On 11/27/2014 11:23 PM, Dan Carpenter wrote:
>> +static inline void rtsx_pci_write_be32(struct rtsx_pcr *pcr, u16 reg, u32 val)
>> >+{
>> >+ rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg, 0xFF, val >> 24);
>> >+ rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg + 1, 0xFF, val >> 16);
>> >+ rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg + 2, 0xFF, val >> 8);
>> >+ rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg + 3, 0xFF, val);
> This assumes the cpu is little endian. First convert to big endian
> using cpu_to_be32() and then write it out.
>
> __be32 be_val = cpu_to_be32()
>
> rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg, 0xFF, be_val);
> rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg + 1, 0xFF, be_val >> 8);
> rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg + 2, 0xFF, be_val >> 16);
> rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, reg + 3, 0xFF, be_val >> 24);
>
> (Written hurredly in my mail client. May be wrong).
>
I think we better not use cpu_to_be32() here, leave the work to caller
may be better.
eg, in sd_ops.c the cmd.arg is constructed bit by bit, we can put the right
byte to the right register by shift, so the endian check is not need.?韬{.n?????%??檩??w?{.n???{炳i?)?骅w* jg????????G??⒏ ⒎?:+v???????? ????＂??????
