Re: Time for Jessie Beta 2?
On 06:02, Cyril Brulebois wrote: I've discovered #763580 though, so I might revert this change for now since I don't want to reintroduce this bug right now, and generate more work for translators because of an errata we could and should have avoided, since that problem was warned against already. We're possibly looking at an upload this wednesday (European) afternoon. Bringing back the bug from d-i Beta 1 (#712907 auto preseed installs get stuck) seems like the lesser evil than (re)introducing #763580. Obviously we can't find a better fix in this short timeframe, but I hope we can try again for Beta 3. Regards, -- Steven Chamberlain ste...@pyro.eu.org -- To UNSUBSCRIBE, email to debian-cd-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20141001133804.gb5...@squeeze.pyro.eu.org
Re: Bug#763127: UEFI corner case - installer booted in UEFI mode, existing system in BIOS mode - PLEASE REVIEW
Control: tag -1 +patch On Wed, Oct 01, 2014 at 02:43:53AM +0100, Steve McIntyre wrote: On Mon, Sep 29, 2014 at 03:34:17PM +0100, Steve McIntyre wrote: On Sun, Sep 28, 2014 at 08:46:41PM +0100, Steve McIntyre wrote: On Sun, Sep 28, 2014 at 07:50:38PM +0100, Colin Watson wrote: On Sun, Sep 28, 2014 at 06:14:30PM +0100, Steve McIntyre wrote: That sounds better to me too, assuming we can sensibly do a question at that point. Is that allowed? I honestly don't know... :-/ While isinstallable scripts can talk to debconf, they should not ask questions, as main-menu will run the isinstallable script for a given package potentially many times. Also, partman-efi isn't a main-menu item so any isinstallable script added to it will never be run. Right. Thanks for warning before we spend too much time on that option! :-/ So, back to other options I guess. I'm hacking on init.d/efi, and it looks like it might do what we need. We can remember a decision taken (choose not to do EFI installation), but we'll need to update either all the places that currently check if we're an EFI platform or the core libdebian-installer code to look for a flag file or similar... I *think* I have a working solution here, testing now. \o/ Right, here's the changes I have. It works OK for me, tested on a few different systems. The core of the logic is in partman-efi/init.d/efi: * Check for EFI System Partitions on the target disks - if we have none but have other filesystems then that suggests there is a BIOS-mode OS already installed. If so, ask the user what they want to do. * Add a flag file /var/lib/partman/ignore_uefi that other bits of d-i can look for to know that although the installer is running in UEFI mode we should stick to installing for BIOS-mode boot instead. Elsewhere in d-i, anywhere we try to work out what system we're on and care about UEFI, I've added checks for the flag file /var/lib/partman/ignore_uefi - if that exists, don't do UEFI installation but instead fall through to BIOS-mode installation. I've also added a similar check in archdetect in libdebian-installer - that and the other changes are overkill, more like belt-and-braces. Which way people would prefer to go, I don't know. The message in partman-efi/init.d/efi is probably horrible - please suggest better! It would also need translating, of course. So, please look at the patches and tell me what you think folks. I'd like to get this fix in for beta 3 if possible... -- Steve McIntyre, Cambridge, UK.st...@einval.com This dress doesn't reverse. -- Alden Spiess From 046dcbe4f1c7880205c81eabed540407b515d68c Mon Sep 17 00:00:00 2001 From: Steve McIntyre st...@einval.com Date: Thu, 2 Oct 2014 00:08:04 +0100 Subject: [PATCH] Better handle mixed UEFI and non-UEFI booting Warn the user if we've booted in UEFI mode but we seem to have only non-UEFI existing OS installations - give them the option to switch the installer to non-UEFI mode from this point forwards so they don't break potential dual-boot setup. Closes: #763127 Check for EFI System Partitions on the target disks - if we have none but have other filesystems then that suggests there is a BIOS-mode OS already installed. If so, ask the user what they want to do. Add a flag file /var/lib/partman/ignore_uefi that other bits of d-i can look for to know that although the installer is running in UEFI mode we should stick to installing for BIOS-mode boot instead. --- check.d/efi |4 choose_method/efi/choices|4 debian/changelog | 10 ++ debian/partman-efi.templates | 16 debian/po/templates.pot | 15 +++ init.d/efi | 25 + update.d/efi_visuals |4 7 files changed, 78 insertions(+) diff --git a/check.d/efi b/check.d/efi index 395688d..9d74bd3 100755 --- a/check.d/efi +++ b/check.d/efi @@ -4,6 +4,10 @@ if [ ! -d /proc/efi ] [ ! -d /sys/firmware/efi ]; then exit 0 fi +if [ -f /var/lib/partman/ignore_uefi ]; then + exit 0 +fi + . /lib/partman/lib/base.sh have_efi=no diff --git a/choose_method/efi/choices b/choose_method/efi/choices index 160a565..b4af018 100755 --- a/choose_method/efi/choices +++ b/choose_method/efi/choices @@ -9,6 +9,10 @@ if [ ! -d /proc/efi ] [ ! -d /sys/firmware/efi ]; then exit 0 fi +if [ -f /var/lib/partman/ignore_uefi ]; then + exit 0 +fi + db_metaget partman-efi/text/efi description printf efi\t${RET}\n diff --git a/debian/changelog b/debian/changelog index 383b5ae..2a96121 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,3 +1,13 @@ +partman-efi (49) unstable; urgency=medium + + [ Steve McIntyre ] + * Warn the user if we've booted in UEFI mode but we seem to have only +non-UEFI existing OS installations - give them the option to switch +the installer to non-UEFI mode from this point forwards so they don't +break potential
Re: Bug#763127: UEFI corner case - installer booted in UEFI mode, existing system in BIOS mode - PLEASE REVIEW
Hang fire - one test just failed. :-( On Thu, Oct 02, 2014 at 12:29:40AM +0100, Steve McIntyre wrote: Control: tag -1 +patch On Wed, Oct 01, 2014 at 02:43:53AM +0100, Steve McIntyre wrote: On Mon, Sep 29, 2014 at 03:34:17PM +0100, Steve McIntyre wrote: On Sun, Sep 28, 2014 at 08:46:41PM +0100, Steve McIntyre wrote: On Sun, Sep 28, 2014 at 07:50:38PM +0100, Colin Watson wrote: On Sun, Sep 28, 2014 at 06:14:30PM +0100, Steve McIntyre wrote: That sounds better to me too, assuming we can sensibly do a question at that point. Is that allowed? I honestly don't know... :-/ While isinstallable scripts can talk to debconf, they should not ask questions, as main-menu will run the isinstallable script for a given package potentially many times. Also, partman-efi isn't a main-menu item so any isinstallable script added to it will never be run. Right. Thanks for warning before we spend too much time on that option! :-/ So, back to other options I guess. I'm hacking on init.d/efi, and it looks like it might do what we need. We can remember a decision taken (choose not to do EFI installation), but we'll need to update either all the places that currently check if we're an EFI platform or the core libdebian-installer code to look for a flag file or similar... I *think* I have a working solution here, testing now. \o/ Right, here's the changes I have. It works OK for me, tested on a few different systems. The core of the logic is in partman-efi/init.d/efi: * Check for EFI System Partitions on the target disks - if we have none but have other filesystems then that suggests there is a BIOS-mode OS already installed. If so, ask the user what they want to do. * Add a flag file /var/lib/partman/ignore_uefi that other bits of d-i can look for to know that although the installer is running in UEFI mode we should stick to installing for BIOS-mode boot instead. Elsewhere in d-i, anywhere we try to work out what system we're on and care about UEFI, I've added checks for the flag file /var/lib/partman/ignore_uefi - if that exists, don't do UEFI installation but instead fall through to BIOS-mode installation. I've also added a similar check in archdetect in libdebian-installer - that and the other changes are overkill, more like belt-and-braces. Which way people would prefer to go, I don't know. The message in partman-efi/init.d/efi is probably horrible - please suggest better! It would also need translating, of course. So, please look at the patches and tell me what you think folks. I'd like to get this fix in for beta 3 if possible... -- Steve McIntyre, Cambridge, UK.st...@einval.com This dress doesn't reverse. -- Alden Spiess From 046dcbe4f1c7880205c81eabed540407b515d68c Mon Sep 17 00:00:00 2001 From: Steve McIntyre st...@einval.com Date: Thu, 2 Oct 2014 00:08:04 +0100 Subject: [PATCH] Better handle mixed UEFI and non-UEFI booting Warn the user if we've booted in UEFI mode but we seem to have only non-UEFI existing OS installations - give them the option to switch the installer to non-UEFI mode from this point forwards so they don't break potential dual-boot setup. Closes: #763127 Check for EFI System Partitions on the target disks - if we have none but have other filesystems then that suggests there is a BIOS-mode OS already installed. If so, ask the user what they want to do. Add a flag file /var/lib/partman/ignore_uefi that other bits of d-i can look for to know that although the installer is running in UEFI mode we should stick to installing for BIOS-mode boot instead. --- check.d/efi |4 choose_method/efi/choices|4 debian/changelog | 10 ++ debian/partman-efi.templates | 16 debian/po/templates.pot | 15 +++ init.d/efi | 25 + update.d/efi_visuals |4 7 files changed, 78 insertions(+) diff --git a/check.d/efi b/check.d/efi index 395688d..9d74bd3 100755 --- a/check.d/efi +++ b/check.d/efi @@ -4,6 +4,10 @@ if [ ! -d /proc/efi ] [ ! -d /sys/firmware/efi ]; then exit 0 fi +if [ -f /var/lib/partman/ignore_uefi ]; then + exit 0 +fi + . /lib/partman/lib/base.sh have_efi=no diff --git a/choose_method/efi/choices b/choose_method/efi/choices index 160a565..b4af018 100755 --- a/choose_method/efi/choices +++ b/choose_method/efi/choices @@ -9,6 +9,10 @@ if [ ! -d /proc/efi ] [ ! -d /sys/firmware/efi ]; then exit 0 fi +if [ -f /var/lib/partman/ignore_uefi ]; then + exit 0 +fi + db_metaget partman-efi/text/efi description printf efi\t${RET}\n diff --git a/debian/changelog b/debian/changelog index 383b5ae..2a96121 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,3 +1,13 @@ +partman-efi (49) unstable; urgency=medium + + [ Steve McIntyre ] + * Warn the user if we've booted in UEFI mode but we seem to have only +non-UEFI existing OS installations - give them
Re: Bug#763127: UEFI corner case - installer booted in UEFI mode, existing system in BIOS mode - PLEASE REVIEW
On Thu, 2014-10-02 at 00:29 +0100, Steve McIntyre wrote: From: Steve McIntyre st...@einval.com Date: Thu, 2 Oct 2014 00:01:28 +0100 Subject: [PATCH] Recognise the new ignore_uefi flag from partman-efi [...] --- a/src/system/efi.c +++ b/src/system/efi.c @@ -28,7 +28,14 @@ int di_system_is_efi(void) { int ret = access(/sys/firmware/efi, R_OK); if (ret == 0) - return 1; + { + /* Have we been told to ignore EFI in partman-efi? */ + ret = access(/var/lib/partman/ignore_uefi, R_OK); + if (ret == 0) + return 1; Should be return 0? + else + return 1; [...] Ben. -- Ben Hutchings Everything should be made as simple as possible, but not simpler. - Albert Einstein signature.asc Description: This is a digitally signed message part