IRC channel logs


back to list of logs

<flavioc>youpi: any suggestion for how to get someone to review my gcc patch for 64 bits on the gcc mailing list?
<flavioc>binutils with 64 bit support on hurd was released in the latest version :)
<damo22>flavioc: youpi1: this commit broke my smp branch * d5e5dd34 Update so that we don't need glibc when running ./configure.
<damo22>it causes the buffer overrun kind of bug
<damo22>proc auth.
<flavioc>damo22: does it work if you revert just that change? I wonder if it's changing the CFLAGS somehow even though it is not supposed to
<flavioc>what is the ./configure command that you are using?
<damo22>configure --enable-ncpus=8 --enable-kdb --enable-apic
<damo22>but on my branch
<flavioc>is the branch in the official repo?
<flavioc>I would look into the gcc commands produced by make before and after the patch. it is likely affecting the cflags being passed or even the result of the configuration
<damo22>this branch with the revert seems to work
<damo22>i am getting a compile error for --disable-apic (default)
<damo22>i need to fix that so i can test with PIC
<damo22>i am running qemu with -smp 1 for now
<flavioc>seems like w/ my patch we no longer pass -O2 -g in gcc. I think those seem to be autoconf defaults
<flavioc>let me know if ./configure CFLAGS="-O2 -g" fixes the problem
<damo22>ok i can test shortly
<damo22>flavioc: that fixed it
<damo22>youpi1: does every processor need its own kernel_pmap?
<damo22>or can they share it
<damo22>there seems to be no support for separate ones