On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org --- config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force case "$(uname -m)" in armv*) - mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire - rm /boot/initramfs-${KVER}-ipfire.img + mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire + rm /boot/initramfs-5.15.35-ipfire.img ;; aarch64) - mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire + mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;; esac
Peter,
How do I tell when this makes it into the CU 168 testing build? (The one that Pakfire will pickup & install)
I will can test this on the Raspberry Pi RPi4B and on the FriendlyArm R2S
Jon
On May 12, 2022, at 12:41 PM, Peter Müller peter.mueller@ipfire.org wrote:
On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org
config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force case "$(uname -m)" in armv*)
mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfirerm /boot/initramfs-${KVER}-ipfire.img
mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfirerm /boot/initramfs-5.15.35-ipfire.img ;; aarch64)
mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire
mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;;esac
2.35.3
Hello Jon,
thanks for your reply.
Usually, I do push fixes to Core Updates straight into "next". However, I do not have permission to push into "master" (Michael or Arne need to do this for me), and as soon as a patch lands there _and_ the nightly builds ran successfully, it is available to the "testing" channel of upcoming Core Updates.
However, I do not feel sufficiently confident with this patch, which is why I sent it to the mailing list, and wait for feedback on it before amending.
On a general notice: Any IPFire installation that ran Core Update 167 from the "stable" channel before (and has been rebooted at least once ever since) can, to my knowledge, update to Core Update 168 without all the hiccups Adolf and Rob observed. (Please do let me know if this is wrong.)
Therefore, you do not necessarily have to wait for this patch to land in "master", presumed your systems ran on Core Update 167 from the "stable" channel before.
Thanks, and best regards, Peter Müller
Peter,
How do I tell when this makes it into the CU 168 testing build? (The one that Pakfire will pickup & install)
I will can test this on the Raspberry Pi RPi4B and on the FriendlyArm R2S
Jon
On May 12, 2022, at 12:41 PM, Peter Müller peter.mueller@ipfire.org wrote:
On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org
config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force case "$(uname -m)" in armv*)
mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfirerm /boot/initramfs-${KVER}-ipfire.img
mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfirerm /boot/initramfs-5.15.35-ipfire.img ;; aarch64)
mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire
mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;;esac
2.35.3
Hi Peter,
On 12/05/2022 21:33, Peter Müller wrote:
Hello Jon,
thanks for your reply.
Usually, I do push fixes to Core Updates straight into "next". However, I do not have permission to push into "master" (Michael or Arne need to do this for me), and as soon as a patch lands there _and_ the nightly builds ran successfully, it is available to the "testing" channel of upcoming Core Updates.
However, I do not feel sufficiently confident with this patch, which is why I sent it to the mailing list, and wait for feedback on it before amending.
On a general notice: Any IPFire installation that ran Core Update 167 from the "stable" channel before (and has been rebooted at least once ever since) can, to my knowledge, update to Core Update 168 without all the hiccups Adolf and Rob observed. (Please do let me know if this is wrong.)
I am afraid that I am going to be disappointing you. I have a running vm with the current stable CU, currently CU167. It will only ever get updated with the next fully released CU.
When a Testing release is issued I create a clone of the running stable CU and then do the upgrade and evaluation on that vm. When the CU is released as a full stable one then I delete that Testing vm and update the existing stable vm.
That is how I have been doing it for a long time now.
Regards,
Adolf.
Therefore, you do not necessarily have to wait for this patch to land in "master", presumed your systems ran on Core Update 167 from the "stable" channel before.
Thanks, and best regards, Peter Müller
Peter,
How do I tell when this makes it into the CU 168 testing build? (The one that Pakfire will pickup & install)
I will can test this on the Raspberry Pi RPi4B and on the FriendlyArm R2S
Jon
On May 12, 2022, at 12:41 PM, Peter Müller peter.mueller@ipfire.org wrote:
On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org
config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force case "$(uname -m)" in armv*) - mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire - rm /boot/initramfs-${KVER}-ipfire.img + mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire + rm /boot/initramfs-5.15.35-ipfire.img ;; aarch64) - mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire + mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;; esac -- 2.35.3
Hello Adolf,
thanks for your reply.
This changes things then, and this patch can be dropped.
As for your issue, the only thing I can think of so far is the linux-firmware update. I will go through the delta and look for anything suspicious, since I am not going to be able to do some VM-based testing myself before next week.
Which VM setup are you running on? VirtualBox?
Thanks, and best regards, Peter Müller
Hi Peter,
On 12/05/2022 21:33, Peter Müller wrote:
Hello Jon,
thanks for your reply.
Usually, I do push fixes to Core Updates straight into "next". However, I do not have permission to push into "master" (Michael or Arne need to do this for me), and as soon as a patch lands there _and_ the nightly builds ran successfully, it is available to the "testing" channel of upcoming Core Updates.
However, I do not feel sufficiently confident with this patch, which is why I sent it to the mailing list, and wait for feedback on it before amending.
On a general notice: Any IPFire installation that ran Core Update 167 from the "stable" channel before (and has been rebooted at least once ever since) can, to my knowledge, update to Core Update 168 without all the hiccups Adolf and Rob observed. (Please do let me know if this is wrong.)
I am afraid that I am going to be disappointing you. I have a running vm with the current stable CU, currently CU167. It will only ever get updated with the next fully released CU.
When a Testing release is issued I create a clone of the running stable CU and then do the upgrade and evaluation on that vm. When the CU is released as a full stable one then I delete that Testing vm and update the existing stable vm.
That is how I have been doing it for a long time now.
Regards,
Adolf.
Therefore, you do not necessarily have to wait for this patch to land in "master", presumed your systems ran on Core Update 167 from the "stable" channel before.
Thanks, and best regards, Peter Müller
Peter,
How do I tell when this makes it into the CU 168 testing build? (The one that Pakfire will pickup & install)
I will can test this on the Raspberry Pi RPi4B and on the FriendlyArm R2S
Jon
On May 12, 2022, at 12:41 PM, Peter Müller peter.mueller@ipfire.org wrote:
On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org
config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force case "$(uname -m)" in armv*) - mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire - rm /boot/initramfs-${KVER}-ipfire.img + mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire + rm /boot/initramfs-5.15.35-ipfire.img ;; aarch64) - mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire + mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;; esac -- 2.35.3
Hi Peter,
On 12/05/2022 22:00, Peter Müller wrote:
Hello Adolf,
thanks for your reply.
This changes things then, and this patch can be dropped.
As for your issue, the only thing I can think of so far is the linux-firmware update. I will go through the delta and look for anything suspicious, since I am not going to be able to do some VM-based testing myself before next week.
Which VM setup are you running on? VirtualBox?
Yes it is VirtualBox.
Regards, Adolf.
Thanks, and best regards, Peter Müller
Hi Peter,
On 12/05/2022 21:33, Peter Müller wrote:
Hello Jon,
thanks for your reply.
Usually, I do push fixes to Core Updates straight into "next". However, I do not have permission to push into "master" (Michael or Arne need to do this for me), and as soon as a patch lands there _and_ the nightly builds ran successfully, it is available to the "testing" channel of upcoming Core Updates.
However, I do not feel sufficiently confident with this patch, which is why I sent it to the mailing list, and wait for feedback on it before amending.
On a general notice: Any IPFire installation that ran Core Update 167 from the "stable" channel before (and has been rebooted at least once ever since) can, to my knowledge, update to Core Update 168 without all the hiccups Adolf and Rob observed. (Please do let me know if this is wrong.)
I am afraid that I am going to be disappointing you. I have a running vm with the current stable CU, currently CU167. It will only ever get updated with the next fully released CU.
When a Testing release is issued I create a clone of the running stable CU and then do the upgrade and evaluation on that vm. When the CU is released as a full stable one then I delete that Testing vm and update the existing stable vm.
That is how I have been doing it for a long time now.
Regards,
Adolf.
Therefore, you do not necessarily have to wait for this patch to land in "master", presumed your systems ran on Core Update 167 from the "stable" channel before.
Thanks, and best regards, Peter Müller
Peter,
How do I tell when this makes it into the CU 168 testing build? (The one that Pakfire will pickup & install)
I will can test this on the Raspberry Pi RPi4B and on the FriendlyArm R2S
Jon
On May 12, 2022, at 12:41 PM, Peter Müller peter.mueller@ipfire.org wrote:
On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org
config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force case "$(uname -m)" in armv*) - mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire - rm /boot/initramfs-${KVER}-ipfire.img + mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire + rm /boot/initramfs-5.15.35-ipfire.img ;; aarch64) - mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire + mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;; esac -- 2.35.3
Can you elaborate more on what this patch is supposed to fix?
On 12 May 2022, at 18:41, Peter Müller peter.mueller@ipfire.org wrote:
On systems that have previously running on testing, kernel 5.15.32 might still be installed. dracut being called with ${KVER} will then build an inital ramdisk for the wrong kernel, as 5.15.32 might still be running, albeit 5.15.35 has been installed due to the Pakfire procedure when upgrading on testing.
Due to lack of hardware, this patch is untested on ARM.
https://lists.ipfire.org/pipermail/development/2022-May/013433.html
Reported-by: Stefan Schantl stefan.schantl@ipfire.org Signed-off-by: Peter Müller peter.mueller@ipfire.org
config/rootfiles/core/168/update.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/config/rootfiles/core/168/update.sh b/config/rootfiles/core/168/update.sh index d21f648dd..7cc8800b2 100644 --- a/config/rootfiles/core/168/update.sh +++ b/config/rootfiles/core/168/update.sh @@ -107,14 +107,14 @@ chmod -v 750 /etc/sudoers.d chmod -v 640 /etc/sudoers.d/*
# Rebuild initial ramdisk to apply microcode updates -dracut --regenerate-all --force +dracut --kver="5.15.35-ipfire" --regenerate-all --force
Isn’t specifying the release and “—-regenerate-all” a conflict?
-Michael
case "$(uname -m)" in armv*)
mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfirerm /boot/initramfs-${KVER}-ipfire.img
mkimage -A arm -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfirerm /boot/initramfs-5.15.35-ipfire.img ;; aarch64)
mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-${KVER}-ipfire.img /boot/uInit-${KVER}-ipfire
mkimage -A arm64 -T ramdisk -C lzma -d /boot/initramfs-5.15.35-ipfire.img /boot/uInit-5.15.35-ipfire # dont remove initramfs because grub need this to boot. ;;esac
2.35.3
-
Adolf Belka -
Jon Murphy -
Michael Tremer -
Peter Müller