5:05 a.m.
Hi,
finally created a LiveDVD ISO automatically booting x86_64 OS on x86_64
(and i686 otherwise). Regular users will not notice there exists any new arch
while they will benefit from the full performance of their PC:
http://people.redhat.com/jkratoch/x86bilive-2009062000.tar.gz (71KB)
It uses live_dir=LiveOS-x86_64 vs. live_dir=LiveOS-i686 to boot the image.
The syslinux patch provides "default-{x86_64,i386}" keywords in isolinux.cfg.
livecd-iso-to-disk is not patched/compatible with such image.
livecd-creator should create such ISO on a single run, not by merging the
output of two livecd-creator runs by a 3rd party app.
Regards,
Jan
Reasons:
* I still did not understand why I have to carry with me two media - both
x86_64 and i386 - when all the data perfectly fit on a single media.
* Why I have to try to boot x86_64 first to find out if the specific machine
is x86_64? Even common programmers do not know it, "Windows XP works here".
* The OS must "just work", it must be fun and easy. Requiring a special
technical decision before even starting the OS download is a showstopper.
* Checked that a regular user will on http://fedoraproject.org/get-fedora
still download terrible performance degradation of 32-bit OS although her
hadware is in 70%-95%(?) of cases x86_64. x86_64 is here for 6 years now.
* Arguing x86 may be faster than x86_64... I did not find any such case,
x86_64 is a more modern arch (more registers, PIC for free, better ABI).
We already hit the 2GB address space limitations. x86_64 is the future.
* All the friends of mine have 8Mbit+ ADSL and TB disks downloading many DVD
disks so some several more hundreds of MB are not something to notice.
mkisofs -f -J -r -hide-rr-moved -hide-joliet-trans-tbl -V Fedora-11-x86bi-Live -o
../x86bilive.iso -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot
-boot-info-table -boot-load-size 4 .
mount -r -o loop Fedora-11-x86_64-Live.iso x86_64/
mount -r -o loop Fedora-11-i686-Live.iso i686/
x86bilive:
total 4
lrwxrwxrwx 1 root root 13 2009-06-18 21:10 GPL -> ../x86_64/GPL
lrwxrwxrwx 1 root root 14 2009-06-18 21:11 LiveOS-i686 -> ../i686/LiveOS/
lrwxrwxrwx 1 root root 16 2009-06-18 21:10 LiveOS-x86_64 -> ../x86_64/LiveOS/
lrwxrwxrwx 1 root root 16 2009-06-18 21:10 README -> ../x86_64/README
drwxr-xr-x 2 root root 4096 2009-06-20 21:44 isolinux/
x86bilive/isolinux:
total 184
lrwxrwxrwx 1 root root 30 2009-06-18 21:13 boot.cat ->
../../x86_64/isolinux/boot.cat
lrwxrwxrwx 1 root root 31 2009-06-18 21:17 ii686 ->
../../i686/isolinux/initrd0.img
-rw-r--r-- 1 root root 14336 2009-06-20 21:45 isolinux.bin
-r--r--r-- 1 root root 1411 2009-06-20 21:44 isolinux.cfg
lrwxrwxrwx 1 root root 33 2009-06-18 21:13 ix8664 ->
../../x86_64/isolinux/initrd0.img
lrwxrwxrwx 1 root root 28 2009-06-18 21:17 ki686 -> ../../i686/isolinux/vmlinuz0
lrwxrwxrwx 1 root root 30 2009-06-18 21:13 kx8664 ->
../../x86_64/isolinux/vmlinuz0
lrwxrwxrwx 1 root root 29 2009-06-18 21:13 memtest ->
../../x86_64/isolinux/memtest
lrwxrwxrwx 1 root root 32 2009-06-18 21:13 splash.jpg ->
../../x86_64/isolinux/splash.jpg
-r--r--r-- 1 root root 159888 2009-06-20 20:48 vesamenu.c32
isolinux.cfg:
default vesamenu.c32
timeout 100
menu background splash.jpg
menu title Welcome to Fedora-11-x86bi-Live!
menu color border 0 #ffffffff #00000000
menu color sel 7 #ffffffff #ff000000
menu color title 0 #ffffffff #00000000
menu color tabmsg 0 #ffffffff #00000000
menu color unsel 0 #ffffffff #00000000
menu color hotsel 0 #ff000000 #ffffffff
menu color hotkey 7 #ffffffff #ff000000
menu color timeout_msg 0 #ffffffff #00000000
menu color timeout 0 #ffffffff #00000000
menu color cmdline 0 #ffffffff #00000000
menu hidden
menu hiddenrow 5
label linux0
menu label x86_64 Boot
kernel kx8664
append initrd=ix8664 root=CDLABEL=Fedora-11-x86bi-Live rootfstype=auto
live_dir=LiveOS-x86_64 ro liveimg quiet rhgb
menu default-x86_64
label check0
menu label x86_64 Verify and Boot
kernel kx8664
append initrd=ix8664 root=CDLABEL=Fedora-11-x86bi-Live rootfstype=auto
live_dir=LiveOS-x86_64 ro liveimg quiet rhgb check
label linux1
menu label i686 Boot
kernel ki686
append initrd=ii686 root=CDLABEL=Fedora-11-x86bi-Live rootfstype=auto
live_dir=LiveOS-i686 ro liveimg quiet rhgb
menu default-i386
label check1
menu label i686 Verify and Boot
kernel ki686
append initrd=ii686 root=CDLABEL=Fedora-11-x86bi-Live rootfstype=auto
live_dir=LiveOS-i686 ro liveimg quiet rhgb check
label memtest
menu label Memory Test
kernel memtest
label local
menu label Boot from local drive
localboot 0xffff
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ defaults.patch 20 Jun 2009 20:10:46 -0000
@@ -0,0 +1,175 @@
+--- syslinux-3.75-orig/com32/menu/Makefile 2009-04-16 06:42:14.000000000 +0200
++++ syslinux-3.75-defaults/com32/menu/Makefile 2009-06-20 19:21:32.000000000 +0200
+@@ -24,7 +24,7 @@ MODULES = menu.c32 vesamenu.c32
+ TESTFILES =
+
+ COMMONOBJS = menumain.o readconfig.o passwd.o drain.o printmsg.o colors.o \
+- background.o refstr.o execute.o
++ background.o refstr.o execute.o verify_cpu.o
+
+ all: $(MODULES) $(TESTFILES)
+
+--- syslinux-3.75-orig/com32/menu/menu.h 2009-04-16 06:42:14.000000000 +0200
++++ syslinux-3.75-defaults/com32/menu/menu.h 2009-06-20 19:21:32.000000000 +0200
+@@ -224,4 +224,9 @@ void execute(const char *cmdline, enum k
+ /* drain.c */
+ void drain_keyboard(void);
+
++/* verify_cpu.S */
++#if defined(__i386__) || defined(__x86_64__)
++int verify_cpu(void);
++#endif
++
+ #endif /* MENU_H */
+--- syslinux-3.75-orig/com32/menu/readconfig.c 2009-04-16 06:42:14.000000000 +0200
++++ syslinux-3.75-defaults/com32/menu/readconfig.c 2009-06-20 20:23:24.000000000 +0200
+@@ -656,11 +656,22 @@ static void parse_config_file(FILE *f)
+ m->parent_entry->displayname = refstr_get(m->title);
+ }
+ }
+- } else if ( looking_at(p, "default") ) {
+- if (ld.label) {
+- ld.menudefault = 1;
+- } else if (m->parent_entry) {
+- m->parent->defentry = m->parent_entry->entry;
++ } else if ( looking_at(p, "default")
++#if defined(__i386__) || defined(__x86_64__)
++ || looking_at(p, "default-i386")
++ || looking_at(p, "default-x86_64")
++#endif
++ ) {
++#if defined(__i386__) || defined(__x86_64__)
++ if (! ( ( looking_at(p, "default-i386") && verify_cpu () == 0
) ||
++ ( looking_at(p, "default-x86_64") && verify_cpu () != 0 ) )
)
++#endif
++ {
++ if (ld.label) {
++ ld.menudefault = 1;
++ } else if (m->parent_entry) {
++ m->parent->defentry = m->parent_entry->entry;
++ }
+ }
+ } else if ( looking_at(p, "hide") ) {
+ ld.menuhide = 1;
+--- syslinux-3.75-orig/com32/menu/verify_cpu.S 1970-01-01 01:00:00.000000000 +0100
++++ syslinux-3.75-defaults/com32/menu/verify_cpu.S 2009-06-20 20:30:27.000000000 +0200
+@@ -0,0 +1,119 @@
++/* https://lists.linux-foundation.org/pipermail/fastboot/2006-November/01188... */
++#if defined(__i386__) || defined(__x86_64__)
++ .globl verify_cpu
++verify_cpu:
++ push %ebx
++ push %esi
++ push %edi
++ push %ebp
++ call Xverify_cpu
++ pop %ebp
++ pop %edi
++ pop %esi
++ pop %ebx
++ ret
++
++/*
++ *
++ * verify_cpu.S
++ *
++ * 14 Nov 2006 Vivek Goyal: Created the file
++ *
++ * This is a common code for verification whether CPU supports
++ * long mode and SSE or not. It is not called directly instead this
++ * file is included at various places and compiled in that context.
++ * Following are the current usage.
++ *
++ * This file is included by both 16bit and 32bit code.
++ *
++ * arch/x86_64/boot/setup.S : Boot cpu verification (16bit)
++ * arch/x86_64/boot/compressed/head.S: Boot cpu verification (32bit)
++ * arch/x86_64/kernel/trampoline.S: secondary processor verfication (16bit)
++ * arch/x86_64/kernel/acpi/wakeup.S:Verfication at resume (16bit)
++ *
++ * verify_cpu, returns the status of cpu check in register %eax.
++ * 0: Success 1: Failure
++ *
++ * The caller needs to check for the error code and take the action
++ * appropriately. Either display a message or halt.
++ */
++
++Xverify_cpu:
++
++ pushfl # Save caller passed flags
++ pushl $0 # Kill any dangerous flags
++ popfl
++
++ /* minimum CPUID flags for x86-64 */
++ /* see http://www.x86-64.org/lists/discuss/msg02971.html */
++#define SSE_MASK ((1<<25)|(1<<26))
++#define REQUIRED_MASK1
((1<<0)|(1<<3)|(1<<4)|(1<<5)|(1<<6)|(1<<8)|\
++ (1<<13)|(1<<15)|(1<<24))
++#define REQUIRED_MASK2 (1<<29)
++ pushfl # standard way to check for cpuid
++ popl %eax
++ movl %eax,%ebx
++ xorl $0x200000,%eax
++ pushl %eax
++ popfl
++ pushfl
++ popl %eax
++ cmpl %eax,%ebx
++ jz verify_cpu_no_longmode # cpu has no cpuid
++
++ movl $0x0,%eax # See if cpuid 1 is implemented
++ cpuid
++ cmpl $0x1,%eax
++ jb verify_cpu_no_longmode # no cpuid 1
++
++ xor %di,%di
++ cmpl $0x68747541,%ebx # AuthenticAMD
++ jnz verify_cpu_noamd
++ cmpl $0x69746e65,%edx
++ jnz verify_cpu_noamd
++ cmpl $0x444d4163,%ecx
++ jnz verify_cpu_noamd
++ mov $1,%di # cpu is from AMD
++
++verify_cpu_noamd:
++ movl $0x1,%eax # Does the cpu have what it takes
++ cpuid
++ andl $REQUIRED_MASK1,%edx
++ xorl $REQUIRED_MASK1,%edx
++ jnz verify_cpu_no_longmode
++
++ movl $0x80000000,%eax # See if extended cpuid is implemented
++ cpuid
++ cmpl $0x80000001,%eax
++ jb verify_cpu_no_longmode # no extended cpuid
++
++ movl $0x80000001,%eax # Does the cpu have what it takes
++ cpuid
++ andl $REQUIRED_MASK2,%edx
++ xorl $REQUIRED_MASK2,%edx
++ jnz verify_cpu_no_longmode
++
++verify_cpu_sse_test:
++ movl $1,%eax
++ cpuid
++ andl $SSE_MASK,%edx
++ cmpl $SSE_MASK,%edx
++ je verify_cpu_sse_ok
++ test %di,%di
++ jz verify_cpu_no_longmode # only try to force SSE on AMD
++ movl $0xc0010015,%ecx # HWCR
++ rdmsr
++ btr $15,%eax # enable SSE
++ wrmsr
++ xor %di,%di # don't loop
++ jmp verify_cpu_sse_test # try again
++
++verify_cpu_no_longmode:
++ popfl # Restore caller passed flags
++ movl $1,%eax
++ ret
++verify_cpu_sse_ok:
++ popfl # Restore caller passed flags
++ xorl %eax, %eax
++ ret
++#endif
--- syslinux.spec 16 Apr 2009 18:11:30 -0000 1.59
+++ syslinux.spec 20 Jun 2009 20:10:46 -0000
@@ -7,6 +7,7 @@ License: GPLv2+
Group: Applications/System
URL: http://syslinux.zytor.com/
Source0:
ftp://ftp.kernel.org/pub/linux/utils/boot/syslinux/%{name}-%{tarball_vers...
+Patch9: defaults.patch
ExclusiveArch: %{ix86} x86_64
Buildroot: %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n)
BuildRequires: nasm >= 0.98.38-1, perl, netpbm-progs
@@ -33,12 +34,18 @@ MEMDISK, which loads legacy operating sy
%prep
%setup -q -n syslinux-%{tarball_version}
+%patch9 -p1
%build
CFLAGS="-Werror -Wno-unused -finline-limit=2000"
export CFLAGS
# If you make clean here, we lose the provided syslinux.exe
#make clean
+rm -f com32/menu/vesamenu.c32
+make -C com32/lib libcom32.a
+make -C com32/gpllib libcom32gpl.a
+make -C com32/menu vesamenu.c32
+test -f com32/menu/vesamenu.c32
make installer
make -C sample tidy
12:53 p.m.
On 06/21/2009 12:05 PM, Jan Kratochvil wrote:
Hi,
finally created a LiveDVD ISO automatically booting x86_64 OS on x86_64
(and i686 otherwise). Regular users will not notice there exists any new arch
while they will benefit from the full performance of their PC:
http://people.redhat.com/jkratoch/x86bilive-2009062000.tar.gz (71KB)
It uses live_dir=LiveOS-x86_64 vs. live_dir=LiveOS-i686 to boot the image.
The syslinux patch provides "default-{x86_64,i386}" keywords in isolinux.cfg.
livecd-iso-to-disk is not patched/compatible with such image.
livecd-creator should create such ISO on a single run, not by merging the
output of two livecd-creator runs by a 3rd party app.
It does not make much sense to me, so maybe you can explain it:
Isn't syslinux 3.72 or later itself capable of choosing one or the other
menu entry using ifcpu64.c32?
Would that not be a better option?
-Jeroen
1:35 p.m.
On Sun, 21 Jun 2009 19:53:58 +0200, Jeroen van Meeuwen wrote:
On 06/21/2009 12:05 PM, Jan Kratochvil wrote:
> The syslinux patch provides "default-{x86_64,i386}" keywords in
isolinux.cfg.
...
It does not make much sense to me, so maybe you can explain it:
Isn't syslinux 3.72 or later itself capable of choosing one or the other
menu entry using ifcpu64.c32?
Would that not be a better option?
yes, definitely, my syslinux patch is useless.
I did not expect such option exists when such ISO still has not been released
while waiting for it the recent years.
Thanks for the notice,
Jan
7:51 a.m.
On 06/21/2009 08:35 PM, Jan Kratochvil wrote:
On Sun, 21 Jun 2009 19:53:58 +0200, Jeroen van Meeuwen wrote:
> On 06/21/2009 12:05 PM, Jan Kratochvil wrote:
>> The syslinux patch provides "default-{x86_64,i386}" keywords in
isolinux.cfg.
...
> It does not make much sense to me, so maybe you can explain it:
>
> Isn't syslinux 3.72 or later itself capable of choosing one or the other
> menu entry using ifcpu64.c32?
>
> Would that not be a better option?
yes, definitely, my syslinux patch is useless.
I did not expect such option exists when such ISO still has not been released
while waiting for it the recent years.
So yeah, the problem we get with i386/x86_64 hybrids is more that we
would reduce the available size to either of the Live images to half a
CD or half a DVD. Some of our spins fit on half a DVD, but not half a
CD. If some spins have separate media for each arch, and others do not,
then that may work confusing.
That said though, I don't think we've had this kinda thing discussed or
decided upon... so it's a good point ;-)
-Jeroen
9:22 a.m.
Jeroen van Meeuwen (kanarip(a)kanarip.com) said:
So yeah, the problem we get with i386/x86_64 hybrids is more that we
would reduce the available size to either of the Live images to half a
CD or half a DVD. Some of our spins fit on half a DVD, but not half a
CD. If some spins have separate media for each arch, and others do not,
then that may work confusing.
That said though, I don't think we've had this kinda thing discussed or
decided upon... so it's a good point ;-)
I think it's safe to assume that attempting to fit any 'real' biarch
spin on CD would be a futile exercise. DVD shouldn't be an issue for
the non-Games live spins yet, should it?
Bill
9:47 a.m.
On Tue, 23 Jun 2009 16:22:07 +0200, Bill Nottingham wrote:
DVD shouldn't be an issue for the non-Games live spins yet,
should it?
* FireFox + OpenOffice.org are (IMO) the two most visible F/OSS applications.
* LiveCD is there (also / primarily?) to make a preview of what is F/OSS.
=> The major F/OSS application success is omitted because of ... why?
There ware recent discussions people no longer download CD media; but to say
the truth I do not find a clear consensus.
BitTorrent downloaders of the split CD sets have been decreasing.
F9: 6.32%. F10 4.58% F11 so far: 2.2%.
+
Do we need split media CDs for F12?
https://www.redhat.com/archives/fedora-devel-list/2009-June/thread.html#0...
+
https://fedoraproject.org/wiki/F12_No_Split_CDs_Proposal
Regards,
Jan
11:14 a.m.
On Tue, Jun 23, 2009 at 10:22:07 -0400,
Bill Nottingham <notting(a)redhat.com> wrote:
Jeroen van Meeuwen (kanarip(a)kanarip.com) said:
> So yeah, the problem we get with i386/x86_64 hybrids is more that we
> would reduce the available size to either of the Live images to half a
> CD or half a DVD. Some of our spins fit on half a DVD, but not half a
> CD. If some spins have separate media for each arch, and others do not,
> then that may work confusing.
>
> That said though, I don't think we've had this kinda thing discussed or
> decided upon... so it's a good point ;-)
I think it's safe to assume that attempting to fit any 'real' biarch
spin on CD would be a futile exercise. DVD shouldn't be an issue for
the non-Games live spins yet, should it?
A biarch rescue image would be useful. For some kinds of recoveries you need
to use a rescue image matching the arch of the system you are trying to fix.
I got bit by this during the F11 devel cycle.
7:58 a.m.
On 06/23/2009 06:14 PM, Bruno Wolff III wrote:
On Tue, Jun 23, 2009 at 10:22:07 -0400,
Bill Nottingham<notting(a)redhat.com> wrote:
> Jeroen van Meeuwen (kanarip(a)kanarip.com) said:
>> So yeah, the problem we get with i386/x86_64 hybrids is more that we
>> would reduce the available size to either of the Live images to half a
>> CD or half a DVD. Some of our spins fit on half a DVD, but not half a
>> CD. If some spins have separate media for each arch, and others do not,
>> then that may work confusing.
>>
>> That said though, I don't think we've had this kinda thing discussed or
>> decided upon... so it's a good point ;-)
> I think it's safe to assume that attempting to fit any 'real' biarch
> spin on CD would be a futile exercise. DVD shouldn't be an issue for
> the non-Games live spins yet, should it?
A biarch rescue image would be useful. For some kinds of recoveries you need
to use a rescue image matching the arch of the system you are trying to fix.
I got bit by this during the F11 devel cycle.
This would actually fit on a CD, too. I'm interested in exploring this.
Would this most suitably fit within the Fedora Project as a spin, or
maybe something else (a Fedora 12 Feature??)?
-Jeroen
10:25 a.m.
New subject: auto-biarch (x86_64 + i686) LiveUSBFlash [Re: auto-biarch (x86_64 + i686) LiveDVD patch + ISO]
Hi,
followup on:
https://www.redhat.com/archives/fedora-livecd-list/2009-June/msg00018.html
On Sun, 21 Jun 2009 19:53:58 +0200, Jeroen van Meeuwen wrote:
Isn't syslinux 3.72 or later itself capable of choosing one or
the
other menu entry using ifcpu64.c32?
updated it to use ifcpu64.c32, thanks for the notice.
It is currently tested only with F-12 (updated) but it should work with any
release. It can take Fedora-12-i686-Live.iso + Fedora-12-x86_64-Live.iso or
it can build two such from hypothetical custom Fedora-12-LiveCD.ks.
The previous version created biarch DVD. But one could not easily convert it
to USBFlash as livecd-iso-to-disk expects a single directory on the CD/DVD.
As CD/DVD are IMO no longer useful as bootable media this script converts:
uni-arch x86_64 CD/DVD + uni-arch i686 CD/DVD -> bi-arch x86_64+i686 USBFlash
Created it now just for my personal needs; there is missing:
* Some proper integration with livecd tools. Either
* livecd-creator should be able to create biarch CD/DVDs AND
livecd-iso-to-disk should convert such biarch CD/DVDs into USBFlashes.
or
* livecd-iso-to-disk should create biarch USBFlash from two uniarch CD/DVDs.
or
* livecd-iso-to-disk should be able to update uniarch USBFlash with second
uniarch CD/DVD creating biarch USBFlash (=reimplementing this script).
* It does not have EFI. Unaware how to test it and when it is needed for x86*.
* It is an ugly hack.
Some people may think Fedora should stop producing i686.
Unfortunately I still know about two boxes running 32bit only CPUs.
(Running i686 on x86_64 box has no technical benefits; besides some easier
compatibility with proprietary i686 binaries as I heard about.)
Regards,
Jan
2:02 a.m.
New subject: mkbiarch state [auto-biarch (x86_64 + i686) LiveUSBFlash]
Hi,
I finally could check the output of mkbiarch, which I believed should be the
upstreamed variant of "auto-biarch" script posted by me:
http://lists.fedoraproject.org/pipermail/livecd/2010-March/005755.html
But it does not serve for the primary purpose at all: Normal user just not
touching the boot on an x86_64 box will still get booted the crippled i686 OS.
I have checked It does not use ifcpu64.c32 at all so it cannot work.
It also does not provide the "Basic Video" / "Memory Test" etc.
options.
It also does not provide an option for a fat filesystem on the flash media,
together with fitting/specifying media size for an available free space there.
It is neither on-par with the original media nor on-par with "auto-biarch".
Should I convert the "auto-biarch" script into Python or try to merge the
features into existing /usr/bin/mkbiarch to get it accepted? This way I am
not sure if it still be in time for a meaningful official F15 bi-arch media.
Thanks,
Jan
4:49 p.m.
New subject: auto-biarch update for F14
On Thu, 30 Dec 2010 09:02:48 +0100, Jan Kratochvil wrote:
But it does not serve for the primary purpose at all: Normal user
just not
touching the boot on an x86_64 box will still get booted the crippled i686 OS.
I have checked It does not use ifcpu64.c32 at all so it cannot work.
It also does not provide the "Basic Video" / "Memory Test" etc.
options.
It also does not provide an option for a fat filesystem on the flash media,
together with fitting/specifying media size for an available free space there.
Attaching an updated shell script with these features. Tested with:
livecd-tools-14.0-1.fc14.x86_64
Should I convert the "auto-biarch" script into Python or
try to merge the
features into existing /usr/bin/mkbiarch to get it accepted?
BTW I would prefer Perl which is not available on some LiveCDs but it should
be always available on the build system. I do not know if/why EFI support
would be needed for such boot-flash. GPT should not be needed on boot-flash.
Thanks,
Jan
5:44 p.m.
New subject: auto-biarch update for F14 [fix]
[ Sorry, one fatality fix. ]
#! /bin/bash
# $Id: OOo-LiveCD.sh,v 1.5 2010/12/30 23:34:08 jkratoch Exp $
set -ex
SIZE=2021654528
BASE=f14-live
BIN=$BASE.bin
ISO64=$BASE-x86_64.iso
ISO32=$BASE-i686.iso
ISO=$BASE.iso
KS=$BASE.ks
SELF=$BASE.sh
# $PWD: rpm: arguments to --root (-r) must begin with a /
MNT=$PWD/$BASE.mnt
LOOPDEV=
set -o pipefail
renice +19 -p $$
ionice -c3 -p $$
test "`uname -m`" = "x86_64"
test ! -e $BIN
test ! -e $ISO
test ! -e $MNT
test -e $SELF
if [ ! -e $ISO64 ];then
test -e "$KS"
livecd-creator --config=$KS --fslabel=$BASE --cache=/var/cache/live
test -e $ISO
mv -f $ISO $ISO64
fi
if [ ! -e $ISO32 ];then
test -e "$KS"
setarch i686 livecd-creator --config=$KS --fslabel=$BASE --cache=/var/cache/live
test -e $ISO
mv -f $ISO $ISO32
fi
trap 'set +e; rm -f $BIN; (losetup -d $LOOPDEV; umount $MNT/{c,b,a,}; rmdir
$MNT/{c,b,a,}) 2>/dev/null' EXIT
dd if=/dev/zero of=$BIN bs=1 seek=$[SIZE-1] count=1
/sbin/parted --script $BIN mklabel msdos
dd bs=1 if=/usr/share/syslinux/mbr.bin of=$BIN conv=notrunc
# 17408=34*512
/sbin/parted --script $BIN unit b mkpart primary fat32 17408 $[$(find $BIN -printf
%s)-17408] set 1 boot on
LOOPDEV=`losetup -o 17408 -v -f $BIN | sed 's/^Loop device is //'`
[ -n "$LOOPDEV" -a "${LOOPDEV#/dev/loop[0-9]}" != "$LOOPDEV"
]
/sbin/mkdosfs -n LIVE $LOOPDEV
UUID="$(dd skip=$[0x43] bs=1 count=4 if=$LOOPDEV|od -An -txI|tr a-z A-Z|sed 's/^
\(....\)\(....\)$/\1-\2/')"
# --overlay-size-mb 100 --home-size-mb 100
echo | livecd-iso-to-disk --unencrypted-home --livedir $BASE-i686 $ISO32 $LOOPDEV
mkdir $MNT
mount $LOOPDEV $MNT
CFG="`cat $MNT/syslinux/syslinux.cfg | sed 's/\( root=live:UUID=\)....-....
/\1'"$UUID /"`"
mv $MNT/syslinux/initrd0.img $MNT/syslinux/initrd1.img
mv $MNT/syslinux/vmlinuz0 $MNT/syslinux/vmlinuz1
umount $MNT
livecd-iso-to-disk --unencrypted-home --livedir $BASE-x86_64 $ISO64 $LOOPDEV
mount $LOOPDEV $MNT
cp -p /usr/share/syslinux/ifcpu64.c32 $MNT/syslinux/
(
echo "$CFG" | sed -n '1,/^menu hiddenrow/p'
cat <<EOH
label detect
menu label Autoselect x86_64 / i686 Boot
kernel ifcpu64.c32
append linux0 -- linux1
menu default
EOH
echo "$CFG" | sed -n -e "s/$BASE-i686/$BASE-x86_64/g"
-e 's/^ menu label .*$/& x86_64/' -e '/^label.*0$/,/^
append/p'
echo "$CFG" | sed -n -e 's/^\(label.*\)0$/\11/' -e
's/\(vmlinuz\|initrd\)0/\11/g' -e 's/^ menu label .*$/& i686/' -e
'/^label.*1$/,/^ append/p'
echo "$CFG" | sed -n '/^label mem/,$p'
) >$MNT/syslinux/syslinux.cfg
cp -p $SELF $MNT/
if [ -e "$KS" ];then
cp -p $KS $MNT/
fi
umount $MNT
mkdir $MNT/{a,b,c}
mount $LOOPDEV $MNT/a
mount -r -o loop $MNT/a/$BASE-x86_64/squashfs.img $MNT/b
mount -r -o loop $MNT/b/LiveOS/ext3fs.img $MNT/c
rpm -r $MNT/c -qa|sort >$MNT/a/$BASE-x86_64/rpm-qa
umount $MNT/{c,b}
mount -r -o loop $MNT/a/$BASE-i686/squashfs.img $MNT/b
mount -r -o loop $MNT/b/LiveOS/ext3fs.img $MNT/c
rpm -r $MNT/c -qa|sort >$MNT/a/$BASE-i686/rpm-qa
diff -u <(sed 's/\.x86_64$//' <$MNT/a/$BASE-x86_64/rpm-qa) <(sed
's/\.i686$//' <$MNT/a/$BASE-i686/rpm-qa) || :
BIN=
echo OK
10:21 p.m.
New subject: auto-biarch update for F14 [fixes#2]
[ Some minor fixes, no longer hacking it now so this one should be final. ]
#! /bin/bash
# $Id: OOo-LiveCD.sh,v 1.6 2010/12/31 21:58:53 jkratoch Exp $
set -ex
SIZE=2021654528
BASE=f14-live
BIN=$BASE.bin
ISO64=$BASE-x86_64.iso
ISO32=$BASE-i686.iso
ISO=$BASE.iso
KS=$BASE.ks
SELF=$BASE.sh
# $PWD: rpm: arguments to --root (-r) must begin with a /
MNT=$PWD/$BASE.mnt
LOOPDEV=
set -o pipefail
renice +19 -p $$
ionice -c3 -p $$
test "`uname -m`" = "x86_64"
test ! -e $BIN
test ! -e $ISO
test ! -e $MNT
test -e $SELF
if [ ! -e $ISO64 ];then
test -e "$KS"
livecd-creator --config=$KS --fslabel=$BASE --cache=/var/cache/live
test -e $ISO
mv -f $ISO $ISO64
fi
if [ ! -e $ISO32 ];then
test -e "$KS"
setarch i686 livecd-creator --config=$KS --fslabel=$BASE --cache=/var/cache/live
test -e $ISO
mv -f $ISO $ISO32
fi
trap 'set +e; rm -f $BIN; (umount $MNT/{c,b,a,}; rmdir $MNT/{c,b,a,}; losetup -d
$LOOPDEV) 2>/dev/null' EXIT
dd if=/dev/zero of=$BIN bs=1 seek=$[SIZE-1] count=1
/sbin/parted --script $BIN mklabel msdos
dd bs=1 if=/usr/share/syslinux/mbr.bin of=$BIN conv=notrunc
# 17408=34*512
/sbin/parted --script $BIN unit b mkpart primary fat32 17408 $[$(find $BIN -printf
%s)-17408] set 1 boot on
LOOPDEV=`losetup -o 17408 -v -f $BIN | sed 's/^Loop device is //'`
[ -n "$LOOPDEV" -a "${LOOPDEV#/dev/loop[0-9]}" != "$LOOPDEV"
]
/sbin/mkdosfs -n LIVE $LOOPDEV
UUID="$(dd skip=$[0x43] bs=1 count=4 if=$LOOPDEV|od -An -txI|tr a-z A-Z|sed 's/^
\(....\)\(....\)$/\1-\2/')"
# --overlay-size-mb 100 --home-size-mb 100
echo | livecd-iso-to-disk --unencrypted-home --livedir $BASE-i686 $ISO32 $LOOPDEV
mkdir $MNT
mount $LOOPDEV $MNT
CFG="`cat $MNT/syslinux/syslinux.cfg | sed 's/\( root=live:UUID=\)....-....
/\1'"$UUID /"`"
mv $MNT/syslinux/initrd0.img $MNT/syslinux/initrd1.img
mv $MNT/syslinux/vmlinuz0 $MNT/syslinux/vmlinuz1
umount $MNT
livecd-iso-to-disk --unencrypted-home --livedir $BASE-x86_64 $ISO64 $LOOPDEV
mount $LOOPDEV $MNT
cp -p /usr/share/syslinux/ifcpu64.c32 $MNT/syslinux/
(
echo "$CFG" | sed -n '1,/^menu hiddenrow/p'
cat <<EOH
label detect
menu label Autoselect x86_64 / i686 Boot
kernel ifcpu64.c32
append linux0 -- linux1
menu default
EOH
echo "$CFG" | sed -n -e "s/$BASE-i686/$BASE-x86_64/g"
-e 's/^ menu label .*$/& x86_64/' -e '/^label.*0$/,/^
append/p'
echo "$CFG" | sed -n -e 's/^\(label.*\)0$/\11/' -e
's/\(vmlinuz\|initrd\)0/\11/g' -e 's/^ menu label .*$/& i686/' -e
'/^label.*1$/,/^ append/p'
echo "$CFG" | sed -n '/^label mem/,$p'
) >$MNT/syslinux/syslinux.cfg
cp -p $SELF $MNT/
if [ -e "$KS" ];then
cp -p $KS $MNT/
fi
umount $MNT
mkdir $MNT/{a,b,c}
mount $LOOPDEV $MNT/a
mount -r -o loop $MNT/a/$BASE-x86_64/squashfs.img $MNT/b
mount -r -o loop $MNT/b/LiveOS/ext3fs.img $MNT/c
rpm -r $MNT/c -qa|sort >$MNT/a/$BASE-x86_64/rpm-qa
umount $MNT/{c,b}
mount -r -o loop $MNT/a/$BASE-i686/squashfs.img $MNT/b
mount -r -o loop $MNT/b/LiveOS/ext3fs.img $MNT/c
rpm -r $MNT/c -qa|sort >$MNT/a/$BASE-i686/rpm-qa
diff -u <(sed 's/\.x86_64$//' <$MNT/a/$BASE-x86_64/rpm-qa) <(sed
's/\.i.86$//' <$MNT/a/$BASE-i686/rpm-qa) || :
BIN=
echo OK
4857
days inactive
5417
days old
livecd@lists.fedoraproject.org
12 comments
4 participants
participants (4)
-
Bill Nottingham -
Bruno Wolff III -
Jan Kratochvil -
Jeroen van Meeuwen