Hi,
The ARM implementation of memset has a bug when the fill-value is negative or outside the [0, 255] range. To reproduce:
int main() {
char array[256];
memset(array, -5, 256);
for (int i = 0; i < 256; ++i) {
printf("%d, ", (int)array[i]);
}
return 0;
}
This is supposed to fill the array with int8 values -5, -5, -5, ... . On ARM, this does not work because the implementation assumes the high bytes of the fill-value argument are already zero. However in this test case they are filled with 1-bits. The other implementations that I checked (aarch64 and x86_64) do not have this problem: they first convert the fill-value to an unsigned byte following the specification of memset.
With GCC one can use `memset(ptr, (-5 & 0xFF), size)` as a workaround, but for clang users that does not work: clang optimizes the `& 0xFF` away because it assumes that memset will do it.
How to fix/patch:
In this file: https://cgit.uclibc-ng.org/cgi/cgit/uclibc-ng.git/tree/libc/string/arm/mems…
Before line 35 (lsl) insert this: uxtb r1, r1
Before line 71 (orr) insert this: uxtb a2, a2
Best regards,
Tom
Hi Tom,
Tom Bannink wrote,
> Hi uClibc-ng developers,
>
> I've stumbled across a bug in the implementation of `lrint` in uClibc-ng. I'm
> not sure where to report this so I'm sending this email. Please let me know if
> I can submit an issue on a bugtracker somewhere. I tried to post an issue on
> the gitlab page (https://gogs.waldemar-brodkorb.de/oss/uclibc-ng/issues) but it
> seems I cannot make an account there.
>
> The bug is simple to reproduce: `lrint(0.5)` and `lrint(-0.5)` should return 0
> (or +-1 depending on the rounding mode) but instead they return 2 and -2
> respectively.
>
> The faulty code is the block at lines 55-64 of this file: https://
> cgit.uclibc-ng.org/cgi/cgit/uclibc-ng.git/tree/libm/s_lrint.c
> Note that the `lround` function works fine and has very similar code: https://
> cgit.uclibc-ng.org/cgi/cgit/uclibc-ng.git/tree/libm/s_lround.c
>
> Let me know if you need more information.
Can you provide a simple test case for the issue?
best regards
Waldemar
