From: Zoltan B. <zb...@fr...> - 2004-11-14 22:28:15
|
Hi, I just wanted to tell my second success story on this Athlon64 machine. :-) MSI K8T Neo FIS2R mainboard, Radeon 9200SE AGP, Radeon 7000 PCI, two PS2 keyboards, two USB mice. I deleted my old 32 bit Fedora Core 1 and installed Fedora Core 3/x86_64 and today I finished setting it up for multihead. I patched the Fedora official kernel with ruby-2.6.9-20041021.diff.bz2, the patch applied smoothly but that didn't compile at first because of this chunk for fs/compat_ioctl.c: ---------------------------------------------- @@ -1609,8 +1611,10 @@ static int do_fontx_ioctl(unsigned int f get_user(data, &user_cfd->chardata)) return -EFAULT; op.data =3D compat_ptr(data); - return con_font_op(fg_console, &op); + return con_font_op(vc, &op); case GIO_FONTX: + if (!cfdarg.chardata) + return 0; op.op =3D KD_FONT_OP_GET; op.flags =3D 0; op.width =3D 8; ---------------------------------------------- The "if (!cfdarg.chardata)" should be "if (!user_cfd->chardata)". Another problem is that I had to disable building the smp kernel because vt.c did not want to link at the very end of the compilation. It was OK during the UP kernel, though. It turned out that "free_bootmem()" was not defined when compiling for SMP kernel. It was only declared in some .h file, so it compiled but didn't link. I looked at the changes in the ruby patch and it's suspicious that the original vt.c does not call free_bootmem() anywhere, though it calls alloc_bootmem(). Commenting out the free_bootmem() call should have fixed it. I have to apply the IsolatePCI patch to FC3's XOrg X11 6.8.1. It works as well as my FC1 before, and it still has the one problem that does not annoy me anymore: logging out on the first display ( :0 ) blanks the second display ( :1 ). Changing resolution by pressing Crtl-Alt-Num+/- brings the second display back. Best regards, Zolt=E1n B=F6sz=F6rm=E9nyi |