[PATCH 02/24] mtd: rawnand: sunxi_spl: fix pointer from integer without a cast

[Richard GENOUD]

Le 17/10/2025 à 09:57, Andre Przywara a écrit :
> Hi,
> 
> On 16/10/2025 15:27, Richard Genoud wrote:
>> Fix pointer from interget warning when compiling for ARM64
>>
>> When compiling for arm64, we get this error:
>> error: passing argument 2 of ‘__memcpy_fromio’ makes pointer from
>>             integer without a cast [-Wint-conversion]
>>
>> Fix that with a cast to (void *)(uintptr_t)
>>
>> Signed-off-by: Richard Genoud <richard.genoud at bootlin.com>
>> ---
>>   drivers/mtd/nand/raw/sunxi_nand_spl.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/mtd/nand/raw/sunxi_nand_spl.c b/drivers/mtd/nand/ 
>> raw/sunxi_nand_spl.c
>> index 4f1e2d9a5775..19091ece632b 100644
>> --- a/drivers/mtd/nand/raw/sunxi_nand_spl.c
>> +++ b/drivers/mtd/nand/raw/sunxi_nand_spl.c
>> @@ -310,7 +310,8 @@ static int nand_read_page(const struct nfc_config 
>> *conf, u32 offs,
>>               return 1;
>>           /* Retrieve the data from SRAM */
>> -        memcpy_fromio(data, SUNXI_NFC_BASE + NFC_RAM0_BASE,
>> +        memcpy_fromio(data,
>> +                  (void *)(uintptr_t)SUNXI_NFC_BASE + NFC_RAM0_BASE,
> 
> I don't think memcpy_fromio() is correct here in the first place. It 
> looks tempting to use it, but the implementation is not safe to use with 
> Allwinner peripherals: we need to stick to the predefined device 
> register access width. Neither the arm nor the arm64 implementation are 
> fully compliant. If you go by the book, the peripheral registers are 32- 
> bit wide, and must only be accessed via a 32-bit wide access (readl/ 
> writel). Empirically 8-bit accesses to a 32-bit aligned address also 
> seem to work, but not to other addresses. Everything else is unsafe and 
> should be avoided.
> So can you please replace this call with a for-loop of readl() 
> invocations? This should be safe with all devices, and since the base 
> address is 32-bit aligned, and I assume ecc_size is as well, we are good 
> here.
Ok, sure.

Thanks!

> 
> Cheers,
> Andre
> 
>>                     conf->ecc_size);
>>           /* Stop the ECC engine */
> 


-- 
Richard Genoud, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com