Re: [PATCH] crypto: qce: Initialize core src clock @100Mhz

2016-09-07 Thread Iaroslav Gridin
> > + ret = clk_set_rate(qce->core_src, 1); > > Could you point me from where you got this number? I got it from codeaurora qce driver: https://android.googlesource.com/kernel/msm/+/android-msm-hammerhead-3.4-kk-r1/drivers/crypto/msm/qce50.c#3386 -- To unsubscribe from this list: se

Re: [PATCH] crypto: qce: Initialize core src clock @100Mhz

2016-09-07 Thread Iaroslav Gridin
On Wed, Sep 07, 2016 at 04:04:01PM +0300, Stanimir Varbanov wrote: > Hi Iaroslav, > > On 09/03/2016 07:45 PM, Iaroslav Gridin wrote: > > Without that, QCE performance is about 2x less. > > On which platform? The clock rates are per SoC. Dragonboard 8074. Should clock ra

[PATCH] crypto: qce: Initialize core src clock @100Mhz

2016-09-03 Thread Iaroslav Gridin
Without that, QCE performance is about 2x less. Signed-off-by: Iaroslav Gridin --- drivers/crypto/qce/core.c | 18 +- drivers/crypto/qce/core.h | 2 +- 2 files changed, 18 insertions(+), 2 deletions(-) diff --git a/drivers/crypto/qce/core.c b/drivers/crypto/qce/core.c index

[PATCH 1/4] crypto: qce: Remove unneeded length check for scatterlist

2016-08-30 Thread Iaroslav Gridin
From: Voker57 Current code avoids supplying scatterlist containing more data than used to DMA. This leads to dropping data from scatterlists which would leave some for next run. Signed-off-by: Iaroslav Gridin --- drivers/crypto/qce/sha.c | 2 -- 1 file changed, 2 deletions(-) diff --git a

[PATCH 3/4] crypto: qce: Ensure QCE receives no zero-sized updates

2016-08-30 Thread Iaroslav Gridin
From: Voker57 Zero-sized updates lock QCE, so ensure there's always some data left for the final update, up to blocksize. Signed-off-by: Iaroslav Gridin --- drivers/crypto/qce/sha.c | 30 -- 1 file changed, 20 insertions(+), 10 deletions(-) diff --git a/dr

[PATCH 4/4] crypto: qce: If total text size is zero, return pre-computed digest

2016-08-30 Thread Iaroslav Gridin
From: Voker57 If total data amount to hash is zero, we cannot submit it to QCE, since it locks up on zero-sized updates. So, return pre-computed SHA256/SHA1 hash. Signed-off-by: Iaroslav Gridin --- drivers/crypto/qce/sha.c | 27 --- 1 file changed, 24 insertions(+), 3

[PATCH 2/4] crypto: qce: Avoid repeat hash finalization

2016-08-30 Thread Iaroslav Gridin
From: Voker57 Calling QCE finalization when hash have already been finalized causes a lockup. Avoid it by introducing finalized flag. Signed-off-by: Iaroslav Gridin --- drivers/crypto/qce/sha.c | 6 ++ drivers/crypto/qce/sha.h | 1 + 2 files changed, 7 insertions(+) diff --git a/drivers

[no subject]

2016-08-30 Thread Iaroslav Gridin
This set of patches fixes QCE digest code, preventing lockups and incorrect results. -- To unsubscribe from this list: send the line "unsubscribe linux-crypto" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html