[sanitizer] Remove #include <linux/fs.h> to resolve fsconfig_command/mount_attr conflict with glibc 2.36

[MaskRay]

Summary

It is generally not a good idea to mix usage of glibc headers and Linux UAPI
headers (https://sourceware.org/glibc/wiki/Synchronizing_Headers). In glibc
since 7eae6a91e9b1670330c9f15730082c91c0b1d570 (milestone: 2.36), sys/mount.h
defines fsconfig_command which conflicts with linux/mount.h:

.../usr/include/linux/mount.h:95:6: error: redeclaration of ‘enum fsconfig_command’

Remove #include <linux/fs.h> which pulls in linux/mount.h. Expand its 4 macros manually.

Fix https://github.com/llvm/llvm-project/issues/56421

[zatrazz]

LGTM, the kernel exported value is same for all architectures. Anther possibility would be to split the sanitizer_platform_limits_posix.cpp and add a sanitizer_platform_limits_posix_linux.cpp which only add linux includes.

[MaskRay]

In D129471#3642148, @zatrazz wrote:

LGTM, the kernel exported value is same for all architectures. Anther possibility would be to split the sanitizer_platform_limits_posix.cpp and add a sanitizer_platform_limits_posix_linux.cpp which only add linux includes.

If we are going to do this, sanitizer_platform_limits_linux.cpp may be the choice.

Do you think other #include <linux/*.h> in this file may cause a potential conflict in the future?

[zatrazz]

In D129471#3643096, @MaskRay wrote:

In D129471#3642148, @zatrazz wrote:

LGTM, the kernel exported value is same for all architectures. Anther possibility would be to split the sanitizer_platform_limits_posix.cpp and add a sanitizer_platform_limits_posix_linux.cpp which only add linux includes.

If we are going to do this, sanitizer_platform_limits_linux.cpp may be the choice.

Do you think other #include <linux/*.h> in this file may cause a potential conflict in the future?

I am not sure, it has bitten us in the past which required either glibc to fixup by including the kernel header
or the kernel to fixup some internal definitions. I personally prefer to keep glibc header not dependent of the
kernel ones, even at cost of extra maintainability (however I think this stance is not shared with
other glibc maintainers).

I think a potential sanitizer_platform_limits_linux.cpp may be a better approach to disantagle the kernel
include from libc one.

[hctim]

Looks like this has broken the sanitizer-android bot build, PTAL: https://lab.llvm.org/buildbot/#/builders/77/builds/19507

[MaskRay]

In D129471#3643146, @zatrazz wrote:

In D129471#3643096, @MaskRay wrote:

In D129471#3642148, @zatrazz wrote:

LGTM, the kernel exported value is same for all architectures. Anther possibility would be to split the sanitizer_platform_limits_posix.cpp and add a sanitizer_platform_limits_posix_linux.cpp which only add linux includes.

If we are going to do this, sanitizer_platform_limits_linux.cpp may be the choice.

Do you think other #include <linux/*.h> in this file may cause a potential conflict in the future?

I am not sure, it has bitten us in the past which required either glibc to fixup by including the kernel header
or the kernel to fixup some internal definitions. I personally prefer to keep glibc header not dependent of the
kernel ones, even at cost of extra maintainability (however I think this stance is not shared with
other glibc maintainers).

musl's stance: it supports the inclusion of the kernel headers after the libc ones
https://git.musl-libc.org/cgit/musl/commit/?id=04983f2272382af92eb8f8838964ff944fbb8258

I prefer that glibc's header not dependent of the kernel ones, too, to support building using older Linux uapi.
And inclusion of libc headers after uapi headers is unsupported.

For this sys/mount.h issue, the commit should be easily cherry picked into older compiler-rt (<15.0.0) to help distributions.

I think a potential sanitizer_platform_limits_linux.cpp may be a better approach to disantagle the kernel
include from libc one.

LG. I can investigate how to place all these linux/*.h headers elsewhere:

#include <linux/if_eql.h>
#include <linux/if_plip.h>
#include <linux/lp.h>
#include <linux/mroute.h>
#include <linux/mroute6.h>
#include <linux/scc.h>
#include <linux/serial.h>

[MaskRay]

In D129471#3643267, @hctim wrote:

Looks like this has broken the sanitizer-android bot build, PTAL: https://lab.llvm.org/buildbot/#/builders/77/builds/19507

Testing and will revert this and push an amended one in one minute.

[MaskRay]

In D129471#3643321, @MaskRay wrote:

In D129471#3643267, @hctim wrote:

Looks like this has broken the sanitizer-android bot build, PTAL: https://lab.llvm.org/buildbot/#/builders/77/builds/19507

Testing and will revert this and push an amended one in one minute.

By the way, do you have instructions testing an Android build without an Android device?
For now, I am going to use

+#if SANITIZER_ANDROID
 #include <linux/fs.h>
+#endif

I plan to do some #include refactoring and having a pre-commit test method will be useful...

[hctim]

In D129471#3643323, @MaskRay wrote:

In D129471#3643321, @MaskRay wrote:

In D129471#3643267, @hctim wrote:

Looks like this has broken the sanitizer-android bot build, PTAL: https://lab.llvm.org/buildbot/#/builders/77/builds/19507

Testing and will revert this and push an amended one in one minute.

By the way, do you have instructions testing an Android build without an Android device?
For now, I am going to use

+#if SANITIZER_ANDROID
 #include <linux/fs.h>
+#endif

I plan to do some #include refactoring and having a pre-commit test method will be useful...

Just the build? https://github.com/google/sanitizers/wiki/SanitizerBotReproduceBuild should help out, use buildbot_android.sh, and you can comment out line 46 (test_android ...).

compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cpp

 #if SANITIZER_LINUX
 #include <malloc.h>
 #include <mntent.h>
 #include <netinet/ether.h>
 #include <sys/sysinfo.h>
 #include <sys/vt.h>
 #include <linux/cdrom.h>
 #include <linux/fd.h>
-#include <linux/fs.h>
 #include <linux/hdreg.h>
 #include <linux/input.h>
 #include <linux/ioctl.h>
 #include <linux/soundcard.h>
 #include <linux/sysctl.h>
 #include <linux/utsname.h>
 #include <linux/posix_types.h>
 #include <net/if_arp.h>
#if EV_VERSION > (0x010000)
   unsigned IOCTL_EVIOCGKEYCODE_V2 = EVIOCGKEYCODE_V2;
   unsigned IOCTL_EVIOCGPROP = EVIOCGPROP(0);
   unsigned IOCTL_EVIOCSKEYCODE_V2 = EVIOCSKEYCODE_V2;
 #else
   unsigned IOCTL_EVIOCGKEYCODE_V2 = IOCTL_NOT_PRESENT;
   unsigned IOCTL_EVIOCGPROP = IOCTL_NOT_PRESENT;
   unsigned IOCTL_EVIOCSKEYCODE_V2 = IOCTL_NOT_PRESENT;
 #endif
-  unsigned IOCTL_FS_IOC_GETFLAGS = FS_IOC_GETFLAGS;
-  unsigned IOCTL_FS_IOC_GETVERSION = FS_IOC_GETVERSION;
-  unsigned IOCTL_FS_IOC_SETFLAGS = FS_IOC_SETFLAGS;
-  unsigned IOCTL_FS_IOC_SETVERSION = FS_IOC_SETVERSION;
+  unsigned IOCTL_FS_IOC_GETFLAGS = _IOR('f', 1, long);
+  unsigned IOCTL_FS_IOC_GETVERSION = _IOR('v', 1, long);
+  unsigned IOCTL_FS_IOC_SETFLAGS = _IOW('f', 2, long);
+  unsigned IOCTL_FS_IOC_SETVERSION = _IOW('v', 2, long);
   unsigned IOCTL_GIO_CMAP = GIO_CMAP;
   unsigned IOCTL_GIO_FONT = GIO_FONT;
   unsigned IOCTL_GIO_UNIMAP = GIO_UNIMAP;
   unsigned IOCTL_GIO_UNISCRNMAP = GIO_UNISCRNMAP;
   unsigned IOCTL_KDADDIO = KDADDIO;
   unsigned IOCTL_KDDELIO = KDDELIO;
   unsigned IOCTL_KDGETKEYCODE = KDGETKEYCODE;
   unsigned IOCTL_KDGKBDIACR = KDGKBDIACR;