Re: [PATCH v4 01/10] spi: Add a driver for the Freescale/NXP QuadSPI controller

[Schrempf Frieder <frieder.schrempf@xxxxxxxxxx>]

Hi Yogesh,

On 13.11.18 09:22, Yogesh Narayan Gaur wrote:
[...]
>> +
>> +static void fsl_qspi_read_ahb(struct fsl_qspi *q, const struct
>> +spi_mem_op *op) {
>> +	/*
>> +	 * We want to avoid needing to invalidate the cache by issueing
>> +	 * a reset to the AHB and Serial Flash domain, as this needs
>> +	 * time. So we change the address on each read to trigger an
>> +	 * actual read operation on the flash. The actual address for
>> +	 * the flash memory is set by programming the LUT.
>> +	 */
> As discussed previously, please go away with this hack and use AHB bus invalidation method with smaller timeout value.
> 
> I would start doing validation of this patch series from next version onward. As you have mentioned in other mail discussion about issue in the break condition for function  fsl_qspi_readl_poll_tout().

Thank you for your comments. I just sent v5 with some fixes, including 
fixed fsl_qspi_readl_poll_tout().

I also removed the hack above and I'm properly resetting the AHB domain 
now. I'm using a 1us delay just like in the old driver. In my tests the 
performance impact was very small (~2%) and it didn't get better when 
using ndelay.

Thanks,
Frieder
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/