Repository navigation
fedora & rhel-8,9,10: drop unused no_timer_check kernel parameter - #1856
Conversation
achilleas-k
left a comment
achilleas-k
left a comment
There was a problem hiding this comment.
Splendid. Thank you!
LGTM
|
Note: Failing integration tests are not required and fixed in |
mvo5
left a comment
mvo5
left a comment
There was a problem hiding this comment.
The change looks fine. I guess it is okay that images for e.g. rhel-8.10 build once this landed are "different" from the previous images, i.e. its not a stability we guarantee or need to document somewhere?
Appreciated the head-up, thanks. So no rebasing needed, we can put it to the queue.
If this was for me, I do not understand at all as I am not familiar with |
|
[..]
More a question for the super experienced people, tbh I still often am not sure about policy details like this myself so I just want to make it explicit and hope that someone like @achilleas-k knows the answer :) |
Taking the bug report at face value, this should not have any functional effect. If that's the case, then I think it should be fine to drop this everywhere. Since it came up though, I did try to find out more about this. I also found suggestions that this is enabled by default when running on KVM, which I think is what the current bug/request is suggesting with the pasted function docstring. It seems to me it's safe to drop the explicit option in images as far back as RHEL 8. |
e23aade to
3daa820
Compare
|
Ok, rebased to resolve conflicts in manifests. Diff: https://gist.github.com/lzap/68130d16741e00895a33a97eab9c3d87 |
|
@lzap this needs a rebase |
3daa820 to
6d9eb7f
Compare
|
Rebased, this is getting ugly so I am proposing a radical change: #1878 |
This change removes the legacy no_timer_check kernel parameter from the README to clean it up. This parameter was needed in the past (circa 2014-2016) to work around issues on older hardware, but it has not been necessary since the RHEL 7 era. As explained in arch/x86/kernel/apic/io_apic.c, its original purpose was to handle buggy SMP boards with incorrect mptable information about the timer IRQ. On modern hypervisors like KVM, VMware, and Hyper-V, the kernel's timer handling is robust, and the behavior already mimics the effect of this parameter, making it redundant. There is no reason to add this legacy parameter on RHEL 8 or later, and Fedora 19 and later. Removing it aligns with best practices by eliminating outdated configuration, which helps prevent future confusion.
This change removes the legacy no_timer_check kernel parameter for Fedora. This parameter was needed in the past (circa 2014-2016) to work around issues on older hardware, but it has not been necessary since the RHEL 7 era. As explained in arch/x86/kernel/apic/io_apic.c, its original purpose was to handle buggy SMP boards with incorrect mptable information about the timer IRQ. On modern hypervisors like KVM, VMware, and Hyper-V, the kernel's timer handling is robust, and the behavior already mimics the effect of this parameter, making it redundant. There is no reason to add this legacy parameter on RHEL 8 or later, and Fedora 19 and later. Removing it aligns with best practices by eliminating outdated configuration, which helps prevent future confusion.
This change removes the legacy no_timer_check kernel parameter for RHEL8. This parameter was needed in the past (circa 2014-2016) to work around issues on older hardware, but it has not been necessary since the RHEL 7 era. As explained in arch/x86/kernel/apic/io_apic.c, its original purpose was to handle buggy SMP boards with incorrect mptable information about the timer IRQ. On modern hypervisors like KVM, VMware, and Hyper-V, the kernel's timer handling is robust, and the behavior already mimics the effect of this parameter, making it redundant. There is no reason to add this legacy parameter on RHEL 8 or later, and Fedora 19 and later. Removing it aligns with best practices by eliminating outdated configuration, which helps prevent future confusion.
This change removes the legacy no_timer_check kernel parameter for EL9. This parameter was needed in the past (circa 2014-2016) to work around issues on older hardware, but it has not been necessary since the RHEL 7 era. As explained in arch/x86/kernel/apic/io_apic.c, its original purpose was to handle buggy SMP boards with incorrect mptable information about the timer IRQ. On modern hypervisors like KVM, VMware, and Hyper-V, the kernel's timer handling is robust, and the behavior already mimics the effect of this parameter, making it redundant. There is no reason to add this legacy parameter on RHEL 8 or later, and Fedora 19 and later. Removing it aligns with best practices by eliminating outdated configuration, which helps prevent future confusion.
This change removes the legacy no_timer_check kernel parameter for EL10. This parameter was needed in the past (circa 2014-2016) to work around issues on older hardware, but it has not been necessary since the RHEL 7 era. As explained in arch/x86/kernel/apic/io_apic.c, its original purpose was to handle buggy SMP boards with incorrect mptable information about the timer IRQ. On modern hypervisors like KVM, VMware, and Hyper-V, the kernel's timer handling is robust, and the behavior already mimics the effect of this parameter, making it redundant. There is no reason to add this legacy parameter on RHEL 8 or later, and Fedora 19 and later. Removing it aligns with best practices by eliminating outdated configuration, which helps prevent future confusion.
6d9eb7f to
c852cec
Compare
|
Rebased again - conflicts. |
6 participants
This parameter was needed in the past (circa 2014-2016) to work around issues
on older hardware, but it has not been necessary since the RHEL 7 era.
As explained in arch/x86/kernel/apic/io_apic.c, its original purpose was to
handle buggy SMP boards with incorrect mptable information about the timer IRQ.
On modern hypervisors like KVM, VMware, and Hyper-V, the kernel's timer
handling is robust, and the behavior already mimics the effect of this
parameter, making it redundant.
There is no reason to add this legacy parameter on RHEL 8 or later, and Fedora
19 and later. Removing it aligns with best practices by eliminating outdated
configuration, which helps prevent future confusion.
https://issues.redhat.com/browse/RHEL-11974