Re: [RFC] bmac:change to use bitrev8() generic function

[yalin wang]

> On Aug 10, 2015, at 20:02, David Laight  wrote:
> 
>> From: Tobias Klauser
>> Sent: 10 August 2015 12:49
>> On 2015-08-10 at 11:53:41 +0200, yalin wang  wrote:
>>> This change to use generic bitrev8() for bmac driver.
> ...
>>> @@ -871,7 +860,7 @@ bmac_addhash(struct bmac_data *bp, unsigned char *addr)
>>> 
>>> if (!(*addr)) return;
>>> crc = bmac_crc((unsigned short *)addr) & 0x3f; /* Big-endian alert! */
> 
> Why not *((u8 *)addr + 1) & 0x3f
> 
>>> -   crc = reverse6[crc];/* Hyperfast bit-reversing algorithm */
>>> +   crc = bitrev8((u8)crc);
>> 
>> No, this won't work. reverse6 works on 6 bit values, while bitrev8 works
>> on 8bit values, e.g. reverse6[1] = 0x20, while bitrev8(1) = 0x80
> 
> You could use bitrev8(n) >> 2.
> 
> But that is a 'strange' map of a 7-bit value to a 6-bit one.
> 
> I thought it was more common for ethernet hardware to use the
> value of the crc register after all 6 bytes of the mac address
> had been processed.
oh, my mistake, i will change it in V2 for review.--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

RE: [RFC] bmac:change to use bitrev8() generic function

[David Laight]

> From: Tobias Klauser
> Sent: 10 August 2015 12:49
> On 2015-08-10 at 11:53:41 +0200, yalin wang  wrote:
> > This change to use generic bitrev8() for bmac driver.
...
> > @@ -871,7 +860,7 @@ bmac_addhash(struct bmac_data *bp, unsigned char *addr)
> >
> > if (!(*addr)) return;
> > crc = bmac_crc((unsigned short *)addr) & 0x3f; /* Big-endian alert! */

Why not *((u8 *)addr + 1) & 0x3f

> > -   crc = reverse6[crc];/* Hyperfast bit-reversing algorithm */
> > +   crc = bitrev8((u8)crc);
> 
> No, this won't work. reverse6 works on 6 bit values, while bitrev8 works
> on 8bit values, e.g. reverse6[1] = 0x20, while bitrev8(1) = 0x80

You could use bitrev8(n) >> 2.

But that is a 'strange' map of a 7-bit value to a 6-bit one.

I thought it was more common for ethernet hardware to use the
value of the crc register after all 6 bytes of the mac address
had been processed.

David

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [RFC] bmac:change to use bitrev8() generic function

[Tobias Klauser]

On 2015-08-10 at 11:53:41 +0200, yalin wang  wrote:
> This change to use generic bitrev8() for bmac driver.
> 
> Signed-off-by: yalin wang 
> ---
>  drivers/net/ethernet/apple/bmac.c | 17 +++--
>  1 file changed, 3 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/net/ethernet/apple/bmac.c 
> b/drivers/net/ethernet/apple/bmac.c
> index a65d7a6..e749747 100644
> --- a/drivers/net/ethernet/apple/bmac.c
> +++ b/drivers/net/ethernet/apple/bmac.c
> @@ -22,6 +22,7 @@
>  #include 
>  #include 
>  #include 
> +#include 
>  #include 
>  #include 
>  #include 
> @@ -805,18 +806,6 @@ static irqreturn_t bmac_txdma_intr(int irq, void *dev_id)
>  }
>  
>  #ifndef SUNHME_MULTICAST
> -/* Real fast bit-reversal algorithm, 6-bit values */
> -static int reverse6[64] = {
> - 0x0,0x20,0x10,0x30,0x8,0x28,0x18,0x38,
> - 0x4,0x24,0x14,0x34,0xc,0x2c,0x1c,0x3c,
> - 0x2,0x22,0x12,0x32,0xa,0x2a,0x1a,0x3a,
> - 0x6,0x26,0x16,0x36,0xe,0x2e,0x1e,0x3e,
> - 0x1,0x21,0x11,0x31,0x9,0x29,0x19,0x39,
> - 0x5,0x25,0x15,0x35,0xd,0x2d,0x1d,0x3d,
> - 0x3,0x23,0x13,0x33,0xb,0x2b,0x1b,0x3b,
> - 0x7,0x27,0x17,0x37,0xf,0x2f,0x1f,0x3f
> -};
> -
>  static unsigned int
>  crc416(unsigned int curval, unsigned short nxtval)
>  {
> @@ -871,7 +860,7 @@ bmac_addhash(struct bmac_data *bp, unsigned char *addr)
>  
>   if (!(*addr)) return;
>   crc = bmac_crc((unsigned short *)addr) & 0x3f; /* Big-endian alert! */
> - crc = reverse6[crc];/* Hyperfast bit-reversing algorithm */
> + crc = bitrev8((u8)crc);

No, this won't work. reverse6 works on 6 bit values, while bitrev8 works
on 8bit values, e.g. reverse6[1] = 0x20, while bitrev8(1) = 0x80
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/