun4i_ss_prng: fix return value of sun4i_ss_prng_generate

Quoted 16 lines · Corentin Labbe
Quoted 15 lines · Artem Savkov

According to crypto/rng.h generate function should return 0 on success and < 0 on error.

Fixes b8ae5c7387ad ("crypto: sun4i-ss - support the Security System PRNG")
---
 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +0 -0
index 0d01d1624252..5754e0b92fb0 100644
@@ -52,5 +52,5 @@ int sun4i_ss_prng_generate(struct crypto_rng *tfm, const u8 *src,
 
Quoted 11 lines
Quoted 8 lines
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +1 -1
 	writel(0, ss->base + SS_CTL);
 	spin_unlock(&ss->slock);
-	return dlen;
+	return 0;
 }
Signature · 1 line
2.15.1

According to Documentation/crypto/api-samples.rst ("Code Example For Random Number Generator Usage") you must return the length of data generated.

I don't think that example is the same as rng_alg.generate, it has a different protottype.

Quoted 1 line

So crypto_rng_generate/crypto_rng_get_bytes documentation in crypto/rng.h must be fixed.

It's not just documentation, every other rng driver returns it this way and it gets aead_init_geniv() (and subsequently crypto_create_tfm()) really confused because they expect the return value to be 0 or < 0.

Quoted 1 line

Herbert could you confirm ?

Signature · 2 lines
Regards,
  Artem
#4 lore raw

Re: [PATCH 1/2] sun4i_ss_prng: fix return value of sun4i_ss_prng_generate

Quoted 29 lines · Artem Savkov
Quoted 16 lines · Corentin Labbe
Quoted 15 lines · Artem Savkov

According to crypto/rng.h generate function should return 0 on success and < 0 on error.

Fixes b8ae5c7387ad ("crypto: sun4i-ss - support the Security System PRNG")
---
 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +0 -0
index 0d01d1624252..5754e0b92fb0 100644
@@ -52,5 +52,5 @@ int sun4i_ss_prng_generate(struct crypto_rng *tfm, const u8 *src,
 
Quoted 11 lines
Quoted 8 lines
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +1 -1
 	writel(0, ss->base + SS_CTL);
 	spin_unlock(&ss->slock);
-	return dlen;
+	return 0;
 }
Signature · 1 line
2.15.1

According to Documentation/crypto/api-samples.rst ("Code Example For Random Number Generator Usage") you must return the length of data generated.

Quoted 2 lines

I don't think that example is the same as rng_alg.generate, it has a different protottype.

crypto_rng_get_bytes() use crypto_rng_generate() which work on crypto_rng

Quoted 7 lines
Quoted 2 lines

So crypto_rng_generate/crypto_rng_get_bytes documentation in crypto/rng.h must be fixed.

It's not just documentation, every other rng driver returns it this way and it gets aead_init_geniv() (and subsequently crypto_create_tfm()) really confused because they expect the return value to be 0 or < 0.

Perhaps the example is wrong because "copied from/inspirated by" hwrng API which have this behavior.

After checking, other crypto_rng driver, you are probably right, so you can add:

Thanks
Regards

#5 lore raw

PATCH 2/2 sun4i_ss_prng: convert lock to _bh in sun4i_ss_prng_generate

In reply to Artem Savkov

Lockdep detects a possible deadlock in sun4i_ss_prng_generate() and throws an "inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage" warning. Disabling softirqs to fix this.

Fixes b8ae5c7387ad ("crypto: sun4i-ss - support the Security System PRNG")
---
 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/crypto/sunxi-ss/sun4i-ss-prng.c +2 -2
index 5754e0b92fb0..63d636424161 100644
@@ -28,7 +28,7 @@ int sun4i_ss_prng_generate(struct crypto_rng *tfm, const u8 *src,
 	algt = container_of(alg, struct sun4i_ss_alg_template, alg.rng);
 	ss = algt->ss;
 
-	spin_lock(&ss->slock);
+	spin_lock_bh(&ss->slock);
 
 	writel(mode, ss->base + SS_CTL);
 
@@ -51,6 +51,6 @@ int sun4i_ss_prng_generate(struct crypto_rng *tfm, const u8 *src,
 	}
 
 	writel(0, ss->base + SS_CTL);
-	spin_unlock(&ss->slock);
+	spin_unlock_bh(&ss->slock);
 	return 0;
 }
Signature · 1 line
2.15.1
maintainer
#6 lore raw
In reply to Artem Savkov
Quoted 9 lines · Artem Savkov

IPSec hasn't been working on my a10 board since 4.14 and it turned out to be caused by sun4i_ss_rng driver.

Artem Savkov (2):
sun4i_ss_prng: fix return value of sun4i_ss_prng_generate
sun4i_ss_prng: convert lock to _bh in sun4i_ss_prng_generate

 drivers/crypto/sunxi-ss/sun4i-ss-prng.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

All applied. Thanks.

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt