-
Notifications
You must be signed in to change notification settings - Fork 1.8k
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
SPL linux 4.0.6 and 4.0.1 compatibility #3523
Comments
I have no problems building version 0.6.4.1 under kernel 4.0.6 . Using Arch Linux, gcc 5.1.0. Which version of ZFS on Linux do you use? |
@djdunn this shouldn't be fatal. If it is please include the |
Im getting it from here is the build log checking whether old 3-argument shrinker exists... no !!! Please attach the following file when seeking support:
|
config.log
|
some of my buildinfo Portage 2.2.20 (python 3.4.3-final-0, default/linux/amd64/13.0/desktop/kde, gcc-4.8.5, glibc-2.20-r2, 4.0.5 x86_64) System uname: Linux-4.0.5-x86_64-AMD_FX-tm-8350_Eight-Core_Processor-with-gentoo-2.2 |
I have not had this problem on gentoo. I have a few systems using the 9999 builds and gentoo-sources-4.0.6 or gentoo-sources-4.1.0. Here is one: vm_gentoo ~ # uname -a
|
@djdunn it appears your hitting a surprising warning in the configure check which is causing the failure. Is there anything particularly special / exotic about your hardware which might have triggered this?
In the meanwhile I suspect you'll be able to build with the following small patch to the SPL. You'll need to be sure to run diff --git a/config/spl-build.m4 b/config/spl-build.m4
index 9999d48..148b60d 100644
--- a/config/spl-build.m4
+++ b/config/spl-build.m4
@@ -740,8 +740,6 @@ AC_DEFUN([SPL_AC_ATOMIC_SPINLOCK], [
])
AC_DEFUN([SPL_AC_SHRINKER_CALLBACK],[
- tmp_flags="$EXTRA_KCFLAGS"
- EXTRA_KCFLAGS="-Werror"
dnl #
dnl # 2.6.23 to 2.6.34 API change
dnl # ->shrink(int nr_to_scan, gfp_t gfp_mask)
@@ -840,7 +838,6 @@ AC_DEFUN([SPL_AC_SHRINKER_CALLBACK],[
])
])
])
- EXTRA_KCFLAGS="$tmp_flags"
])
dnl # |
Perhaps -mfpmath=sse flag? I compile my kernel with -march=native maybe? |
Most definitly having the -mfpmath=sse flag in the kernel makefile did it |
@djdunn was that something you added to the kernel makefile? |
@behlendorf Yes it was, most of the kernel source disables sse math even if you set it explicitly, and spits out that sse disabled error, using 387 arithmetic error, I really wouldn't consider it an error tho more of a warning |
@djdunn ahh I see. Normally it is a warning, but the configure checks promote all warnings to errors since we want to be strict about the interface we're using. If we could suppress the warning with gcc that would resolve things. |
checking whether ->count_objects callback exists... configure: error: error
starts happening after 4.0.5, as 4.0.5 is last that works using the .config from 4.0.5 with 4.0.6
gentoo system using spl-9999
The text was updated successfully, but these errors were encountered: