<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="/rss.xsl.xml"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
    <title>Changes in Makefile</title>
    <description></description>
    <language>en</language>
    <copyright>Copyright 2015</copyright>
    <generator>Java</generator><item>
        <title>ed3994ac - selftests/fchmodat2: fix clang build failure due to -static-libasan</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/tools/testing/selftests/fchmodat2/Makefile#ed3994ac</link>
        <description>selftests/fchmodat2: fix clang build failure due to -static-libasangcc requires -static-libasan in order to ensure that Address Sanitizer&apos;slibrary is the first one loaded. However, this leads to build failureson clang, when building via:    make LLVM=1 -C tools/testing/selftestsHowever, clang already does the right thing by default: it staticallylinks the Address Sanitizer if -fsanitize is specified. Therefore,simply omit -static-libasan for clang builds. And leave behind acomment, because the whole reason for static linking might not beobvious.Cc: Ryan Roberts &lt;ryan.roberts@arm.com&gt;Signed-off-by: John Hubbard &lt;jhubbard@nvidia.com&gt;Reviewed-by: Nathan Chancellor &lt;nathan@kernel.org&gt;Signed-off-by: Shuah Khan &lt;skhan@linuxfoundation.org&gt;

            List of files:
            /linux-6.15/tools/testing/selftests/fchmodat2/Makefile</description>
        <pubDate>Sat, 01 Jun 2024 01:45:34 +0000</pubDate>
        <dc:creator>John Hubbard &lt;jhubbard@nvidia.com&gt;</dc:creator>
    </item>
<item>
        <title>c652df8a - selftests: link libasan statically for tests with -fsanitize=address</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/tools/testing/selftests/fchmodat2/Makefile#c652df8a</link>
        <description>selftests: link libasan statically for tests with -fsanitize=addressWhen dynamically linking, Address Sanitizer requires its library to be thefirst one to be loaded; this is apparently to ensure that every call tomalloc is intercepted.  If using LD_PRELOAD, those listed libraries willbe loaded before the libraries listed in the program&apos;s ELF and willtherefore violate this requirement, leading to the below failure andoutput from ASan.commit 58e2847ad2e6 (&quot;selftests: line buffer test program&apos;s stdout&quot;)modified the kselftest runner to force line buffering by forcing the testprograms to run through `stdbuf`.  It turns out that stdbuf implementsline buffering by injecting a library via LD_PRELOAD.  Therefore selfteststhat use ASan started failing.Fix this by statically linking libasan in the affected test programs,using the `-static-libasan` option.  Note this is already the default forClang, but not got GCC.Test output sample for failing case:  TAP version 13  1..3  # timeout set to 300  # selftests: openat2: openat2_test  # ==4052==ASan runtime does not come first in initial library list;  you should either link runtime to your application or manually preload  it with LD_PRELOAD.  not ok 1 selftests: openat2: openat2_test # exit=1  # timeout set to 300  # selftests: openat2: resolve_test  # ==4070==ASan runtime does not come first in initial library list;  you should either link runtime to your application or manually preload  it with LD_PRELOAD.  not ok 2 selftests: openat2: resolve_test # exit=1Link: https://lkml.kernel.org/r/20230912135048.1755771-1-ryan.roberts@arm.comSigned-off-by: Ryan Roberts &lt;ryan.roberts@arm.com&gt;Fixes: 58e2847ad2e6 (&quot;selftests: line buffer test program&apos;s stdout&quot;)Reported-by: kernel test robot &lt;oliver.sang@intel.com&gt;Closes: https://lore.kernel.org/oe-lkp/202309121342.97e2f008-oliver.sang@intel.comCc: David Hildenbrand &lt;david@redhat.com&gt;Cc: Florent Revest &lt;revest@chromium.org&gt;Cc: J&#233;r&#244;me Glisse &lt;jglisse@redhat.com&gt;Cc: John Hubbard &lt;jhubbard@nvidia.com&gt;Cc: Mark Brown &lt;broonie@kernel.org&gt;Cc: Peter Xu &lt;peterx@redhat.com&gt;Cc: Shuah Khan &lt;shuah@kernel.org&gt;Cc: Tom Rix &lt;trix@redhat.com&gt;Signed-off-by: Andrew Morton &lt;akpm@linux-foundation.org&gt;

            List of files:
            /linux-6.15/tools/testing/selftests/fchmodat2/Makefile</description>
        <pubDate>Tue, 12 Sep 2023 13:50:48 +0000</pubDate>
        <dc:creator>Ryan Roberts &lt;ryan.roberts@arm.com&gt;</dc:creator>
    </item>
<item>
        <title>71214379 - selftests: fchmodat2: remove duplicate unneeded defines</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/tools/testing/selftests/fchmodat2/Makefile#71214379</link>
        <description>selftests: fchmodat2: remove duplicate unneeded definesThese duplicate defines should automatically be picked up from kernelheaders. Use KHDR_INCLUDES to add kernel header files.Signed-off-by: Muhammad Usama Anjum &lt;usama.anjum@collabora.com&gt;Message-Id: &lt;20230805073809.1753462-4-usama.anjum@collabora.com&gt;Signed-off-by: Christian Brauner &lt;brauner@kernel.org&gt;

            List of files:
            /linux-6.15/tools/testing/selftests/fchmodat2/Makefile</description>
        <pubDate>Sat, 05 Aug 2023 07:38:02 +0000</pubDate>
        <dc:creator>Muhammad Usama Anjum &lt;usama.anjum@collabora.com&gt;</dc:creator>
    </item>
<item>
        <title>4859c257 - selftests: Add fchmodat2 selftest</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/tools/testing/selftests/fchmodat2/Makefile#4859c257</link>
        <description>selftests: Add fchmodat2 selftestThe test marks as skipped if a syscall with the AT_SYMLINK_NOFOLLOW flagfails. This is because not all filesystems support changing the modebits of symlinks properly. These filesystems return an error but changethe mode bits:newfstatat(4, &quot;regfile&quot;, {st_mode=S_IFREG|0640, st_size=0, ...}, AT_SYMLINK_NOFOLLOW) = 0newfstatat(4, &quot;symlink&quot;, {st_mode=S_IFLNK|0777, st_size=7, ...}, AT_SYMLINK_NOFOLLOW) = 0syscall_0x1c3(0x4, 0x55fa1f244396, 0x180, 0x100, 0x55fa1f24438e, 0x34) = -1 EOPNOTSUPP (Operation not supported)newfstatat(4, &quot;regfile&quot;, {st_mode=S_IFREG|0640, st_size=0, ...}, AT_SYMLINK_NOFOLLOW) = 0This happens with btrfs and xfs: $ tools/testing/selftests/fchmodat2/fchmodat2_test TAP version 13 1..1 ok 1 # SKIP fchmodat2(symlink) # Totals: pass:0 fail:0 xfail:0 xpass:0 skip:1 error:0 $ stat /tmp/ksft-fchmodat2.*/symlink   File: /tmp/ksft-fchmodat2.3NCqlE/symlink -&gt; regfile   Size: 7               Blocks: 0          IO Block: 4096   symbolic link Device: 7,0     Inode: 133         Links: 1 Access: (0600/lrw-------)  Uid: (    0/    root)   Gid: (    0/    root)Signed-off-by: Alexey Gladkov &lt;legion@kernel.org&gt;Message-Id: &lt;4532a04a870ff589ba62ceeacf76f0bd81b9ba01.1689092120.git.legion@kernel.org&gt;Signed-off-by: Christian Brauner &lt;brauner@kernel.org&gt;

            List of files:
            /linux-6.15/tools/testing/selftests/fchmodat2/Makefile</description>
        <pubDate>Tue, 11 Jul 2023 16:16:07 +0000</pubDate>
        <dc:creator>Alexey Gladkov &lt;legion@kernel.org&gt;</dc:creator>
    </item>
</channel>
</rss>
