Compilation error: kmod/core/core.c:992:27: error: passing argument 2 of ‘set_bit’ from incompatible pointer type [-Werror=incompatible-pointer-types]

[marcin-github]

Hi!
I'm getting compilation error:

CC [M] /var/tmp/kpatch/kmod/core/core.o
/var/tmp/kpatch/kmod/core/core.c: In function ‘kpatch_register’:
/var/tmp/kpatch/kmod/core/core.c:992:27: error: passing argument 2 of ‘set_bit’ from incompatible pointer type [-Werror=incompatible-pointer-types]
set_bit(TAINT_LIVEPATCH, &kpmod->mod->taints);
^
In file included from ./include/linux/bitops.h:36:0,
from ./include/linux/kernel.h:10,
from ./include/linux/list.h:8,
from ./include/linux/module.h:9,
from /var/tmp/kpatch/kmod/core/core.c:37:
./arch/x86/include/asm/bitops.h:72:1: note: expected ‘volatile long unsigned int *’ but argument is of type ‘unsigned int *’
set_bit(long nr, volatile unsigned long *addr)
^
cc1: some warnings being treated as errors

# gcc -v
Using built-in specs.
COLLECT_GCC=/usr/x86_64-pc-linux-gnu/gcc-bin/5.4.0/gcc
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-pc-linux-gnu/5.4.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: /var/tmp/portage/sys-devel/gcc-5.4.0-r2/work/gcc-5.4.0/configure --host=x86_64-pc-linux-gnu --build=x86_64-pc-linux-gnu --prefix=/usr --bindir=/usr/x86_64-pc-linux-gnu/gcc-bin/5.4.0 --includedir=/usr/lib/gcc/x86_64-pc-linux-gnu/5.4.0/include --datadir=/usr/share/gcc-data/x86_64-pc-linux-gnu/5.4.0 --mandir=/usr/share/gcc-data/x86_64-pc-linux-gnu/5.4.0/man --infodir=/usr/share/gcc-data/x86_64-pc-linux-gnu/5.4.0/info --with-gxx-include-dir=/usr/lib/gcc/x86_64-pc-linux-gnu/5.4.0/include/g++-v5 --with-python-dir=/share/gcc-data/x86_64-pc-linux-gnu/5.4.0/python --enable-languages=c,c++,objc,fortran --enable-obsolete --enable-secureplt --disable-werror --with-system-zlib --enable-nls --without-included-gettext --enable-checking=release --with-bugurl=https://bugs.gentoo.org/ --with-pkgversion='Gentoo 5.4.0-r2 p1.2, pie-0.6.5' --enable-libstdcxx-time --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-multilib --with-multilib-list=m32,m64 --disable-altivec --disable-fixed-point --enable-targets=all --disable-libgcj --enable-libgomp --disable-libmudflap --disable-libssp --disable-libcilkrts --disable-libmpx --enable-vtable-verify --enable-libvtv --enable-lto --with-isl --disable-isl-version-check --disable-libsanitizer
Thread model: posix
gcc version 5.4.0 (Gentoo 5.4.0-r2 p1.2, pie-0.6.5)

[jpoimboe]

@marcin-github Thanks for reporting it. I'm guessing you have Linux 4.9?

Linux 4.9 has the module-specific livepatch taint (2992ef29ae01), but the conversion of mod->taints to long (7fd8329ba) isn't until 4.10. So we shouldn't be doing the set_bit() with 4.9.

Maybe we can just skip setting the module-specific taint on 4.9, and change the setting of HAS_MODULE_TAINT to >= 4.10. @joe-lawrence, thoughts?

[marcin-github]

You are right, I'm using 4.9.

[joe-lawrence]

@jpoimboe -- yup, double checking now and the commit/releases timeline is correct.

So really there are three! possible taints - kernel, module (int), and module (long). Would it make sense to convert HAS_RHEL7_MODULE_TAINT into HAS_INT_MODULE_TAINT ?

[jpoimboe]

@joe-lawrence Yeah, that sounds like a good idea.

[marcin-github]

With this patch kpatch build without problem.
I'm reading readme and trying to do my first live kernel patch. kpatch-build throws:
ERROR: can't find special struct size.

It looks that gawk can't find needed information from output of readelf -wi. Should I continue looking for help of the problem here, in new issue or at https://www.redhat.com/mailman/listinfo/kpatch ?

[joe-lawrence]

@marcin-github - thanks for testing the fix! As far as the ERROR: can't find special struct size msg goes, can you provide the kpatch-build command line and if you've got a stock or custom kernel installed?

[marcin-github]

Thanks for fix:) I'm using custom kernel. I've copied source tree to temporary directory, then I run:
./kpatch-build -s /var/tmp/portage/linux-bcache-kpatch/ /tmp/meminfo-test.patch

[joe-lawrence]

That kpatch-build invocation looks correct, so open a new issue and add: gcc version, output from that readelf | gawk script in kpatch-build that you've noted. Attaching your kernel .config might be useful as well. Thanks.

[ric96]

applied the above mentioned patch.... anyone getting this:
[ric@localhost kpatch-stuff]$ kpatch-build -t vmlinux meminfo-string.patch
Using cache at /home/ric/.kpatch/src
Testing patch file
checking file fs/proc/meminfo.c
Hunk #1 FAILED at 95.
1 out of 1 hunk FAILED
ERROR: source patch file failed to apply.
[ric@localhost kpatch-stuff]$

the patch applies fine using git apply, same patch as that of the example
kernel source: 4.9.6-200.fc25.x86_64

[jpoimboe]

@ric96 the meminfo_proc_show() function changed in Linux 4.9, so the old meminfo-string.patch no longer applies.

[ric96]

@jpoimboe oh, I am kinda new to this kpatch stuff, can you suggest another patch that's suitable for show and tell.

[jpoimboe]

@ric96 I'd recommend to just port the meminfo patch for your newer kernel, like:

diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index 8a42849..d8cf031 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -132,7 +132,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
        seq_printf(m, "VmallocTotal:   %8lu kB\n",
                   (unsigned long)VMALLOC_TOTAL >> 10);
        show_val_kb(m, "VmallocUsed:    ", 0ul);
-       show_val_kb(m, "VmallocChunk:   ", 0ul);
+       show_val_kb(m, "VMALLOCCHUNK:   ", 0ul);
 
 #ifdef CONFIG_MEMORY_FAILURE
        seq_printf(m, "HardwareCorrupted: %5lu kB\n",