Re: [PATCH v2] riscv: fix incorrect use of __user pointer

[Clément Léger]

On 25/11/2023 16:37, David Laight wrote:
> ...
>> @@ -491,7 +486,7 @@ int handle_misaligned_load(struct pt_regs *regs)
>>
>>  	val.data_u64 = 0;
>>  	for (i = 0; i < len; i++) {
>> -		if (load_u8(regs, (void *)(addr + i), &val.data_bytes[i]))
>> +		if (load_u8(regs, addr + i, &val.data_bytes[i]))
>>  			return -1;
>>  	}
> 
> I'd really have thought that you'd want to pull the kernel/user
> check way outside the loop?

Hi David,

I hope the compiler is able to extract that 'if' out of the loop since
regs isn't modified in the loop. Nevertheless, that could be more
"clear" if put outside indeed.

> In any case, for a misaligned read why not just read (addr & ~7)[0]
> and (if needed) (addr & ~7)[1] and then ahift and or together?

Makes sense, the original code was like that but probably copy/pasted
from openSBI I guess. (?)

Let's keep that for the moment (this patch is about fixing wrong __user
address space). I will try to submit another series using what you proposed.

Regards,

Clément

> 
> clang will do it for misaligned structure members with known
> misalignment, but it is almost certainly also better for reads
> with unknown misalignment.
> 
> 	David
> 
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)