[arch-releng] [PATCH 1/3] [syslinux-iso] Comment IPAPPEND option and add comment about usage.
IPAPPEND option is only used by PXELINUX. Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar> --- archiso2dual/isolinux.cfg | 6 ++++-- .../syslinux-iso/boot-files/isolinux/isolinux.cfg | 3 ++- 2 files changed, 6 insertions(+), 3 deletions(-) diff --git a/archiso2dual/isolinux.cfg b/archiso2dual/isolinux.cfg index 5ed26e1..6dd4bbb 100644 --- a/archiso2dual/isolinux.cfg +++ b/archiso2dual/isolinux.cfg @@ -39,7 +39,8 @@ ENDTEXT MENU LABEL Boot Arch Linux (i686) KERNEL /boot/i686/vmlinuz26 APPEND initrd=/boot/i686/archiso.img archisolabel=XXX tmpfs_size=75% locale=en_US.UTF-8 -IPAPPEND 0 +# Next line should be uncommented when prepare enviroment to boot from PXE. +#IPAPPEND 3 LABEL arch64 TEXT HELP @@ -49,7 +50,8 @@ ENDTEXT MENU LABEL Boot Arch Linux (x86_64) KERNEL /boot/x86_64/vmlinuz26 APPEND initrd=/boot/x86_64/archiso.img archisolabel=XXX tmpfs_size=75% locale=en_US.UTF-8 -IPAPPEND 0 +# Next line should be uncommented when prepare enviroment to boot from PXE. +#IPAPPEND 3 LABEL existing TEXT HELP diff --git a/configs/syslinux-iso/boot-files/isolinux/isolinux.cfg b/configs/syslinux-iso/boot-files/isolinux/isolinux.cfg index 2d604ba..dd9ef86 100644 --- a/configs/syslinux-iso/boot-files/isolinux/isolinux.cfg +++ b/configs/syslinux-iso/boot-files/isolinux/isolinux.cfg @@ -39,7 +39,8 @@ ENDTEXT MENU LABEL Boot Arch Linux KERNEL /boot/vmlinuz26 APPEND initrd=/boot/archiso.img archisolabel=XXX tmpfs_size=75% locale=en_US.UTF-8 -IPAPPEND 0 +# Next line should be uncommented when prepare enviroment to boot from PXE. +#IPAPPEND 3 LABEL existing TEXT HELP -- 1.7.1
Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar> --- configs/syslinux-iso/packages.i686 | 7 +------ configs/syslinux-iso/packages.x86_64 | 7 +------ 2 files changed, 2 insertions(+), 12 deletions(-) diff --git a/configs/syslinux-iso/packages.i686 b/configs/syslinux-iso/packages.i686 index 36ae337..6833707 100644 --- a/configs/syslinux-iso/packages.i686 +++ b/configs/syslinux-iso/packages.i686 @@ -13,13 +13,10 @@ gnu-netcat inetutils ipw2100-fw ipw2200-fw -iwlwifi-3945-ucode -iwlwifi-4965-ucode -iwlwifi-5000-ucode -iwlwifi-5150-ucode joe lftp lilo +linux-firmware memtest86+ ndiswrapper ndiswrapper-utils @@ -32,8 +29,6 @@ openssh openvpn parted pptpclient -rt2x00-rt61-fw -rt2x00-rt71w-fw speedtouch squashfs-tools syslinux diff --git a/configs/syslinux-iso/packages.x86_64 b/configs/syslinux-iso/packages.x86_64 index 36ae337..6833707 100644 --- a/configs/syslinux-iso/packages.x86_64 +++ b/configs/syslinux-iso/packages.x86_64 @@ -13,13 +13,10 @@ gnu-netcat inetutils ipw2100-fw ipw2200-fw -iwlwifi-3945-ucode -iwlwifi-4965-ucode -iwlwifi-5000-ucode -iwlwifi-5150-ucode joe lftp lilo +linux-firmware memtest86+ ndiswrapper ndiswrapper-utils @@ -32,8 +29,6 @@ openssh openvpn parted pptpclient -rt2x00-rt61-fw -rt2x00-rt71w-fw speedtouch squashfs-tools syslinux -- 1.7.1
Latest init script mounts a /dev FS, so /dev/loop0 will disappear. Create loop0 on /lib/udev/devices so udev copy to /dev on startup. Signed-off-by: Gerardo Exequiel Pozzi <vmlinuz386@yahoo.com.ar> --- archiso/install/archiso | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/archiso/install/archiso b/archiso/install/archiso index ad1c1d0..9b45080 100644 --- a/archiso/install/archiso +++ b/archiso/install/archiso @@ -14,7 +14,7 @@ install () add_dir /tmpfs.sqfs add_dir /bootmnt - add_device /dev/loop0 b 7 0 + add_device /lib/udev/devices/loop0 b 7 0 add_binary /lib/udev/cdrom_id -- 1.7.1
Patches applied, plus a few more fixes, please review them as well.
On Wed, 23 Jun 2010 10:12:31 +0200 Thomas Bächler <thomas@archlinux.org> wrote:
Patches applied, plus a few more fixes, please review them as well.
thanks Thomas, and Gerardo, of course. Dieter
On 06/23/2010 05:12 AM, Thomas Bächler wrote:
Patches applied, plus a few more fixes, please review them as well.
Many thanks. OK. Tested (using mkinitcpio.git and syslinux 4.00-pre60 (not in repos)) * Build OK * Boot normal OK * Boot from PXE+NBD OK * Boot from MEMDISK OK * Boot from PXE+MEMDISK OK Warning: in syslinux-4.00-pre50 memdiskfind command is missing, I open a FS if still missing in next package. -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
On 06/25/2010 05:38 PM, Gerardo Exequiel Pozzi wrote:
On 06/23/2010 05:12 AM, Thomas Bächler wrote:
Patches applied, plus a few more fixes, please review them as well.
Many thanks.
OK. Tested (using mkinitcpio.git and syslinux 4.00-pre60 (not in repos))
* Build OK * Boot normal OK * Boot from PXE+NBD OK * Boot from MEMDISK OK * Boot from PXE+MEMDISK OK
Warning: in syslinux-4.00-pre50 memdiskfind command is missing, I open a FS if still missing in next package.
btw * MEMDISK was tested directly using: qemu-system-x86_64 -m 512 -kernel /tmp/archiso-tftp/memdisk -initrd /tmp/archiso-tftp/boot/archlinux.iso -append "archisolabel=ARCH_201006" * PXE+MEMDISK (add a menu entry to /tmp/archiso-tftp/pxelinux.cfg/default like this) LABEL memdisk MENU LABEL memdisk LINUX memdisk INITRD /boot/archlinux.iso APPEND iso then booting via PXE any machine. -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
Am 25.06.2010 22:50, schrieb Gerardo Exequiel Pozzi:
OK. Tested (using mkinitcpio.git and syslinux 4.00-pre60 (not in repos))
* Build OK * Boot normal OK * Boot from PXE+NBD OK * Boot from MEMDISK OK * Boot from PXE+MEMDISK OK
Wow, many thanks.
Warning: in syslinux-4.00-pre50 memdiskfind command is missing, I open a FS if still missing in next package.
Yes, it didn't exist in that version. And in a later version, it was broken. We'll have syslinux 4.00 on Monday if all goes to plan, so I can move that to extra then.
btw
* MEMDISK was tested directly using: qemu-system-x86_64 -m 512 -kernel /tmp/archiso-tftp/memdisk -initrd /tmp/archiso-tftp/boot/archlinux.iso -append "archisolabel=ARCH_201006"
WTF, that didn't work for me ... well, my "initrd" was a 700MB hard disk image and no a 160MB ISO. Still, it simply hung and didn't continue.
* PXE+MEMDISK (add a menu entry to /tmp/archiso-tftp/pxelinux.cfg/default like this)
LABEL memdisk MENU LABEL memdisk LINUX memdisk INITRD /boot/archlinux.iso APPEND iso
then booting via PXE any machine.
You should mention that iputils' tftpd has a 16MB size limitation! You need to use either tftp-hpa or dnsmasq. You can probably also use qemu's builtin tftpd. You can also use gpxelinux instead of pxelinux and then download the ISO via ftp or http (gpxelinux will accept INITRD http://...)!
On 06/26/2010 06:52 AM, Thomas Bächler wrote: > Am 25.06.2010 22:50, schrieb Gerardo Exequiel Pozzi: >>> OK. Tested (using mkinitcpio.git and syslinux 4.00-pre60 (not in repos)) >>> >>> * Build OK >>> * Boot normal OK >>> * Boot from PXE+NBD OK >>> * Boot from MEMDISK OK >>> * Boot from PXE+MEMDISK OK > Wow, many thanks. No problem :) >>> Warning: in syslinux-4.00-pre50 memdiskfind command is missing, I open >>> a FS if still missing in next package. > Yes, it didn't exist in that version. And in a later version, it was > broken. We'll have syslinux 4.00 on Monday if all goes to plan, so I can > move that to extra then. OK >> * PXE+MEMDISK (add a menu entry to >> /tmp/archiso-tftp/pxelinux.cfg/default like this) >> >> LABEL memdisk >> MENU LABEL memdisk >> LINUX memdisk >> INITRD /boot/archlinux.iso >> APPEND iso >> >> then booting via PXE any machine. > You should mention that iputils' tftpd has a 16MB size limitation! You > need to use either tftp-hpa or dnsmasq. You can probably also use qemu's > builtin tftpd. Yes I am using dnsmasq tftp server. About qemu internal tftpd does not work, hung after downloaded some MBs. > You can also use gpxelinux instead of pxelinux and then download the ISO > via ftp or http (gpxelinux will accept INITRD http://...)! > True :) It seems that using pxelinux.0 also works with http://, but under qemu (that uses GPXE rom). -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
Am 25.06.2010 22:50, schrieb Gerardo Exequiel Pozzi:
OK. Tested (using mkinitcpio.git and syslinux 4.00-pre60 (not in repos))
* Build OK * Boot normal OK * Boot from PXE+NBD OK * Boot from MEMDISK OK * Boot from PXE+MEMDISK OK
Warning: in syslinux-4.00-pre50 memdiskfind command is missing, I open a FS if still missing in next package.
I cannot boot the image via CD. udev does not create /dev/disk/ symlinks for CD devices (any more?). That's why we had the archiso_early hook in the first place. How did you make yours boot? We need an extra udev rule for this to work as far as I can see, otherwise udev won't scan CDs. memdisk works fine though.
On 06/26/2010 05:22 PM, Thomas Bächler wrote:
OK. Tested (using mkinitcpio.git and syslinux 4.00-pre60 (not in repos))
* Build OK * Boot normal OK * Boot from PXE+NBD OK * Boot from MEMDISK OK * Boot from PXE+MEMDISK OK
Warning: in syslinux-4.00-pre50 memdiskfind command is missing, I open a FS if still missing in next package. I cannot boot the image via CD. udev does not create /dev/disk/ symlinks for CD devices (any more?). That's why we had the archiso_early hook in
Am 25.06.2010 22:50, schrieb Gerardo Exequiel Pozzi: the first place. How did you make yours boot? We need an extra udev rule for this to work as far as I can see, otherwise udev won't scan CDs.
memdisk works fine though.
oops, only tested booting as disk. Yes, booting as cdrom fails. The issue is that 60-persistent-storage.rules that does not match cdrom devices like in previous version of udev, so blkid is not executed. The problem is on cdrom_id helper program, is does not dump some data for udev rules ( ID_CDROM_MEDIA_SESSION_NEXT=, ID_CDROM_MEDIA_SESSION_COUNT=, ID_CDROM_MEDIA_TRACK_COUNT= ID_CDROM_MEDIA_TRACK_COUNT_DATA= ) -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
On 06/26/2010 06:03 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 05:22 PM, Thomas Bächler wrote:
I cannot boot the image via CD. udev does not create /dev/disk/ symlinks for CD devices (any more?). That's why we had the archiso_early hook in the first place. How did you make yours boot? We need an extra udev rule for this to work as far as I can see, otherwise udev won't scan CDs.
memdisk works fine though.
oops, only tested booting as disk. Yes, booting as cdrom fails.
The issue is that 60-persistent-storage.rules that does not match cdrom devices like in previous version of udev, so blkid is not executed. The problem is on cdrom_id helper program, is does not dump some data for udev rules ( ID_CDROM_MEDIA_SESSION_NEXT=, ID_CDROM_MEDIA_SESSION_COUNT=, ID_CDROM_MEDIA_TRACK_COUNT= ID_CDROM_MEDIA_TRACK_COUNT_DATA= )
I tested on real machine and booting from cdrom works OK. Some tests: 1) boot current instllation media 2) /lib/udev/cdrom_id --debug /dev/sr0 and works OK 3) update to latest udev 4) execute (2) then FAIL # /lib/udev/cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: size of features buffer 0x0010 cd_profiles: GET CONFIGURATION: feature 'profiles', with 2 entries feature_profiles: profile 0x10 dvd_rom feature_profiles: profile 0x08 cd_rom cd_profiles: current profile 0x08 cd_profiles: profile 0x08 media_cd_rom info_scsi_cmd_err: READ DISC INFORMATION failed with SK=5h/ASC=20h/ACQ=00h ID_CDROM=1 ID_CDROM_CD=1 ID_CDROM_DVD=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1 This is how looks correctly # /lib/udev/cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: number of profiles 20 cd_profiles: current profile 0x08 cd_media_toc: READ TOC: len: 20 cd_media_toc: track=1 info=0x4(data) start_block=0 cd_media_toc: last track 1 starts at block 0 cd_media_info: disk type d8 ID_CDROM=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1 ID_CDROM_MEDIA_SESSION_NEXT=38876 ID_CDROM_MEDIA_SESSION_COUNT=15884 ID_CDROM_MEDIA_TRACK_COUNT=48900 ID_CDROM_MEDIA_TRACK_COUNT_DATA=1 -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
On 06/26/2010 06:41 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 06:03 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 05:22 PM, Thomas Bächler wrote:
I cannot boot the image via CD. udev does not create /dev/disk/ symlinks for CD devices (any more?). That's why we had the archiso_early hook in the first place. How did you make yours boot? We need an extra udev rule for this to work as far as I can see, otherwise udev won't scan CDs.
memdisk works fine though.
oops, only tested booting as disk. Yes, booting as cdrom fails.
The issue is that 60-persistent-storage.rules that does not match cdrom devices like in previous version of udev, so blkid is not executed. The problem is on cdrom_id helper program, is does not dump some data for udev rules ( ID_CDROM_MEDIA_SESSION_NEXT=, ID_CDROM_MEDIA_SESSION_COUNT=, ID_CDROM_MEDIA_TRACK_COUNT= ID_CDROM_MEDIA_TRACK_COUNT_DATA= )
I tested on real machine and booting from cdrom works OK. And also tested under Virtualbox and also works fine booting from cdrom (iso). So if you testwas on qemu, I guess that something is wrong with new cdrom_id (there are several changes) and qemu.
http://git.kernel.org/?p=linux/hotplug/udev.git;a=history;f=extras/cdrom_id/...
Some tests:
1) boot current instllation media 2) /lib/udev/cdrom_id --debug /dev/sr0 and works OK 3) update to latest udev 4) execute (2) then FAIL
# /lib/udev/cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: size of features buffer 0x0010 cd_profiles: GET CONFIGURATION: feature 'profiles', with 2 entries feature_profiles: profile 0x10 dvd_rom feature_profiles: profile 0x08 cd_rom cd_profiles: current profile 0x08 cd_profiles: profile 0x08 media_cd_rom info_scsi_cmd_err: READ DISC INFORMATION failed with SK=5h/ASC=20h/ACQ=00h ID_CDROM=1 ID_CDROM_CD=1 ID_CDROM_DVD=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1
This is how looks correctly # /lib/udev/cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: number of profiles 20 cd_profiles: current profile 0x08 cd_media_toc: READ TOC: len: 20 cd_media_toc: track=1 info=0x4(data) start_block=0 cd_media_toc: last track 1 starts at block 0 cd_media_info: disk type d8 ID_CDROM=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1 ID_CDROM_MEDIA_SESSION_NEXT=38876 ID_CDROM_MEDIA_SESSION_COUNT=15884 ID_CDROM_MEDIA_TRACK_COUNT=48900 ID_CDROM_MEDIA_TRACK_COUNT_DATA=1
-- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
On 06/26/2010 06:53 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 06:41 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 06:03 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 05:22 PM, Thomas Bächler wrote:
I cannot boot the image via CD. udev does not create /dev/disk/ symlinks for CD devices (any more?). That's why we had the archiso_early hook in the first place. How did you make yours boot? We need an extra udev rule for this to work as far as I can see, otherwise udev won't scan CDs.
memdisk works fine though.
oops, only tested booting as disk. Yes, booting as cdrom fails.
The issue is that 60-persistent-storage.rules that does not match cdrom devices like in previous version of udev, so blkid is not executed. The problem is on cdrom_id helper program, is does not dump some data for udev rules ( ID_CDROM_MEDIA_SESSION_NEXT=, ID_CDROM_MEDIA_SESSION_COUNT=, ID_CDROM_MEDIA_TRACK_COUNT= ID_CDROM_MEDIA_TRACK_COUNT_DATA= )
I tested on real machine and booting from cdrom works OK. And also tested under Virtualbox and also works fine booting from cdrom (iso). So if you testwas on qemu, I guess that something is wrong with new cdrom_id (there are several changes) and qemu.
http://git.kernel.org/?p=linux/hotplug/udev.git;a=history;f=extras/cdrom_id/...
Some tests:
1) boot current instllation media 2) /lib/udev/cdrom_id --debug /dev/sr0 and works OK 3) update to latest udev 4) execute (2) then FAIL
# /lib/udev/cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: size of features buffer 0x0010 cd_profiles: GET CONFIGURATION: feature 'profiles', with 2 entries feature_profiles: profile 0x10 dvd_rom feature_profiles: profile 0x08 cd_rom cd_profiles: current profile 0x08 cd_profiles: profile 0x08 media_cd_rom info_scsi_cmd_err: READ DISC INFORMATION failed with SK=5h/ASC=20h/ACQ=00h ID_CDROM=1 ID_CDROM_CD=1 ID_CDROM_DVD=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1
This is how looks correctly # /lib/udev/cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: number of profiles 20 cd_profiles: current profile 0x08 cd_media_toc: READ TOC: len: 20 cd_media_toc: track=1 info=0x4(data) start_block=0 cd_media_toc: last track 1 starts at block 0 cd_media_info: disk type d8 ID_CDROM=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1 ID_CDROM_MEDIA_SESSION_NEXT=38876 ID_CDROM_MEDIA_SESSION_COUNT=15884 ID_CDROM_MEDIA_TRACK_COUNT=48900 ID_CDROM_MEDIA_TRACK_COUNT_DATA=1
OK, the problem is in this commit [#1], reverting back fixes the issue, ID_CDROM_MEDIA_TRACK_COUNT_DATA appears again then udev rule work again. Another way is reverting this commit [#2], error is reported but ID_CDROM_MEDIA_TRACK_COUNT_DATA is returned. # ./cdrom_id --debug /dev/sr0 main: probing: '/dev/sr0' cd_inquiry: INQUIRY: [QEMU ][QEMU DVD-ROM ][0.12] cd_profiles: GET CONFIGURATION: size of features buffer 0x0010 cd_profiles: GET CONFIGURATION: feature 'profiles', with 2 entries feature_profiles: profile 0x10 dvd_rom feature_profiles: profile 0x08 cd_rom cd_profiles: current profile 0x08 cd_profiles: profile 0x08 media_cd_rom cd_media_info: disk type 00 cd_media_toc: READ TOC: len: 20 cd_media_toc: track=1 info=0x4(data) start_block=0 cd_media_toc: last track 1 starts at block 0 ID_CDROM=1 ID_CDROM_CD=1 ID_CDROM_DVD=1 ID_CDROM_MRW=1 ID_CDROM_MRW_W=1 ID_CDROM_MEDIA=1 ID_CDROM_MEDIA_CD=1 ID_CDROM_MEDIA_TRACK_COUNT_DATA=1 [#1] http://git.kernel.org/?p=linux/hotplug/udev.git;a=commitdiff;h=58e178894bfc0... [#2] http://git.kernel.org/?p=linux/hotplug/udev.git;a=commitdiff;h=30e3b1a0d3a3e... -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
Am 27.06.2010 00:46, schrieb Gerardo Exequiel Pozzi:
I tested on real machine and booting from cdrom works OK. And also tested under Virtualbox and also works fine booting from cdrom (iso). So if you testwas on qemu, I guess that something is wrong with new cdrom_id (there are several changes) and qemu.
Yes, I tested on qemu. A simple udev rule like this one in the initramfs will also do the trick: ACTION=="add|change", SUBSYSTEM=="block", KERNEL=="sr*", GOTO="cdboot_start" GOTO="cdboot_end" LABEL="cdboot_start" IMPORT{program}="/sbin/blkid -o udev -p $tempnode" ENV{ID_FS_USAGE}=="filesystem|other|crypto", ENV{ID_FS_UUID_ENC}=="?*", SYMLINK+="disk/by-uuid/$env{ID_FS_UUID_ENC}" ENV{ID_FS_USAGE}=="filesystem|other", ENV{ID_FS_LABEL_ENC}=="?*", SYMLINK+="disk/by-label/$env{ID_FS_LABEL_ENC}" LABEL="cdboot_end" About memdisk: It works fine on 64 Bit, but not on 32: http://bsfh.ath.cx/memdisk_32.png Appending vmalloc=256M to the kernel command line fixes that, but it is apparently not without side effects - should be good enough for the ISO though. I am uploading ISOs that work to http://alberich.archlinux.org/misc/testiso/, they should be there in about an hour.
On 06/26/2010 08:14 PM, Thomas Bächler wrote:
Am 27.06.2010 00:46, schrieb Gerardo Exequiel Pozzi:
I tested on real machine and booting from cdrom works OK. And also tested under Virtualbox and also works fine booting from cdrom (iso). So if you testwas on qemu, I guess that something is wrong with new cdrom_id (there are several changes) and qemu. Yes, I tested on qemu. A simple udev rule like this one in the initramfs will also do the trick:
ACTION=="add|change", SUBSYSTEM=="block", KERNEL=="sr*", GOTO="cdboot_start" GOTO="cdboot_end"
LABEL="cdboot_start" IMPORT{program}="/sbin/blkid -o udev -p $tempnode" ENV{ID_FS_USAGE}=="filesystem|other|crypto", ENV{ID_FS_UUID_ENC}=="?*", SYMLINK+="disk/by-uuid/$env{ID_FS_UUID_ENC}" ENV{ID_FS_USAGE}=="filesystem|other", ENV{ID_FS_LABEL_ENC}=="?*", SYMLINK+="disk/by-label/$env{ID_FS_LABEL_ENC}" LABEL="cdboot_end" Looks good for temporary solution, I sent an email to Kay (udev author) about problem with cdrom_id. So we need to wait for response.
About memdisk: It works fine on 64 Bit, but not on 32: http://bsfh.ath.cx/memdisk_32.png
Appending vmalloc=256M to the kernel command line fixes that, but it is apparently not without side effects - should be good enough for the ISO though. Wheel seems that depends on amount of RAM (in inverse, yeah), for example: Works OK under 32 bit 512MB RAM and no vmalloc. vmalloc= is needed if machine have >~ 800M of RAM. (I guess related to LOWMEM area defined at 32 bits).
-- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
On 06/26/2010 08:14 PM, Thomas Bächler wrote:
Am 27.06.2010 00:46, schrieb Gerardo Exequiel Pozzi:
I tested on real machine and booting from cdrom works OK. And also tested under Virtualbox and also works fine booting from cdrom (iso). So if you testwas on qemu, I guess that something is wrong with new cdrom_id (there are several changes) and qemu. Yes, I tested on qemu. A simple udev rule like this one in the initramfs will also do the trick: Ok seems to be a qemu-kvm issue.
I am doing some tests, and qemu cdrom seems to works as expected if more than one cd-rom drive is used but fails with only one cd-rom drive. To track details about this: http://www.spinics.net/lists/hotplug/msg03936.html https://bugzilla.redhat.com/show_bug.cgi?id=609049 -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
On 06/29/2010 08:53 PM, Gerardo Exequiel Pozzi wrote:
On 06/26/2010 08:14 PM, Thomas Bächler wrote:
Am 27.06.2010 00:46, schrieb Gerardo Exequiel Pozzi:
I tested on real machine and booting from cdrom works OK. And also tested under Virtualbox and also works fine booting from cdrom (iso). So if you testwas on qemu, I guess that something is wrong with new cdrom_id (there are several changes) and qemu. Yes, I tested on qemu. A simple udev rule like this one in the initramfs will also do the trick: Ok seems to be a qemu-kvm issue.
I am doing some tests, and qemu cdrom seems to works as expected if more than one cd-rom drive is used but fails with only one cd-rom drive. ignore me.
Fails on IDE cdrom (default used when -cdrom parameter is used), but works with SCSI cdrom (-drive file=archlinux-2010.06.29-netinstall-i686.iso,if=scsi,media=cdrom,boot)
To track details about this:
http://www.spinics.net/lists/hotplug/msg03936.html https://bugzilla.redhat.com/show_bug.cgi?id=609049
-- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
Am 30.06.2010 02:13, schrieb Gerardo Exequiel Pozzi:
I am doing some tests, and qemu cdrom seems to works as expected if more than one cd-rom drive is used but fails with only one cd-rom drive. ignore me.
Fails on IDE cdrom (default used when -cdrom parameter is used), but works with SCSI cdrom (-drive file=archlinux-2010.06.29-netinstall-i686.iso,if=scsi,media=cdrom,boot)
Does it work on virtio cdrom?
On 06/30/2010 05:45 AM, Thomas Bächler wrote:
Am 30.06.2010 02:13, schrieb Gerardo Exequiel Pozzi:
I am doing some tests, and qemu cdrom seems to works as expected if more than one cd-rom drive is used but fails with only one cd-rom drive. ignore me.
Fails on IDE cdrom (default used when -cdrom parameter is used), but works with SCSI cdrom (-drive file=archlinux-2010.06.29-netinstall-i686.iso,if=scsi,media=cdrom,boot) Does it work on virtio cdrom?
virtio cd-rom is not currently implemented in qemu. (I "discovered" while using new testiso script :P) Booting with if=virtio,media=cdrom is like doing with media=disk. -- Gerardo Exequiel Pozzi \cos^2\alpha + \sin^2\alpha = 1
participants (3)
-
Dieter Plaetinck
-
Gerardo Exequiel Pozzi
-
Thomas Bächler