Re: [PATCH 1/2] hwrng: iproc-rng200: Fix disable of the block.

From: Matthias Brugger <mbrugger@suse.com>
To: Scott Branden <scott.branden@broadcom.com>,
	matthias.bgg@kernel.org, mpm@selenic.com,
	herbert@gondor.apana.org.au, rjui@broadcom.com,
	sbranden@broadcom.com, f.fainelli@gmail.com
Cc: linux-kernel@vger.kernel.org, Julia.Lawall@inria.fr,
	bcm-kernel-feedback-list@broadcom.com,
	linux-arm-kernel@lists.infradead.org, nsaenzjulienne@suse.de,
	linux-crypto@vger.kernel.org
Subject: Re: [PATCH 1/2] hwrng: iproc-rng200: Fix disable of the block.
Date: Fri, 18 Dec 2020 11:49:55 +0100	[thread overview]
Message-ID: <ab41dfe2-e244-f920-5d1f-bb34ac16fe8f@suse.com> (raw)
In-Reply-To: <24e22e5a-7371-02b4-1636-2e03def420e3@broadcom.com>

On 14/12/2020 18:45, Scott Branden wrote:
> 
> 
> On 2020-12-14 8:04 a.m., matthias.bgg@kernel.org wrote:
>> From: Matthias Brugger <mbrugger@suse.com>
>>
>> When trying to disable the block we bitwise or the control
>> register with value zero. This will leave the block always turned on.
>> Fix this by setting the corresponding bit to zero.
>>
>> Fixes: c83d45d5685f ("hwrng: iproc-rng200 - Add Broadcom IPROC RNG driver")
> Commit message needs to be re-written.
> I don't think this is an actual fix as the ~RNG_CTL_RNG_RBGEN_MASK already zeros the bit.  This is just a code change, which is fine because it makes things clearer

Right, I'll reword it.

>> Signed-off-by: Matthias Brugger <mbrugger@suse.com>
>> ---
>>
>>  drivers/char/hw_random/iproc-rng200.c | 5 ++---
>>  1 file changed, 2 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/char/hw_random/iproc-rng200.c b/drivers/char/hw_random/iproc-rng200.c
>> index 01583faf9893..e106ce3c0146 100644
>> --- a/drivers/char/hw_random/iproc-rng200.c
>> +++ b/drivers/char/hw_random/iproc-rng200.c
>> @@ -28,7 +28,6 @@
>>  #define RNG_CTRL_OFFSET					0x00
>>  #define RNG_CTRL_RNG_RBGEN_MASK				0x00001FFF
>>  #define RNG_CTRL_RNG_RBGEN_ENABLE			0x00000001
>> -#define RNG_CTRL_RNG_RBGEN_DISABLE			0x00000000
>>  
>>  #define RNG_SOFT_RESET_OFFSET				0x04
>>  #define RNG_SOFT_RESET					0x00000001
>> @@ -61,7 +60,7 @@ static void iproc_rng200_restart(void __iomem *rng_base)
>>  	/* Disable RBG */
>>  	val = ioread32(rng_base + RNG_CTRL_OFFSET);
>>  	val &= ~RNG_CTRL_RNG_RBGEN_MASK;
> This mask will already zero the enable bit.
>> -	val |= RNG_CTRL_RNG_RBGEN_DISABLE;
>> +	val &= ~RNG_CTRL_RNG_RBGEN_ENABLE;

Correct, so no need for this extra line. I'll delete it.

>>  	iowrite32(val, rng_base + RNG_CTRL_OFFSET);
>>  
>>  	/* Clear all interrupt status */
>> @@ -174,7 +173,7 @@ static void iproc_rng200_cleanup(struct hwrng *rng)
>>  	/* Disable RNG hardware */
>>  	val = ioread32(priv->base + RNG_CTRL_OFFSET);
>>  	val &= ~RNG_CTRL_RNG_RBGEN_MASK;
>> -	val |= RNG_CTRL_RNG_RBGEN_DISABLE;
>> +	val &= ~RNG_CTRL_RNG_RBGEN_ENABLE;
>>  	iowrite32(val, priv->base + RNG_CTRL_OFFSET);
>>  }
>>  
> 