Max Filippov

AFAICS movsf_internal pattern for xtensa does not have suitable constraints to match loading SF mode constant from a literal pool into an FP register, that's what offending RTL pattern tries...

Please consider using prebuilt toolchain for Windows, like [this one](https://programs74.ru/udkew-en.html), or [this one](https://www.esp8266.com/wiki/doku.php?id=setup-windows-compiler-esp8266). It will save you a lot of trouble if all you need is the toolchain. If you...

> How can I find the configure file? Debug what the script does. - check if the configure file is actually there (it should be in the ./dl/gmp-6.0.0a, if not...

> trying to get the "xtensa-ctng-7.2.0" branch to work Why do you think it's the branch that you need? > Is there a way to get a recent version of...

So I've added dummy ASAN support to the xtensa gcc: 1ebcd8481b0762f2126c1478d66a1f4aee369798 It will appear in gcc-7.3 and gcc-8. Using it I was able to enable KASAN support for xtensa in...

That looks interesting. One workaround that I see is the switch -fomit-frame-pointer, which frees a15. But the assembly only wants 14 registers, which we have even when the frame pointer...

> Any idea ? I haven't found an ultimate answer to what happens here, seems to be bad interaction of newlib and libstdc++. I used this workaround: https://github.com/jcmvbkbc/crosstool-NG/blob/b38a1d4cb9f0179a1da5c210e476aa2c9229bb94/local-patches/gcc/5.1.0/0001-WIP-don-t-bring-extra-u-int_least32_t-into-std.patch

Yes, I'm running nm-ssh-service with this change for a few weeks now on a machine with sh symlinked to dash and tap0 used for qemu network.

Ping. I wonder if there's anything that prevents merging this simple fix?

> Or we just compile the code using the latest xtensa GCC with stack smashing protection support It's backported to gcc-5.2.0 and the patch is in the xtensa-1.22.x branch of...