Fixes build errors on Linux, tested on CentOS.
The problem was pointed out by Valery Ushakov.
Note <sys/param.h> will be prepared by src/tools/compat/configure.
XXX: not sure if we still need this dumb tool for a TierII port...
the TOOLDIR version of libnbcompat, associated include files,
and associated defs.mk file, instead of the version from the
.OBJDIR of src/tools/compat. This should fix PR 47188.
on non-NetBSD hosts. Tested on CYGWIN_NT-5.1-1.5.25-i686.
XXX: It's enough to have generated boothfs.uue binary
XXX: (for mkisofs -boot-hfs-file option), rather than
XXX: building a tool and generating it at run time.
CD-ROM as the root file system. It contains a limited subset of $DESTDIR
along with sysinst. This will help to massively reduce the amount of memory
needed to install from CD or netboot, and allow us to continue using the
generic kernel on the CDs.
- Make MULTIPROCESSOR mandatory on i386.
Installation changes:
- Update installation section of release notes to match reality.
- Rename INSTALL to INSTALL_FLOPPY, retire INSTALL_LARGE.
- Build INSTALL kernel from GENERIC, like on amd64.
- Update boot menu to allow disabling ACPI and/or SMP.
- Remove GENERIC.NOACPI from the installed kernel list.
- TODO: install default boot.cfg in etc.tgz.
- TODO: possibly enable PCI fixup stuff at runtime if ACPI is disabled.
Build changes:
- No longer build ALL, it's for verification, is slow to build and the
build process is already crippled by the number of kernels built.
- No longer build GENERIC.NOACPI.
RSYNC_HOST
specify rsync host (default: rsync.NetBSD.org)
RSYNC_PATH_PREFIX
specify extra prefix path of NetBSD module names for rsync mirrors
ex. "/pub", "/netbsd" etc. (default: empty)
DAILY_DIR
specify directory name on using NetBSD-daily snapshot for -current
ex. "200712060002Z" etc. (no default; mandatory for -current)
RELEASE_SUFFIX
specify suffix of directory names used for pre-releases
ex. "_BETA2", "_RC5" etc. (default: empty)
RSYNC_PATH
specify path used to rsync hosts (default: set from the above variables)
Examples:
make RELEASE=3.1 TARGET_CD_IMAGE=mac68kcd fetch
-> fetch 3.1 mac68k sets from
rsync.NetBSD.org:/NetBSD/NetBSD-3.1/
make RELEASE=4.0 TARGET_CD_IMAGE=mac68kcd RELEASE_SUFFIX=_RC5 fetch
-> fetch 4.0_RC5 mac68k sets from
rsync.NetBSD.org:/NetBSD/NetBSD-4.0_RC5/
make RELEASE=current TARGET_CD_IMAGE=mac68kcd DAILY_DIR=200712060002Z fetch
-> fetch -current mac68k sets from
rsync.NetBSD.org:/NetBSD-daily/HEAD/200712060002Z/
make RELEASE=4.0 TARGET_CD_IMAGE=mac68kcd RELEASE_SUFFIX=_RC5 \
RSYNC_HOST=rsync3.jp.NetBSD.org RSYNC_PATH_PREFIX=/pub fetch
-> fetch 4.0_RC5 mac68k sets from
rsync3.jp.NetBSD.org:/pub/NetBSD/NetBSD-4.0_RC5/
make RELEASE=3.1 TARGET_CD_IMAGE=mac68kcd \
RSYNC_PATH=/NetBSD-daily/netbsd-3/200712060002Z fetch
-> fetch mac68k sets of netbsd-3 branch from
rsync.NetBSD.org:/pub/NetBSD-daily/netbsd-3/200712060002Z
("dd if=/dev/zero bs=512 count=0" seems to cause unexpected results)
- use ${TOOL_STAT} to get file size instead of ls and awk
(1.6.x don't have ${TOOL_STAT} but they have been obsolete anyway)
Both are taken from sys/arch/i386/stand/bootxx/Makefile.bootxx.
Use x86 native boot for current.
Swap i386 and amd64 in multiarch CD sets (cd1 would overflow by 30M otherwise)
and leave vax out (cd3 would overflow by more than 100M - time for a
multi-cd4 ?)
A bootable sgimips has a volume header prepended, and so is useless for
anything but sgimips. The issue was not found before because the
logic to make an ISO bootable on sgimips was broken, and has been fixed by
tsutsui@ on 2007/03/04.
- introduce SGIMIPS_SGIVOL variable which specifies sgivol(8) path
because nbsgivol tool is available only in TOOLDIR for sgimips and
possibly exists in a different directory from default TOOLDIR path
- use fixed (4096) value for SGIVOLHDR.size because default value
taken from <sys/bootblock.h> is 3135 (not multiples of 4) so it
doesn't match 2048bytes/sector CD images
- also pass SGIVOLHDR.size to sgivol(8) on creating volume
- rename raw iso9660 image before creating SGIVOLHDR to avoid
incomplete (non-bootable) image left on failure of sgivol(8)
Problem (i.e. sgimipscd-3.0.2.iso is not bootable) was reported
by Thierry Lacoste on port-sgimips, and fixed images are also
tested by him on IP22 (and by me on IP32).
christos