linux: 5.2-rc7 -> 5.3-rc8 #67111
linux: 5.2-rc7 -> 5.3-rc8 #67111
Conversation
NeQuissimus
force-pushed
the
kernel_5_3
branch
from
40b0f01
to
863b902
Compare
|
|
||
| -modules := $(sort $(shell cat $(if $(KBUILD_EXTMOD),$(KBUILD_EXTMOD)/)modules.order)) | ||
| +modules := $(sort $(foreach f,$(wildcard $(MODVERDIR)/*.mod),$(shell \ | ||
| + grep -h '\.ko$$' '$f'))) |
There was a problem hiding this comment.
Hm, did you check whether this patch is still needed? Like eg. build the kernel by adding a bunch of environment variables and see whether modules are properly installed.
There was a problem hiding this comment.
Hm, did you check whether this patch is still needed? Like eg. build the kernel by adding a bunch of environment variables and see whether modules are properly installed.
As it is now, AFAICT no modules actually get installed :(. Found out when went to reboot and expected modules weren't found... (somewhat surprised initrd generator didn't die during build or something,but shrug).
Removing this patch (64829de) builds and installs the modules. I'm not sure what construction was previously running into argument length limits, but via notes here: 98842f8 (which links to kernel commit changing Makefile.modinst, hence the conflict) it might be possible it's no longer a problem. Or maybe only shows up on Darwin or when cross-compiling or something O:).
There was a problem hiding this comment.
@dtzWill: Okay, can you please verify if it really works by adding some more environment variables to the build?
NeQuissimus
force-pushed
the
kernel_5_3
branch
from
863b902
to
0be8207
Compare
|
@GrahamcOfBorg test kernel-testing |
|
Sorry guys, I was running the test overnight, only now saw that it wasn't quite right. Figured last night I'd get early feedback :) |
|
The test passes, I can't boot into it because my root drive is ZFS and ZOL does not yet support 5.3, so the module won't build. |
|
I think |
|
I assume you have checked that the compile test for succeeds? |
|
I have not attempted to make ZOL work at all |
NeQuissimus
force-pushed
the
kernel_5_3
branch
from
4fe29ca
to
0f286be
Compare
|
I got ZFS to work by updating to the proposed 0.8.2 patchset (openzfs/zfs#9161). I had to revert one path-related commit to get it working. my tree is here: https://github.com/jb55/zfs/tree/zfs-0.8.2 I'm currently running 5.3-rc7 nixos and nothing has exploded yet |
NeQuissimus
force-pushed
the
kernel_5_3
branch
from
0f286be
to
6255a4b
Compare
|
@GrahamcOfBorg test kernel-testing |
|
I am going to merge this once the tests pass |
|
FWIW 5.3-rc8 was released yesterday: https://lwn.net/Articles/798743/
…On Mon, 09 Sep 2019 06:24:37 -0700, Tim Steinbach ***@***.***> wrote:
I am going to merge this once the tests pass
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#67111 (comment) part: text/html
|
|
@dtzWill Already included in this PR :) |
|
Will this hit 19.09 also? |
Motivation for this change
Get ready for next mainstream kernel
Things done
sandboxinnix.confon non-NixOS)nix-shell -p nix-review --run "nix-review wip"./result/bin/)nix path-info -Sbefore and after)I had to remove the
--exclude='/Documentation/'clause because I could not figure out how to make the build happy.It will look for
Documentation/KConfigand try to build that.Somehow, I could not determine where it is actually told to look for that...
The fpu functions are now exported (under different names) and openzfs/zfs@becdcec picks them up from there