Performing substitutions during post-patch breaks tools such as mkpatches,
making it very difficult to regenerate correct patches after making changes,
and often leading to substituted string replacements being committed.
Doublequote the eval arguments to ensure whitespace is preserved, so
a cpu_brand of "Genuine Intel(R) CPU T2500 @ 2.00GHz"
is not converted down to "Genuine Intel(R) CPU T2500 @ 2.00GHz"
- Add "Intel(r) Xeon(r) CPU E5430" and strip leading whitespace from
Solaris psrinfo output - from Jim Wise
- Add "Intel(R) Xeon(R) CPU E5310" - from Felix Resch
- Add "Genuine Intel(R) CPU T2500" - from Daniel Carosone
Intel appear to have re-used cpu branding strings between Northwood
and Prescott pentium4s. Thats just... special.
Handle this by explicitly testing for SSE3 support to distinguish between
'-march=prescott' and '-march=pentium4'
Add:
"Intel(R) Celeron(R) M processor "* - thanks to Daniel Horecki
"Intel(R) Pentium(R) 4 CPU 3.20GHz" - thanks to John Nemeth
"VIA Nehemiah" and "Intel(R) Atom(TM) CPU N270 " - thanks to Joel CARNAT
Add map_x86_brand_string to subr_x86 and use in NetBSD,
Linux and Solaris to determine ARCH for Intel CPUs after PIII.
(Still need to rework AMD CPU detection to use family and
extended family)
Added 'AMD K6-2 (586-class)' - thanks Brian Ginsbach
Added 'Pentium(R) Dual-Core CPU E5200' - thanks Andreas Burghardt
- Updates for NetBSD:
- Report OS version in verbose mode
- Cleanup the parsing of "cpuctl identify 0" / "grep ^cpu0: dmesg"
- Add cases for:
- 'AMD Athlon 64 X2 (686-class)' - thanks Ryo HAYASAKA
- 'Intel Pentium M (Yonah) (686-class)' - thanks Juho Juopperi
- 'Intel Pentium III (Katmai) (686-class)'
- Adjust '-march=core2' facllback to pentium-m - thanks Thomas E. Spanjaard
- Pickup -'msse3'
- Updates for gcc:
- Add some more x86 -march cases from gcc 4.3
Completely rework i386 and x86_64 processor checking on NetBSD:
If cpuctl is present, use the first list of its output, otherwise
rely on the first cpu0: line in /var/run/dmesg.boot
Unfortunately this means we will need to reenter some of the older CPU
details.
- return 0 in verbose() to fix issue with 'set -e'
- ONLY_FOR_PLATFORM += FreeBSD-*-*
- Use `` rather than $() to unbreak Solaris
- Add x86 arch fixup/corrections in files/subr_x86
All from Yakovetsky Vladimir
