On 22/09/2015 15:48, Felix Fietkau wrote:
> On 2015-09-22 15:26, Michael Lee wrote:
>> From: michael lee <[email protected]>
>>
>> only support spi flash command (half duplex).
>> no need chunk io patch. done by driver.
>>
>> test results on mt7621. use dd read bs=512 with 32128 records
>> old driver : 30.52s
>> new driver : 34.31s
>> new driver + no chunk io : 16.65s
>> new driver + no chunk io + fast read clock from 10Mhz to 50Mhz : 5.00s
> Where does the performance gain without the clock change come from?
>
> - Felix
>
Hi,
i am working on a different patch that does several consecutive
fullduplex transactions which improves the speed by more than 10%. i
have not measure it but subjective feeling is more like 50-100%.
also this patch has a lot of stuff all mixed into a single patch making
it near impossible to review the code.
and, you make the driver m25p80 aware which is SDK style. we dont want
that in our kernel.
John
_______________________________________________
> openwrt-devel mailing list
> [email protected]
> https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
>
_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
