-
Notifications
You must be signed in to change notification settings - Fork 65
Drop unused support for block aperture, fixing build with kernel 5.18 #194
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Quoting Linux commit f8669f1d6a86a6b17104ceca9340ded280307ac1: > Block Aperture Window support was an attempt to layer an error model > over PMEM for platforms that did not support machine-check-recovery. > However, it was abandoned before it ever shipped, and only ever existed > in the ACPI specification. Meanwhile Linux has carried a large pile of > dead code for non-shipping infrastructure. [...] We fetched block sizes but never used that info for anything. Now that the required defines have been dropped in kernel 5.18, let's purge that code. Signed-off-by: Adam Borowski <adam.borowski@intel.com>
Thanks, we will take a look. |
Here's the original report, BTW. |
The name ND_DEVICE_NAMESPACE_BLK is defined in ndctl.h as is ND_DEVICE_NAMESPACE_IO. Looking at the report linked to it appears like the file should be coming from one of these packages
which I believe is looking in http://deb.debian.org/debian/pool/main/n/ndctl/ndctl_73.orig.tar.gz it appears like ND_DEVICE_NAMESPACE_BLK is defined So I think the code as exists should compile. I am missing something? |
The define comes from But even if |
You are correct, we don't expose the values to any external API. Making your change as well as some other sections. It'll take a while, since we need to start the change internally. Thanks! |
I'll use my patch as-is for Debian packaging then (as maintainers get shouted a lot for "does not build" bugs), you can take your time. |
OE-core has updated kernel headers to 5.19 and that is resulting in failures: | /build/cje/workspace/poky/build/tmp/work/corei7-64-poky-linux/ipmctl/03.00.00.0439-r0/git/src/os/linux/lnx_system.c:336:52: error: 'ND_DEVICE_NAMESPACE_BLK' undeclared (first use in this function); did you mean 'ND_DEVICE_NAMESPACE_IO'? | 336 | (nstype == ND_DEVICE_NAMESPACE_BLK)) | | ^~~~~~~~~~~~~~~~~~~~~~~ | | ND_DEVICE_NAMESPACE_IO | compilation terminated due to -Wfatal-errors. For more details: intel/ipmctl#194 Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Hello, is this issue going to be fixed any time soon ?? |
@vynu , we are not planning to spin new version of ipmctl at this time. |
Every downstream project that builds ipmctl from source needs to apply the patch. For some projects, applying patches is a routine matter, for others it requires tedious manual hacks. But in every case, it's unnecessary work that's multiplied for every downstream. |
Done, thanks @kilobyte! Let me know if there is another step I should do. |
OE-core has updated kernel headers to 5.19 and that is resulting in failures: | /build/cje/workspace/poky/build/tmp/work/corei7-64-poky-linux/ipmctl/03.00.00.0439-r0/git/src/os/linux/lnx_system.c:336:52: error: 'ND_DEVICE_NAMESPACE_BLK' undeclared (first use in this function); did you mean 'ND_DEVICE_NAMESPACE_IO'? | 336 | (nstype == ND_DEVICE_NAMESPACE_BLK)) | | ^~~~~~~~~~~~~~~~~~~~~~~ | | ND_DEVICE_NAMESPACE_IO | compilation terminated due to -Wfatal-errors. For more details: intel/ipmctl#194 Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
OE-core has updated kernel headers to 5.19 and that is resulting in failures: | /build/cje/workspace/poky/build/tmp/work/corei7-64-poky-linux/ipmctl/03.00.00.0439-r0/git/src/os/linux/lnx_system.c:336:52: error: 'ND_DEVICE_NAMESPACE_BLK' undeclared (first use in this function); did you mean 'ND_DEVICE_NAMESPACE_IO'? | 336 | (nstype == ND_DEVICE_NAMESPACE_BLK)) | | ^~~~~~~~~~~~~~~~~~~~~~~ | | ND_DEVICE_NAMESPACE_IO | compilation terminated due to -Wfatal-errors. For more details: intel/ipmctl#194 Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
With current userspace kernel headers, the build fails:
As block sizes we fetch have never been used for anything, let's just drop the query.