[PATCH RFC v02 00/45] Compile test and fixes for exported header files

[PATCH RFC v02 00/45] Compile test and fixes for exported header files

[Mikko Rapeli]

v2:
added cross compiler support with CROSS_COMPILE,
detecting libc and GCC headers from compiler,
more header file fixes

v1:
https://lkml.org/lkml/2014/8/21/665

Users of kernel header files would be happier if they did not contain
kernel specific parts and would contain #include statements for all
other header files that they depend on, and in general would compile.

This patch set fixes some headers_install issues, then introduces a
compile test for headers exported to userspace and then fixes many of the
userspace header compilation failures.

Changes are also available in headers_test_v02 branch at github:
https://github.com/mcfrisk/linux/tree/headers_test_v02

Mikko Rapeli (45):
  Makefile: propagate verbose options
  scripts/headers.sh: add verbose option to make calls if defined
  headers_install.sh: enhance error handling
  scripts/headers_compile_test.sh: compile test script for exported
    headers
  drm.h: include stdlib.h in userspace
  drm_mode.h: include stdint.h and linux/types.h in userspace
  exynos_drm.h: include stdint.h in userspace
  nouveau_drm.h: include stdint.h in userspace
  radeon_drm.h: include stdint.h in userspace
  r128_drm.h: include drm/drm.h
  via_drm.h: include linux/types.h instead of non-existing
    via_drmclient.h
  via_drm.h: hide struct via_file_private in userspace
  savage_drm.h: include <drm/drm.h>
  sis_drm.h: hide sis_file_private in userspace
  dm-log-userspace.h: include stdint.h in userspace
  hsi_char.h: include stdint.h in userspace
  ebtables.h: include stdint.h in userspace
  cld.h: include stdint.h in userspace
  rds.h: include stdint.h in userspace
  sctp.h: include stdint.h in userspace
  scsi_bsg_fc.h: include stdint.h in userspace
  scsi_netlink.h: include stdint.h in userspace
  scsi_netlink_fc.h: include stdint.h in userspace
  hdspm.h: include stdint.h in userspace
  gntalloc.h: include stdint.h in userspace
  gntdev.h: include stdint.h in userspace
  include/uapi/linux/sysctl.h: include stdlib.h in userspace
  include/uapi/asm-generic/ipcbuf.h: include linux/posix_types.h
  include/uapi/asm-generic/msgbuf.h: include asm/ipcbuf.h
  include/uapi/asm-generic/shmbuf.h: include fixes
  asm-generic/sembuf.h: include asm/posix_types.h and asm/ipcbuf.h
  drm/i810_drm.h: include drm/drm.h
  include/uapi/asm-generic/signal.h: include stdlib.h in userspace
  include/uapi/asm-generic/signal.h: hide sigset_t definition in
    userspace
  include/uapi/sound/asound.h: include stdlib.h in userspace
  include/uapi/linux/netfilter.h: include in.h and in6.h
  include/uapi/linux/socket.h: include sys/socket.h in userspace
  include/uapi/linux/agpgart.h: include stdlib.h in userspace
  include/uapi/sound/asequencer.h: include sound/asound.h
  include/uapi/sound/emu10k1.h: include sound/asound.h
  include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and
    code_valid in userspace
  include/uapi/linux/rds.h: include linux/socket.h and linux/types.h
    always
  include/uapi/linux/netfilter_bridge.h: include if.h
  nf_conntrack_tuple_common.h: include linux/types.h and
    linux/netfilter.h
  include/uapi/asm-generic/ucontext.h: include signal.h and sigcontext.h

 Makefile                                           |   8 +-
 include/uapi/asm-generic/ipcbuf.h                  |   2 +
 include/uapi/asm-generic/msgbuf.h                  |   2 +
 include/uapi/asm-generic/sembuf.h                  |   2 +
 include/uapi/asm-generic/shmbuf.h                  |   6 +
 include/uapi/asm-generic/signal.h                  |   7 +
 include/uapi/asm-generic/ucontext.h                |   3 +
 include/uapi/drm/drm.h                             |   4 +
 include/uapi/drm/drm_mode.h                        |   5 +
 include/uapi/drm/exynos_drm.h                      |   5 +
 include/uapi/drm/i810_drm.h                        |   2 +
 include/uapi/drm/nouveau_drm.h                     |   6 +
 include/uapi/drm/r128_drm.h                        |   2 +
 include/uapi/drm/radeon_drm.h                      |   5 +
 include/uapi/drm/savage_drm.h                      |   2 +
 include/uapi/drm/sis_drm.h                         |   2 +
 include/uapi/drm/via_drm.h                         |   6 +-
 include/uapi/linux/agpgart.h                       |   1 +
 include/uapi/linux/dm-log-userspace.h              |   5 +
 include/uapi/linux/hsi/hsi_char.h                  |   7 +-
 include/uapi/linux/netfilter.h                     |   3 +-
 .../linux/netfilter/nf_conntrack_tuple_common.h    |   3 +
 include/uapi/linux/netfilter_bridge.h              |   1 +
 include/uapi/linux/netfilter_bridge/ebtables.h     |   5 +
 include/uapi/linux/nfsd/cld.h                      |   6 +
 include/uapi/linux/rds.h                           |   4 +
 include/uapi/linux/sctp.h                          |   4 +
 include/uapi/linux/socket.h                        |   4 +
 include/uapi/linux/sysctl.h                        |   4 +
 include/uapi/scsi/scsi_bsg_fc.h                    |   6 +
 include/uapi/scsi/scsi_netlink.h                   |   6 +-
 include/uapi/scsi/scsi_netlink_fc.h                |   5 +
 include/uapi/sound/asequencer.h                    |   1 +
 include/uapi/sound/asound.h                        |   3 +
 include/uapi/sound/emu10k1.h                       |   9 +-
 include/uapi/sound/hdspm.h                         |   6 +
 include/uapi/xen/gntalloc.h                        |   6 +
 include/uapi/xen/gntdev.h                          |   6 +
 scripts/headers.sh                                 |   2 +-
 scripts/headers_compile_test.sh                    | 143 +++++++++++++++++++++
 scripts/headers_install.sh                         |  15 ++-
 41 files changed, 306 insertions(+), 18 deletions(-)
 create mode 100755 scripts/headers_compile_test.sh

-- 
2.1.4

[PATCH 01/45] Makefile: propagate verbose options

[Mikko Rapeli]

Usefull when debugging the make scripts.

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 Makefile | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/Makefile b/Makefile
index 4e93284..c1b515d 100644
--- a/Makefile
+++ b/Makefile
@@ -912,13 +912,13 @@ quiet_cmd_link-vmlinux = LINK    $@
 # execute if the rest of the kernel build went well.
 vmlinux: scripts/link-vmlinux.sh $(vmlinux-deps) FORCE
 ifdef CONFIG_HEADERS_CHECK
-	$(Q)$(MAKE) -f $(srctree)/Makefile headers_check
+	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) -f $(srctree)/Makefile headers_check
 endif
 ifdef CONFIG_SAMPLES
-	$(Q)$(MAKE) $(build)=samples
+	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(build)=samples
 endif
 ifdef CONFIG_BUILD_DOCSRC
-	$(Q)$(MAKE) $(build)=Documentation
+	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(build)=Documentation
 endif
 	+$(call if_changed,link-vmlinux)
 
@@ -1069,7 +1069,7 @@ headers_check_all: headers_install_all
 PHONY += headers_check
 headers_check: headers_install
 	$(Q)$(MAKE) $(hdr-inst)=include/uapi HDRCHECK=1
-	$(Q)$(MAKE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
+	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
 
 # ---------------------------------------------------------------------------
 # Kernel selftest
-- 
2.1.4

Re: [PATCH 01/45] Makefile: propagate verbose options

[Masahiro Yamada]

Hi Mikko,

On Tue, 17 Feb 2015 00:05:04 +0100
Mikko Rapeli <mikko.rapeli@iki.fi> wrote:

> Usefull when debugging the make scripts.
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  Makefile | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/Makefile b/Makefile
> index 4e93284..c1b515d 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -912,13 +912,13 @@ quiet_cmd_link-vmlinux = LINK    $@
>  # execute if the rest of the kernel build went well.
>  vmlinux: scripts/link-vmlinux.sh $(vmlinux-deps) FORCE
>  ifdef CONFIG_HEADERS_CHECK
> -	$(Q)$(MAKE) -f $(srctree)/Makefile headers_check
> +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) -f $(srctree)/Makefile headers_check
>  endif
>  ifdef CONFIG_SAMPLES
> -	$(Q)$(MAKE) $(build)=samples
> +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(build)=samples
>  endif
>  ifdef CONFIG_BUILD_DOCSRC
> -	$(Q)$(MAKE) $(build)=Documentation
> +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(build)=Documentation
>  endif
>  	+$(call if_changed,link-vmlinux)
>  
> @@ -1069,7 +1069,7 @@ headers_check_all: headers_install_all
>  PHONY += headers_check
>  headers_check: headers_install
>  	$(Q)$(MAKE) $(hdr-inst)=include/uapi HDRCHECK=1
> -	$(Q)$(MAKE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
> +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
>  
>  # ---------------------------------------------------------------------------
>  # Kernel selftest



If you give V=... from the command line, it is automatically propagated
to its sub-makes.

I think this patch has no point.


Best Regards
Masahiro Yamada

Re: [PATCH 01/45] Makefile: propagate verbose options

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 04:18:12PM +0900, Masahiro Yamada wrote:
> Hi Mikko,
> 
> On Tue, 17 Feb 2015 00:05:04 +0100
> Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> 
> > Usefull when debugging the make scripts.
> > 
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > ---
> >  Makefile | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/Makefile b/Makefile
> > index 4e93284..c1b515d 100644
> > --- a/Makefile
> > +++ b/Makefile
> > @@ -912,13 +912,13 @@ quiet_cmd_link-vmlinux = LINK    $@
> >  # execute if the rest of the kernel build went well.
> >  vmlinux: scripts/link-vmlinux.sh $(vmlinux-deps) FORCE
> >  ifdef CONFIG_HEADERS_CHECK
> > -	$(Q)$(MAKE) -f $(srctree)/Makefile headers_check
> > +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) -f $(srctree)/Makefile headers_check
> >  endif
> >  ifdef CONFIG_SAMPLES
> > -	$(Q)$(MAKE) $(build)=samples
> > +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(build)=samples
> >  endif
> >  ifdef CONFIG_BUILD_DOCSRC
> > -	$(Q)$(MAKE) $(build)=Documentation
> > +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(build)=Documentation
> >  endif
> >  	+$(call if_changed,link-vmlinux)
> >  
> > @@ -1069,7 +1069,7 @@ headers_check_all: headers_install_all
> >  PHONY += headers_check
> >  headers_check: headers_install
> >  	$(Q)$(MAKE) $(hdr-inst)=include/uapi HDRCHECK=1
> > -	$(Q)$(MAKE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
> > +	$(Q)$(MAKE) V=$(KBUILD_VERBOSE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
> >  
> >  # ---------------------------------------------------------------------------
> >  # Kernel selftest
> 
> 
> 
> If you give V=... from the command line, it is automatically propagated
> to its sub-makes.
> 
> I think this patch has no point.

You are right, I'll drop this patch.

-Mikko

[PATCH 02/45] scripts/headers.sh: add verbose option to make calls if defined

[Mikko Rapeli]

Useful when debugging the scripts.

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 scripts/headers.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/headers.sh b/scripts/headers.sh
index 95ece06..aa62fd8 100755
--- a/scripts/headers.sh
+++ b/scripts/headers.sh
@@ -7,7 +7,7 @@ set -e
 do_command()
 {
 	if [ -f ${srctree}/arch/$2/include/asm/Kbuild ]; then
-		make ARCH=$2 KBUILD_HEADERS=$1 headers_$1
+		make V=$KBUILD_VERBOSE ARCH=$2 KBUILD_HEADERS=$1 headers_$1
 	else
 		printf "Ignoring arch: %s\n" ${arch}
 	fi
-- 
2.1.4

Re: [PATCH 02/45] scripts/headers.sh: add verbose option to make calls if defined

[Masahiro Yamada]

On Tue, 17 Feb 2015 00:05:05 +0100
Mikko Rapeli <mikko.rapeli@iki.fi> wrote:

> Useful when debugging the scripts.
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  scripts/headers.sh | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/scripts/headers.sh b/scripts/headers.sh
> index 95ece06..aa62fd8 100755
> --- a/scripts/headers.sh
> +++ b/scripts/headers.sh
> @@ -7,7 +7,7 @@ set -e
>  do_command()
>  {
>  	if [ -f ${srctree}/arch/$2/include/asm/Kbuild ]; then
> -		make ARCH=$2 KBUILD_HEADERS=$1 headers_$1
> +		make V=$KBUILD_VERBOSE ARCH=$2 KBUILD_HEADERS=$1 headers_$1
>  	else
>  		printf "Ignoring arch: %s\n" ${arch}
>  	fi
> -- 


KBUILD_VERBOSE is already exported.

Why do you need to set V=... again here?


Best Regards
Masahiro Yamada

Re: [PATCH 02/45] scripts/headers.sh: add verbose option to make calls if defined

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 04:18:59PM +0900, Masahiro Yamada wrote:
> 
> On Tue, 17 Feb 2015 00:05:05 +0100
> Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> 
> > Useful when debugging the scripts.
> > 
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > ---
> >  scripts/headers.sh | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/scripts/headers.sh b/scripts/headers.sh
> > index 95ece06..aa62fd8 100755
> > --- a/scripts/headers.sh
> > +++ b/scripts/headers.sh
> > @@ -7,7 +7,7 @@ set -e
> >  do_command()
> >  {
> >  	if [ -f ${srctree}/arch/$2/include/asm/Kbuild ]; then
> > -		make ARCH=$2 KBUILD_HEADERS=$1 headers_$1
> > +		make V=$KBUILD_VERBOSE ARCH=$2 KBUILD_HEADERS=$1 headers_$1
> >  	else
> >  		printf "Ignoring arch: %s\n" ${arch}
> >  	fi
> > -- 
> 
> 
> KBUILD_VERBOSE is already exported.
> 
> Why do you need to set V=... again here?

You're right, I'll drop this too. Thanks for the review!

-Mikko

[PATCH 03/45] headers_install.sh: enhance error handling

[Mikko Rapeli]

Exit with error if using undefined variables or if any sub command fails
with error return value. unidef needs special handling since but this can
be done without the trap. Enables exaniming intermediate files if some
commands failed.

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 scripts/headers_install.sh | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/scripts/headers_install.sh b/scripts/headers_install.sh
index fdebd66..072f7d3 100755
--- a/scripts/headers_install.sh
+++ b/scripts/headers_install.sh
@@ -1,4 +1,7 @@
 #!/bin/sh
+set -e
+set -u
+#set -x
 
 if [ $# -lt 2 ]
 then
@@ -26,7 +29,6 @@ shift
 # Iterate through files listed on command line
 
 FILE=
-trap 'rm -f "$OUTDIR/$FILE" "$OUTDIR/$FILE.sed"' EXIT
 for i in "$@"
 do
 	FILE="$(basename "$i")"
@@ -37,10 +39,13 @@ do
 		-e 's/(^|[^a-zA-Z0-9])__packed([^a-zA-Z0-9_]|$)/\1__attribute__((packed))\2/g' \
 		-e 's/(^|[ \t(])(inline|asm|volatile)([ \t(]|$)/\1__\2__\3/g' \
 		-e 's@#(ifndef|define|endif[ \t]*/[*])[ \t]*_UAPI@#\1 @' \
-		"$SRCDIR/$i" > "$OUTDIR/$FILE.sed" || exit 1
+		"$SRCDIR/$i" > "$OUTDIR/$FILE.sed" || \
+		( rm -f "$OUTDIR/$FILE.sed" ; exit 1 )
 	scripts/unifdef -U__KERNEL__ -D__EXPORTED_HEADERS__ "$OUTDIR/$FILE.sed" \
-		> "$OUTDIR/$FILE"
-	[ $? -gt 1 ] && exit 1
+		> "$OUTDIR/$FILE" || \
+		( if [ $? -gt 1 ]; then \
+				rm -f "$OUTDIR/$FILE" "$OUTDIR/$FILE.sed" ; \
+				exit 1 ; \
+		  fi )
 	rm -f "$OUTDIR/$FILE.sed"
 done
-trap - EXIT
-- 
2.1.4

Re: [PATCH 03/45] headers_install.sh: enhance error handling

[Michal Marek]

On 2015-02-17 00:05, Mikko Rapeli wrote:
> Exit with error if using undefined variables or if any sub command fails
> with error return value. unidef needs special handling since but this can
> be done without the trap. Enables exaniming intermediate files if some
> commands failed.

The intermediate files are removed by your patch as well, aren't they?
Actually, I don't see any change in behavior after your patch.


> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  scripts/headers_install.sh | 15 ++++++++++-----
>  1 file changed, 10 insertions(+), 5 deletions(-)
> 
> diff --git a/scripts/headers_install.sh b/scripts/headers_install.sh
> index fdebd66..072f7d3 100755
> --- a/scripts/headers_install.sh
> +++ b/scripts/headers_install.sh
> @@ -1,4 +1,7 @@
>  #!/bin/sh
> +set -e
> +set -u
> +#set -x

The commented-out statement should be removed.

Michal

Re: [PATCH 03/45] headers_install.sh: enhance error handling

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 02:17:51PM +0100, Michal Marek wrote:
> On 2015-02-17 00:05, Mikko Rapeli wrote:
> > Exit with error if using undefined variables or if any sub command fails
> > with error return value. unidef needs special handling since but this can
> > be done without the trap. Enables exaniming intermediate files if some
> > commands failed.
> 
> The intermediate files are removed by your patch as well, aren't they?
> Actually, I don't see any change in behavior after your patch.

Yes, if everything succeeded without errors, then intermediate files
are removed. But the files are left around if something fails which is
what I like in scripts.

-Mikko

Re: [PATCH 03/45] headers_install.sh: enhance error handling

[Michal Marek]

On 2015-03-10 23:44, Mikko Rapeli wrote:
> On Tue, Feb 17, 2015 at 02:17:51PM +0100, Michal Marek wrote:
>> On 2015-02-17 00:05, Mikko Rapeli wrote:
>>> Exit with error if using undefined variables or if any sub command fails
>>> with error return value. unidef needs special handling since but this can
>>> be done without the trap. Enables exaniming intermediate files if some
>>> commands failed.
>>
>> The intermediate files are removed by your patch as well, aren't they?
>> Actually, I don't see any change in behavior after your patch.
> 
> Yes, if everything succeeded without errors, then intermediate files
> are removed. But the files are left around if something fails

They are not:

+		"$SRCDIR/$i" > "$OUTDIR/$FILE.sed" || \
+		( rm -f "$OUTDIR/$FILE.sed" ; exit 1 )

This removes "$OUTDIR/$FILE.sed" if the sed command files. Just like the
trap command has been doing.

Michal

Re: [PATCH 03/45] headers_install.sh: enhance error handling

[Mikko Rapeli]

On Wed, Mar 11, 2015 at 11:16:11AM +0100, Michal Marek wrote:
> On 2015-03-10 23:44, Mikko Rapeli wrote:
> > On Tue, Feb 17, 2015 at 02:17:51PM +0100, Michal Marek wrote:
> >> On 2015-02-17 00:05, Mikko Rapeli wrote:
> >>> Exit with error if using undefined variables or if any sub command fails
> >>> with error return value. unidef needs special handling since but this can
> >>> be done without the trap. Enables exaniming intermediate files if some
> >>> commands failed.
> >>
> >> The intermediate files are removed by your patch as well, aren't they?
> >> Actually, I don't see any change in behavior after your patch.
> > 
> > Yes, if everything succeeded without errors, then intermediate files
> > are removed. But the files are left around if something fails
> 
> They are not:
> 
> +		"$SRCDIR/$i" > "$OUTDIR/$FILE.sed" || \
> +		( rm -f "$OUTDIR/$FILE.sed" ; exit 1 )
> 
> This removes "$OUTDIR/$FILE.sed" if the sed command files. Just like the
> trap command has been doing.

Heh, you're right. I might as well change this to leave the sed script
around if something failed.

-Mikko

[PATCH 04/45] scripts/headers_compile_test.sh: compile test script for exported headers

[Mikko Rapeli]

Users of kernel header files would be happier if they did not contain
kernel specific parts and would contain #include statements for all
other header files that they depend on, and in general would compile.

For each header file exported to userspace, this script creates
a simple .c file which just includes the header file. Then it
tries to compile it together with minimal header files from GCC
and libc, and reports results.

Default libc and GCC header file locations are parsed from compiler
configuration.

Kernel headers depend on GCC headers so their path is included in
the test compiler command line.

Some gcc and kernel headers depend on libc headers which are made available
by copying from the compiler default location to a temporary location and
removing possibly existing kernel headers from this directory. This is a bit
of a hack but seems to work in multiple environments.

Tested natively on:

Debian unstable, i586-linux-gnu and gcc 4.9.2
Raspbian Wheezy, arm-linux-gnueabihf and gcc 4.6.3
Ubuntu 12.04 LTS, x86_64-linux-gnu and gcc 4.6.3

Tested cross compilation using standard CROSS_COMPILE=/path/to/gcc with:

arm-linux-gnueabihf-gcc (crosstool-NG linaro-1.13.1-4.9-2014.09 - Linaro GCC 4.9-2014.09) 4.9.2 20140904 (prerelease)

Execute the script in the directory where kernel header files are installed.
For example:

$ make headers_install
$ cd usr/include
$ ../../scripts/headers_compile_test.sh

Example statistics from 3.18.4 kernel:

130 files failed the compile test.
609 files passed the compile test.

Example error types from 3.18.4 kernel:

$ ../../scripts/headers_compile_test.sh 2>&1 | \
grep error: | sed -e 's/.*error://g' | sort | uniq -c | sort -rn

    352  unknown type name ‘uint32_t’
    134  unknown type name ‘uint64_t’
    103  unknown type name ‘size_t’
     43  unknown type name ‘__kernel_ulong_t’
     38  unknown type name ‘uint8_t’
     24  unknown type name ‘int32_t’
     22  field ‘addr’ has incomplete type
     18  field ‘tstamp’ has incomplete type
     16  unknown type name ‘__kernel_time_t’
     16  field ‘in’ has incomplete type
     16  field ‘in6’ has incomplete type
     14  unknown type name ‘__be16’
     13  ‘IFNAMSIZ’ undeclared here (not in a function)
     12  unknown type name ‘uint16_t’
      9  field ‘ifru_netmask’ has incomplete type
      9  field ‘ifru_hwaddr’ has incomplete type
      9  field ‘ifru_dstaddr’ has incomplete type
      9  field ‘ifru_broadaddr’ has incomplete type
      9  field ‘ifru_addr’ has incomplete type
      8  unknown type name ‘__kernel_pid_t’
      7  unknown type name ‘u_short’
      7  unknown type name ‘pid_t’
      6  invalid application of ‘sizeof’ to incomplete type ‘struct timespec’
      6  field ‘src’ has incomplete type
      6  field ‘audio_tstamp’ has incomplete type
      6  array type has incomplete element type
      5  unknown type name ‘__kernel_long_t’
      5  requested alignment is not an integer constant
      5  field ‘smsk’ has incomplete type
      4  unknown type name ‘__kernel_uid32_t’
      4  unknown type name ‘__kernel_gid32_t’
      4  ‘ETH_ALEN’ undeclared here (not in a function)
      3  unknown type name ‘int64_t’
      3  unknown type name ‘caddr_t’
      3  ‘IPSET_ERR_TYPE_SPECIFIC’ undeclared here (not in a function)
      3  field ‘trigger_tstamp’ has incomplete type
      3  field ‘src_addr’ has incomplete type
      3  field ‘sin_addr’ has incomplete type
      3  field ‘laddr’ has incomplete type
      3  field ‘id’ has incomplete type
      3  field ‘dmsk’ has incomplete type
      3  field ‘bssid’ has incomplete type
      3  expected specifier-qualifier-list before ‘uint64_t’
      2  unknown type name ‘u_long’
      2  unknown type name ‘stack_t’
      2  unknown type name ‘sigset_t’
      2  unknown type name ‘sa_family_t’
      2  unknown type name ‘__kernel_mode_t’
      2  unknown type name ‘__kernel_key_t’
      2  unknown type name ‘elf_gregset_t’
      2  unknown type name ‘bool’
      2  ‘uint64_t’ undeclared here (not in a function)
      2  ‘true’ undeclared (first use in this function)
      2  ‘NAME_MAX’ undeclared here (not in a function)
      2  ‘__kernel_mode_t’ undeclared here (not in a function)
      2  invalid application of ‘sizeof’ to incomplete type ‘struct sockaddr’
      2  field ‘uc_mcontext’ has incomplete type
      2  field ‘tmsk’ has incomplete type
      2  field ‘tgt’ has incomplete type
      2  field ‘shm_perm’ has incomplete type
      2  field ‘sem_perm’ has incomplete type
      2  field ‘raddr’ has incomplete type
      2  field ‘msg_perm’ has incomplete type
      2  field ‘grp’ has incomplete type
      2  field ‘dst’ has incomplete type
      2  field ‘dst_addr’ has incomplete type
      2  field ‘arp_pa’ has incomplete type
      2  field ‘arp_netmask’ has incomplete type
      2  field ‘arp_ha’ has incomplete type
      2  ‘false’ undeclared (first use in this function)
      1  xen/interface/xen.h: No such file or directory
      1  via_drmclient.h: No such file or directory
      1  unknown type name ‘wait_queue_head_t’
      1  unknown type name ‘snd_seq_client_type_t’
      1  unknown type name ‘int16_t’
      1  unknown type name ‘ino_t’
      1  unknown type name ‘elf_greg_t’
      1  unknown type name ‘elf_fpxregset_t’
      1  unknown type name ‘elf_fpregset_t’
      1  unknown type name ‘__be32’
      1  ‘SIOCDEVPRIVATE’ undeclared here (not in a function)
      1  ‘sa_family_t’ undeclared here (not in a function)
      1  ‘NULL’ undeclared (first use in this function)
      1  ‘MSG_FIN’ undeclared here (not in a function)
      1  ‘MAX_IPOPTLEN’ undeclared here (not in a function)
      1  ‘MAX_ADDR_LEN’ undeclared here (not in a function)
      1  ‘IFHWADDRLEN’ undeclared here (not in a function)
      1  field ‘vmask’ has incomplete type
      1  field ‘vifc_rmt_addr’ has incomplete type
      1  field ‘vifc_lcl_addr’ has incomplete type
      1  field ‘vaddr’ has incomplete type
      1  field ‘uc_chain’ has incomplete type
      1  field ‘tgt_ip’ has incomplete type
      1  field ‘tcp’ has incomplete type
      1  field ‘sw_reserved’ has incomplete type
      1  field ‘sspp_addr’ has incomplete type
      1  field ‘ssp_addr’ has incomplete type
      1  field ‘src_mask’ has incomplete type
      1  field ‘src_ip’ has incomplete type
      1  field ‘spt_address’ has incomplete type
      1  field ‘spp_address’ has incomplete type
      1  field ‘spinfo_address’ has incomplete type
      1  field ‘spc_aaddr’ has incomplete type
      1  field ‘sas_obs_rto_ipaddr’ has incomplete type
      1  field ‘saddr’ has incomplete type
      1  field ‘rtmsg_src’ has incomplete type
      1  field ‘rtmsg_gateway’ has incomplete type
      1  field ‘rtmsg_dst’ has incomplete type
      1  field ‘rt_genmask’ has incomplete type
      1  field ‘rt_gateway’ has incomplete type
      1  field ‘rt_dst’ has incomplete type
      1  field ‘real’ has incomplete type
      1  field ‘prefix’ has incomplete type
      1  field ‘obj_list’ has incomplete type
      1  field ‘mfcc_origin’ has incomplete type
      1  field ‘mfcc_mcastgrp’ has incomplete type
      1  field ‘mf6cc_origin’ has incomplete type
      1  field ‘mf6cc_mcastgrp’ has incomplete type
      1  field ‘mask’ has incomplete type
      1  field ‘ival2’ has incomplete type
      1  field ‘ival1’ has incomplete type
      1  field ‘iph’ has incomplete type
      1  field ‘ip’ has incomplete type
      1  field ‘im_src’ has incomplete type
      1  field ‘im_dst’ has incomplete type
      1  field ‘im6_src’ has incomplete type
      1  field ‘im6_dst’ has incomplete type
      1  field ‘gw’ has incomplete type
      1  field ‘expected’ has incomplete type
      1  field ‘dst_mask’ has incomplete type
      1  field ‘dest_addr’ has incomplete type
      1  field ‘daddr’ has incomplete type
      1  field ‘ap_addr’ has incomplete type
      1  field ‘a6’ has incomplete type
      1  field ‘a4’ has incomplete type
      1  expected specifier-qualifier-list before ‘DECLARE_BITMAP’
      1  expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘snd_seq_client_type_t’
      1  #error "patchkey.h included directly"
      1  ‘DLM_RESNAME_MAXLEN’ undeclared here (not in a function)

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 scripts/headers_compile_test.sh | 143 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 143 insertions(+)
 create mode 100755 scripts/headers_compile_test.sh

diff --git a/scripts/headers_compile_test.sh b/scripts/headers_compile_test.sh
new file mode 100755
index 0000000..9a73945
--- /dev/null
+++ b/scripts/headers_compile_test.sh
@@ -0,0 +1,143 @@
+#!/bin/bash
+# bash due to arithmetics and pipefail
+set -euo pipefail
+
+# Debugging
+#set -x
+
+echo Simple compile test for header files meant for userspace.
+
+# sanity test
+if [ ! -d ./linux ]; then
+	echo Sanity check error: ./linux directory not found
+	echo Should be called in usr/include after \'make headers_install\'.
+	echo Returns number of failed files, 0 if none.
+	exit 1
+fi
+
+# Support CC variable for compiler and ccache, and cross compiling.
+# CC is used without quotes to support CC="ccache gcc".
+set +u
+if [ "$CC"foobar == "foobar" ]; then
+	CC=cc
+fi
+
+if [ "$CROSS_COMPILE"foobar != "foobar" ]; then
+	# Using gcc name since some cross compiler tool chains don't provide
+	# the cc symlink
+	CC="$CROSS_COMPILE"gcc
+fi
+set -u
+
+# Kernel headers refer to some gcc and libc headers so make them available.
+set +u
+if [ "$ARCH_TRIPLET"foobar == "foobar" ]; then
+	# Taking triplet from gcc/cpp
+	ARCH_TRIPLET="$( $CC -v -x c -E - < /dev/null 2>&1 | \
+			grep Target | sed -e 's/Target: //' )"
+fi
+
+if [ "$LIBC"foobar == "foobar" ]; then
+	# trying to grep libc includes from gcc/cpp defaults
+	_TEMP="$( $CC -v -x c -E - < /dev/null 2>&1 | \
+		sed -n -e '/^#include <...> search starts here:$/,/^End of search list.$/{//!p}' | \
+		sed -e 's/^\ \//\//g' | \
+		grep '/usr/include' )"
+
+	# sanity check and prepare LIBC dirs
+	for d in $_TEMP; do
+		if [ ! -d "$d" ]; then
+			echo "$d not a directory"
+			exit 1
+		fi
+		LIBC="$LIBC $d"
+	done
+fi
+set -u
+
+# Copy libc include files to temp directory for the tests.
+
+COMPILE_TEST_INC=../headers_compile_test_include
+rm -rf "$COMPILE_TEST_INC"
+mkdir -p "$COMPILE_TEST_INC"
+
+for d in $LIBC; do
+	# check if last part of dir is the arch triplet, cross compile paths
+	# can have it also elsewhere so just the last one counts.
+	if !( echo "$d" | egrep "$ARCH_TRIPLET$" > /dev/null ); then
+		# hopefully just main libc dir, e.g. /usr/include
+		cp -a "$d"/* "$COMPILE_TEST_INC"/
+	elif ( echo "$d" | egrep "$ARCH_TRIPLET$" > /dev/null ); then
+		# hopefully the arch specific dir, e.g. /usr/include/x86_64-linux-gnu
+		cp -ar "$d"/* "$COMPILE_TEST_INC/"
+	else
+		echo "$d unexpected, bailing out"
+		exit 1
+	fi
+done
+
+# Simulate libc headers without kernel headers by removing
+# all known kernel header dirs from the copied libc ones.
+# This seems to magically work.
+_KERNEL_DIRS="$( find . -type d | grep -v '^\.$' )"
+( cd "$COMPILE_TEST_INC" && rm -rf $_KERNEL_DIRS )
+
+# GCC headers
+set +u
+if [ "$GCC_INC"foobar == "foobar" ]; then
+	# Take from $CC default system include paths, filter out
+	# /usr/local/include and /usr/include stuff first, then try to match
+	# for gcc.
+	_TEMP="$( $CC -v -x c -E - < /dev/null 2>&1 | \
+		sed -n -e '/^#include <...> search starts here:$/,/^End of search list.$/{//!p}' | \
+		sed -e 's/^\ \//\//g' | \
+		egrep -v '/usr/local/include' | \
+		egrep -v '/usr/include' | \
+		grep gcc | \
+		xargs )"
+
+	# merge and prepare for use with $CC
+	for d in $_TEMP; do
+		# sanity test
+		if [ ! -d "$d" ]; then
+			echo "$d: is not a directory"
+			exit 1
+		fi
+		GCC_INC="$GCC_INC -I $d"
+	done
+fi
+set -u
+
+# For each header file, create a .c which includes the header file
+# and try to compile it using only current directory for searching other
+# included header files.
+_FAILED=0
+_PASSED=0
+for f in $( find . -name "*\.h" | xargs ); do
+	_FAIL=0
+	CFILE="$( echo "$( dirname "$f" )"/"$( basename "$f" .h )".c )"
+
+	# create .c file
+	echo "#include <"$( echo "$f" | sed -e 's|^.\/||' )">" \
+		> "$(dirname "$f")"/"$(basename "$f" .h)".c
+
+	# compile test, CC not quoted to support ccache
+	echo $CC -Wall -c -nostdinc $GCC_INC -I . -I "$COMPILE_TEST_INC" -I "$COMPILE_TEST_INC/$ARCH_TRIPLET" "$CFILE"
+	$CC -Wall -c -nostdinc $GCC_INC -I . -I "$COMPILE_TEST_INC" -I "$COMPILE_TEST_INC/$ARCH_TRIPLET" "$CFILE" \
+		|| _FAIL=1
+
+	# report errors
+	if [ "$_FAIL" -gt 0 ]; then
+		echo "FAILED: $f"
+		_FAILED="$(( $_FAILED + 1 ))"
+	else
+		echo "PASSED: $f"
+		_PASSED="$(( $_PASSED + 1))"
+	fi
+done
+
+echo Statistics:
+echo "$_FAILED files failed the compile test."
+echo "$_PASSED files passed the compile test."
+
+exit "$_FAILED"
-- 
2.1.4

[PATCH 05/45] drm.h: include stdlib.h in userspace

[Mikko Rapeli]

Fixes <drm/drm.h> compilation error:

drm/drm.h:132:2: error: unknown type name ‘size_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/drm.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/uapi/drm/drm.h b/include/uapi/drm/drm.h
index b0b8556..8f4488f 100644
--- a/include/uapi/drm/drm.h
+++ b/include/uapi/drm/drm.h
@@ -42,6 +42,10 @@
 #include <asm/ioctl.h>
 typedef unsigned int drm_handle_t;
 
+#ifndef __KERNEL__
+#include <stdlib.h>
+#endif
+
 #else /* One of the BSDs */
 
 #include <sys/ioccom.h>
-- 
2.1.4

Re: [PATCH 05/45] drm.h: include stdlib.h in userspace

[Emil Velikov]

On 16/02/15 23:05, Mikko Rapeli wrote:
> Fixes <drm/drm.h> compilation error:
> 
> drm/drm.h:132:2: error: unknown type name ‘size_t’
> 
Hi Mikko,

Can you let us know how you're getting these (series-wise) errors ? I've
been meaning to sync the uapi/drm and libdrm headers and would be nice
to have an extra step to test things.

Thanks
Emil

Re: [PATCH 05/45] drm.h: include stdlib.h in userspace

[Mikko Rapeli]

On Mon, Feb 23, 2015 at 10:26:58AM +0000, Emil Velikov wrote:
> On 16/02/15 23:05, Mikko Rapeli wrote:
> > Fixes <drm/drm.h> compilation error:
> > 
> > drm/drm.h:132:2: error: unknown type name ‘size_t’
> > 
> Hi Mikko,
> 
> Can you let us know how you're getting these (series-wise) errors ? I've
> been meaning to sync the uapi/drm and libdrm headers and would be nice
> to have an extra step to test things.

This should have everything needed to reproduce these compile errors,
though some of the errors hide behind other errors and fixes:

https://lkml.org/lkml/2015/2/16/525

-Mikko

Re: [PATCH 05/45] drm.h: include stdlib.h in userspace

[Emil Velikov]

On 23 February 2015 at 10:35, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> On Mon, Feb 23, 2015 at 10:26:58AM +0000, Emil Velikov wrote:
>> On 16/02/15 23:05, Mikko Rapeli wrote:
>> > Fixes <drm/drm.h> compilation error:
>> >
>> > drm/drm.h:132:2: error: unknown type name ‘size_t’
>> >
>> Hi Mikko,
>>
>> Can you let us know how you're getting these (series-wise) errors ? I've
>> been meaning to sync the uapi/drm and libdrm headers and would be nice
>> to have an extra step to test things.
>
> This should have everything needed to reproduce these compile errors,
> though some of the errors hide behind other errors and fixes:
>
> https://lkml.org/lkml/2015/2/16/525
>
Thanks for the link Mikko.

Afaict the general consensus seems to be that one should avoid using
stdint's uint8_t, but stick to __u8 and friends. Did you had the
chance to roll out another series that does so ?

That aside I'm not 100% sure that doing the UAPI split, as is, was the
perfect solution. Afaik drm used to live as an out of tree userspace
library(libdrm). Not sure at which point the major restructuring took
part, but one is certain - libdrm remains the only authoritative
sources of the headers. It's possible that some buggy programs pull
the UAPI headers while linking against the library, but I'd say that
won't end up well in the long term. Additionally since the UAPI split
the `make update-headers' target used to sync libdrm's headers have
been broken leading people to copy misc. hunks and/or files. Leading
to greater chance of things going sour.

All that said, I will need to gather some opinions for drm developers
and maintainers if the idea of part revering 718dcedd7e8(UAPI:
(Scripted) Disintegrate include/drm) will be the way forward.

Thanks
Emil

Re: [PATCH 05/45] drm.h: include stdlib.h in userspace

[Mikko Rapeli]

On Fri, Mar 20, 2015 at 08:25:40PM +0000, Emil Velikov wrote:
> On 23 February 2015 at 10:35, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> > On Mon, Feb 23, 2015 at 10:26:58AM +0000, Emil Velikov wrote:
> >> On 16/02/15 23:05, Mikko Rapeli wrote:
> >> > Fixes <drm/drm.h> compilation error:
> >> >
> >> > drm/drm.h:132:2: error: unknown type name ‘size_t’
> >> >
> >> Hi Mikko,
> >>
> >> Can you let us know how you're getting these (series-wise) errors ? I've
> >> been meaning to sync the uapi/drm and libdrm headers and would be nice
> >> to have an extra step to test things.
> >
> > This should have everything needed to reproduce these compile errors,
> > though some of the errors hide behind other errors and fixes:
> >
> > https://lkml.org/lkml/2015/2/16/525
> >
> Thanks for the link Mikko.
> 
> Afaict the general consensus seems to be that one should avoid using
> stdint's uint8_t, but stick to __u8 and friends. Did you had the
> chance to roll out another series that does so ?

Yes, new series with these changes is on the way. I'm trying to follow up to
all other review comments as well and get down to 100% compiling uapi
headers; 35 failures to go...

> That aside I'm not 100% sure that doing the UAPI split, as is, was the
> perfect solution. Afaik drm used to live as an out of tree userspace
> library(libdrm). Not sure at which point the major restructuring took
> part, but one is certain - libdrm remains the only authoritative
> sources of the headers. It's possible that some buggy programs pull
> the UAPI headers while linking against the library, but I'd say that
> won't end up well in the long term. Additionally since the UAPI split
> the `make update-headers' target used to sync libdrm's headers have
> been broken leading people to copy misc. hunks and/or files. Leading
> to greater chance of things going sour.
>
> All that said, I will need to gather some opinions for drm developers
> and maintainers if the idea of part revering 718dcedd7e8(UAPI:
> (Scripted) Disintegrate include/drm) will be the way forward.

Ok, I'll follow what is available in Linus' tree (or -next, not shure which
one I should track for these changes).

-Mikko

Re: [PATCH 05/45] drm.h: include stdlib.h in userspace

[Emil Velikov]

Hi Mikko

Pardon for the late response,

On 21 March 2015 at 12:17, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> On Fri, Mar 20, 2015 at 08:25:40PM +0000, Emil Velikov wrote:
>> On 23 February 2015 at 10:35, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
>> > On Mon, Feb 23, 2015 at 10:26:58AM +0000, Emil Velikov wrote:
>> >> On 16/02/15 23:05, Mikko Rapeli wrote:
>> >> > Fixes <drm/drm.h> compilation error:
>> >> >
>> >> > drm/drm.h:132:2: error: unknown type name ‘size_t’
>> >> >
>> >> Hi Mikko,
>> >>
>> >> Can you let us know how you're getting these (series-wise) errors ? I've
>> >> been meaning to sync the uapi/drm and libdrm headers and would be nice
>> >> to have an extra step to test things.
>> >
>> > This should have everything needed to reproduce these compile errors,
>> > though some of the errors hide behind other errors and fixes:
>> >
>> > https://lkml.org/lkml/2015/2/16/525
>> >
>> Thanks for the link Mikko.
>>
>> Afaict the general consensus seems to be that one should avoid using
>> stdint's uint8_t, but stick to __u8 and friends. Did you had the
>> chance to roll out another series that does so ?
>
> Yes, new series with these changes is on the way. I'm trying to follow up to
> all other review comments as well and get down to 100% compiling uapi
> headers; 35 failures to go...
>
Glad to hear that it's getting there. Might be a bit slower than
expected but we'll get there :-) Can you please Cc me on the next
iteration ?

>> That aside I'm not 100% sure that doing the UAPI split, as is, was the
>> perfect solution. Afaik drm used to live as an out of tree userspace
>> library(libdrm). Not sure at which point the major restructuring took
>> part, but one is certain - libdrm remains the only authoritative
>> sources of the headers. It's possible that some buggy programs pull
>> the UAPI headers while linking against the library, but I'd say that
>> won't end up well in the long term. Additionally since the UAPI split
>> the `make update-headers' target used to sync libdrm's headers have
>> been broken leading people to copy misc. hunks and/or files. Leading
>> to greater chance of things going sour.
>>
>> All that said, I will need to gather some opinions for drm developers
>> and maintainers if the idea of part revering 718dcedd7e8(UAPI:
>> (Scripted) Disintegrate include/drm) will be the way forward.
>
> Ok, I'll follow what is available in Linus' tree (or -next, not shure which
> one I should track for these changes).
>
After discussing (more like annoying) our DRM maintainer we got to the
conclusion that the headers will stay exposed to userspace. So any and
all of your work will be greatly appreciated.

Thanks
Emil

Re: [PATCH 05/45] drm.h: include stdlib.h in userspace

[Mikko Rapeli]

On Thu, Apr 09, 2015 at 05:00:48PM +0100, Emil Velikov wrote:
> Hi Mikko
> 
> Pardon for the late response,
> 
> On 21 March 2015 at 12:17, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> > On Fri, Mar 20, 2015 at 08:25:40PM +0000, Emil Velikov wrote:
> >> On 23 February 2015 at 10:35, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> >> > On Mon, Feb 23, 2015 at 10:26:58AM +0000, Emil Velikov wrote:
> >> >> On 16/02/15 23:05, Mikko Rapeli wrote:
> >> >> > Fixes <drm/drm.h> compilation error:
> >> >> >
> >> >> > drm/drm.h:132:2: error: unknown type name ‘size_t’
> >> >> >
> >> >> Hi Mikko,
> >> >>
> >> >> Can you let us know how you're getting these (series-wise) errors ? I've
> >> >> been meaning to sync the uapi/drm and libdrm headers and would be nice
> >> >> to have an extra step to test things.
> >> >
> >> > This should have everything needed to reproduce these compile errors,
> >> > though some of the errors hide behind other errors and fixes:
> >> >
> >> > https://lkml.org/lkml/2015/2/16/525
> >> >
> >> Thanks for the link Mikko.
> >>
> >> Afaict the general consensus seems to be that one should avoid using
> >> stdint's uint8_t, but stick to __u8 and friends. Did you had the
> >> chance to roll out another series that does so ?
> >
> > Yes, new series with these changes is on the way. I'm trying to follow up to
> > all other review comments as well and get down to 100% compiling uapi
> > headers; 35 failures to go...
> >
> Glad to hear that it's getting there. Might be a bit slower than
> expected but we'll get there :-) Can you please Cc me on the next
> iteration ?

Yes will Cc you too, and sorry life is getting on the way of this work. Draft
version is visible at

https://github.com/torvalds/linux/compare/master...mcfrisk:headers_test_v03

-Mikko

[PATCH 06/45] drm_mode.h: include stdint.h and linux/types.h in userspace

[Mikko Rapeli]

Fixes compilation error:

drm/drm_mode.h:472:2: error: unknown type name ‘uint32_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/drm_mode.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index a0db2d4a..c55dfdf 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -27,7 +27,12 @@
 #ifndef _DRM_MODE_H
 #define _DRM_MODE_H
 
+#ifdef __KERNEL__
 #include <linux/types.h>
+#else
+#include <stdint.h>
+#include <linux/types.h>
+#endif
 
 #define DRM_DISPLAY_INFO_LEN	32
 #define DRM_CONNECTOR_NAME_LEN	32
-- 
2.1.4

[PATCH 07/45] exynos_drm.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation errors like:

drm/exynos_drm.h:30:2: error: unknown type name ‘uint64_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/exynos_drm.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/uapi/drm/exynos_drm.h b/include/uapi/drm/exynos_drm.h
index 5575ed1..03d094d 100644
--- a/include/uapi/drm/exynos_drm.h
+++ b/include/uapi/drm/exynos_drm.h
@@ -15,6 +15,11 @@
 #ifndef _UAPI_EXYNOS_DRM_H_
 #define _UAPI_EXYNOS_DRM_H_
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 #include <drm/drm.h>
 
 /**
-- 
2.1.4

[PATCH 08/45] nouveau_drm.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

drm/nouveau_drm.h:41:2: error: unknown type name ‘uint32_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/nouveau_drm.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/drm/nouveau_drm.h b/include/uapi/drm/nouveau_drm.h
index 0d7608d..4baf959 100644
--- a/include/uapi/drm/nouveau_drm.h
+++ b/include/uapi/drm/nouveau_drm.h
@@ -35,6 +35,12 @@
 #define NOUVEAU_ABI16_DEVICE   0xdddddddd
 #define NOUVEAU_ABI16_CHAN(n) (0xcccc0000 | (n))
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 #define NOUVEAU_GEM_DOMAIN_CPU       (1 << 0)
 #define NOUVEAU_GEM_DOMAIN_VRAM      (1 << 1)
 #define NOUVEAU_GEM_DOMAIN_GART      (1 << 2)
-- 
2.1.4

[PATCH 09/45] radeon_drm.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compiler error:

drm/radeon_drm.h:794:2: error: unknown type name ‘uint64_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/radeon_drm.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h
index 50d0fb4..4d8ed03 100644
--- a/include/uapi/drm/radeon_drm.h
+++ b/include/uapi/drm/radeon_drm.h
@@ -33,6 +33,11 @@
 #ifndef __RADEON_DRM_H__
 #define __RADEON_DRM_H__
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 #include <drm/drm.h>
 
 /* WARNING: If you change any of these defines, make sure to change the
-- 
2.1.4

[PATCH 10/45] r128_drm.h: include drm/drm.h

[Mikko Rapeli]

Fixes compile error:

drm/r128_drm.h:156:23: error: array type has incomplete element type
  struct drm_clip_rect boxes[R128_NR_SAREA_CLIPRECTS];

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/r128_drm.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/drm/r128_drm.h b/include/uapi/drm/r128_drm.h
index 8d8878b..76b0aa3 100644
--- a/include/uapi/drm/r128_drm.h
+++ b/include/uapi/drm/r128_drm.h
@@ -33,6 +33,8 @@
 #ifndef __R128_DRM_H__
 #define __R128_DRM_H__
 
+#include <drm/drm.h>
+
 /* WARNING: If you change any of these defines, make sure to change the
  * defines in the X server file (r128_sarea.h)
  */
-- 
2.1.4

[PATCH 11/45] via_drm.h: include linux/types.h instead of non-existing via_drmclient.h

[Mikko Rapeli]

Fixes compiler error:

drm/via_drm.h:36:27: fatal error: via_drmclient.h: No such file or directory

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/via_drm.h | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/include/uapi/drm/via_drm.h b/include/uapi/drm/via_drm.h
index 8b0533c..791531e 100644
--- a/include/uapi/drm/via_drm.h
+++ b/include/uapi/drm/via_drm.h
@@ -24,6 +24,7 @@
 #ifndef _VIA_DRM_H_
 #define _VIA_DRM_H_
 
+#include <linux/types.h>
 #include <drm/drm.h>
 
 /* WARNING: These defines must be the same as what the Xserver uses.
@@ -33,9 +34,6 @@
 #ifndef _VIA_DEFINES_
 #define _VIA_DEFINES_
 
-#ifndef __KERNEL__
-#include "via_drmclient.h"
-#endif
 
 #define VIA_NR_SAREA_CLIPRECTS		8
 #define VIA_NR_XVMC_PORTS               10
-- 
2.1.4

[PATCH 12/45] via_drm.h: hide struct via_file_private in userspace

[Mikko Rapeli]

Fixes compiler error since list_head is not exported to userspace headers.

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/via_drm.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/drm/via_drm.h b/include/uapi/drm/via_drm.h
index 791531e..34ce658 100644
--- a/include/uapi/drm/via_drm.h
+++ b/include/uapi/drm/via_drm.h
@@ -272,8 +272,10 @@ typedef struct drm_via_dmablit {
 	drm_via_blitsync_t sync;
 } drm_via_dmablit_t;
 
+#ifdef __KERNEL__
 struct via_file_private {
 	struct list_head obj_list;
 };
+#endif
 
 #endif				/* _VIA_DRM_H_ */
-- 
2.1.4

[PATCH 13/45] savage_drm.h: include <drm/drm.h>

[Mikko Rapeli]

Fixes compiler error:

drm/savage_drm.h:50:24: error: array type has incomplete element type
  struct drm_tex_region texList[SAVAGE_NR_TEX_HEAPS][SAVAGE_NR_TEX_REGIONS +

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/savage_drm.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/drm/savage_drm.h b/include/uapi/drm/savage_drm.h
index 818d49b..9dc9dc1 100644
--- a/include/uapi/drm/savage_drm.h
+++ b/include/uapi/drm/savage_drm.h
@@ -26,6 +26,8 @@
 #ifndef __SAVAGE_DRM_H__
 #define __SAVAGE_DRM_H__
 
+#include <drm/drm.h>
+
 #ifndef __SAVAGE_SAREA_DEFINES__
 #define __SAVAGE_SAREA_DEFINES__
 
-- 
2.1.4

[PATCH 14/45] sis_drm.h: hide sis_file_private in userspace

[Mikko Rapeli]

Fixes compiler error:

drm/sis_drm.h:68:19: error: field ‘obj_list’ has incomplete type
  struct list_head obj_list;

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/sis_drm.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/drm/sis_drm.h b/include/uapi/drm/sis_drm.h
index df37632..3d0903f 100644
--- a/include/uapi/drm/sis_drm.h
+++ b/include/uapi/drm/sis_drm.h
@@ -64,8 +64,10 @@ typedef struct {
 	unsigned long offset, size;
 } drm_sis_fb_t;
 
+#ifdef __KERNEL__
 struct sis_file_private {
 	struct list_head obj_list;
 };
+#endif
 
 #endif				/* __SIS_DRM_H__ */
-- 
2.1.4

[PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

linux/dm-log-userspace.h:416:2: error: unknown type name ‘uint64_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/dm-log-userspace.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/uapi/linux/dm-log-userspace.h b/include/uapi/linux/dm-log-userspace.h
index 0fa0d9e..7b9684e 100644
--- a/include/uapi/linux/dm-log-userspace.h
+++ b/include/uapi/linux/dm-log-userspace.h
@@ -7,6 +7,11 @@
 #ifndef __DM_LOG_USERSPACE_H__
 #define __DM_LOG_USERSPACE_H__
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 #include <linux/dm-ioctl.h> /* For DM_UUID_LEN */
 
 /*
-- 
2.1.4

Re: [PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Mike Snitzer]

On Mon, Feb 16 2015 at  6:05pm -0500,
Mikko Rapeli <mikko.rapeli@iki.fi> wrote:

> Fixes compilation error:
> 
> linux/dm-log-userspace.h:416:2: error: unknown type name ‘uint64_t’

What userspace code are you compiling?  Do you have a feel for when this
stopped working?

Re: [PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Mikko Rapeli]

On Mon, Feb 16, 2015 at 06:32:54PM -0500, Mike Snitzer wrote:
> On Mon, Feb 16 2015 at  6:05pm -0500,
> Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> 
> > Fixes compilation error:
> > 
> > linux/dm-log-userspace.h:416:2: error: unknown type name ‘uint64_t’
> 
> What userspace code are you compiling?  Do you have a feel for when this
> stopped working?

See Message-Id: <1424127948-22484-1-git-send-email-mikko.rapeli@iki.fi>
or https://lkml.org/lkml/2015/2/16/521

The failure comes from a test which tries to compile each exported header file
in userspace one at a time.

-Mikko

Re: [PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Arnd Bergmann]

On Tuesday 17 February 2015 00:05:18 Mikko Rapeli wrote:
> --- a/include/uapi/linux/dm-log-userspace.h
> +++ b/include/uapi/linux/dm-log-userspace.h
> @@ -7,6 +7,11 @@
>  #ifndef __DM_LOG_USERSPACE_H__
>  #define __DM_LOG_USERSPACE_H__
>  
> +#ifdef __KERNEL__
> +#include <linux/types.h>
> +#else
> +#include <stdint.h>
> +#endif
>  #include <linux/dm-ioctl.h> /* For DM_UUID_LEN */
> 

The normal way to do this in kernel headers is to use linux/types.h
but change the data structures to use __u64 instead of uint64_t
to avoid the build error.

It's possible that the maintainers of this code have a strong opinion
on this matter, but try doing that first.

	Arnd

Re: [PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Alasdair G Kergon]

On Tue, Feb 17, 2015 at 10:08:56AM +0100, Arnd Bergmann wrote:
> The normal way to do this in kernel headers is to use linux/types.h
> but change the data structures to use __u64 instead of uint64_t
> to avoid the build error.
 
That's what happened to dm-ioctl.h.
(Or someone could adjust linux/types.h to include these as standard.)

Alasdair

Re: [PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Arnd Bergmann]

On Tuesday 17 February 2015 13:38:06 Alasdair G Kergon wrote:
> On Tue, Feb 17, 2015 at 10:08:56AM +0100, Arnd Bergmann wrote:
> > The normal way to do this in kernel headers is to use linux/types.h
> > but change the data structures to use __u64 instead of uint64_t
> > to avoid the build error.
>  
> That's what happened to dm-ioctl.h.

Ah, indeed. It turns out that it was my own change that did this:


commit 9adfbfb611307060db54691bc7e6d53fdc12312b
Author: Arnd Bergmann <arnd@arndb.de>
Date:   Thu Feb 26 00:51:40 2009 +0100

    make most exported headers use strict integer types
    
    This takes care of all files that have only a small number
    of non-strict integer type uses.

> (Or someone could adjust linux/types.h to include these as standard.)

No, that wouldn't work. The C user space headers are not meant to
be included implicitly by any standard headers, which might pull
in linux/types.h implicitly.

I think it would be best to change all patches in the new series
in the same way for consistency and try to avoid using stdint.h
as much as we can.

	Arnd

Re: [PATCH 15/45] dm-log-userspace.h: include stdint.h in userspace

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 02:55:49PM +0100, Arnd Bergmann wrote:
> On Tuesday 17 February 2015 13:38:06 Alasdair G Kergon wrote:
> > On Tue, Feb 17, 2015 at 10:08:56AM +0100, Arnd Bergmann wrote:
> > > The normal way to do this in kernel headers is to use linux/types.h
> > > but change the data structures to use __u64 instead of uint64_t
> > > to avoid the build error.
> >  
> > That's what happened to dm-ioctl.h.
> 
> Ah, indeed. It turns out that it was my own change that did this:
> 
> 
> commit 9adfbfb611307060db54691bc7e6d53fdc12312b
> Author: Arnd Bergmann <arnd@arndb.de>
> Date:   Thu Feb 26 00:51:40 2009 +0100
> 
>     make most exported headers use strict integer types
>     
>     This takes care of all files that have only a small number
>     of non-strict integer type uses.
> 
> > (Or someone could adjust linux/types.h to include these as standard.)
> 
> No, that wouldn't work. The C user space headers are not meant to
> be included implicitly by any standard headers, which might pull
> in linux/types.h implicitly.
> 
> I think it would be best to change all patches in the new series
> in the same way for consistency and try to avoid using stdint.h
> as much as we can.

Ok, I'll change all patches to use __u8 etc types instead of uint8_t etc via
stdint.h.

-Mikko

[PATCH 16/45] hsi_char.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compiler error:

linux/hsi/hsi_char.h:51:2: error: unknown type name ‘uint32_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/hsi/hsi_char.h | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/include/uapi/linux/hsi/hsi_char.h b/include/uapi/linux/hsi/hsi_char.h
index 76160b4..9fa8f91 100644
--- a/include/uapi/linux/hsi/hsi_char.h
+++ b/include/uapi/linux/hsi/hsi_char.h
@@ -20,10 +20,15 @@
  * 02110-1301 USA
  */
 
-
 #ifndef __HSI_CHAR_H
 #define __HSI_CHAR_H
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 #define HSI_CHAR_MAGIC		'k'
 #define HSC_IOW(num, dtype)	_IOW(HSI_CHAR_MAGIC, num, dtype)
 #define HSC_IOR(num, dtype)	_IOR(HSI_CHAR_MAGIC, num, dtype)
-- 
2.1.4

[PATCH 17/45] ebtables.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

linux/netfilter_bridge/ebtables.h:38:2: error: unknown type name ‘uint64_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/netfilter_bridge/ebtables.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/uapi/linux/netfilter_bridge/ebtables.h b/include/uapi/linux/netfilter_bridge/ebtables.h
index ba99336..9e8d6b4 100644
--- a/include/uapi/linux/netfilter_bridge/ebtables.h
+++ b/include/uapi/linux/netfilter_bridge/ebtables.h
@@ -12,6 +12,11 @@
 
 #ifndef _UAPI__LINUX_BRIDGE_EFF_H
 #define _UAPI__LINUX_BRIDGE_EFF_H
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 #include <linux/if.h>
 #include <linux/netfilter_bridge.h>
 #include <linux/if_ether.h>
-- 
2.1.4

[PATCH 18/45] cld.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

linux/nfsd/cld.h:40:2: error: unknown type name ‘uint16_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/nfsd/cld.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/linux/nfsd/cld.h b/include/uapi/linux/nfsd/cld.h
index f14a9ab..9f60f9c 100644
--- a/include/uapi/linux/nfsd/cld.h
+++ b/include/uapi/linux/nfsd/cld.h
@@ -22,6 +22,12 @@
 #ifndef _NFSD_CLD_H
 #define _NFSD_CLD_H
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 /* latest upcall version available */
 #define CLD_UPCALL_VERSION 1
 
-- 
2.1.4

[PATCH 19/45] rds.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

linux/rds.h:96:2: error: unknown type name ‘uint8_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/rds.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/uapi/linux/rds.h b/include/uapi/linux/rds.h
index 9195095..7ff0c70 100644
--- a/include/uapi/linux/rds.h
+++ b/include/uapi/linux/rds.h
@@ -34,7 +34,11 @@
 #ifndef _LINUX_RDS_H
 #define _LINUX_RDS_H
 
+#ifdef __KERNEL__
 #include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 
 #define RDS_IB_ABI_VERSION		0x301
 
-- 
2.1.4

[PATCH 20/45] sctp.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

linux/sctp.h:652:2: error: unknown type name ‘uint32_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/sctp.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
index ce70fe6..9fd31cf 100644
--- a/include/uapi/linux/sctp.h
+++ b/include/uapi/linux/sctp.h
@@ -53,7 +53,11 @@
 #ifndef _UAPI_SCTP_H
 #define _UAPI_SCTP_H
 
+#ifdef __KERNEL__
 #include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 #include <linux/socket.h>
 
 typedef __s32 sctp_assoc_t;
-- 
2.1.4

Re: [PATCH 20/45] sctp.h: include stdint.h in userspace

[Neil Horman]

On Tue, Feb 17, 2015 at 12:05:23AM +0100, Mikko Rapeli wrote:
> Fixes compilation error:
> 
> linux/sctp.h:652:2: error: unknown type name ‘uint32_t’
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
Acked-by: Neil Horman <nhorman@tuxdriver.com>

> ---
>  include/uapi/linux/sctp.h | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
> index ce70fe6..9fd31cf 100644
> --- a/include/uapi/linux/sctp.h
> +++ b/include/uapi/linux/sctp.h
> @@ -53,7 +53,11 @@
>  #ifndef _UAPI_SCTP_H
>  #define _UAPI_SCTP_H
>  
> +#ifdef __KERNEL__
>  #include <linux/types.h>
> +#else
> +#include <stdint.h>
> +#endif
>  #include <linux/socket.h>
>  
>  typedef __s32 sctp_assoc_t;
> -- 
> 2.1.4
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: [PATCH 20/45] sctp.h: include stdint.h in userspace

[Daniel Borkmann]

On 02/17/2015 12:05 AM, Mikko Rapeli wrote:
> Fixes compilation error:
>
> linux/sctp.h:652:2: error: unknown type name ‘uint32_t’
>
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>

I have no idea where the rest of this series went (neither netdev, nor
lkml, nor linux-sctp) and what user space application making use of the
header you are trying to fix !?

> ---
>   include/uapi/linux/sctp.h | 4 ++++
>   1 file changed, 4 insertions(+)
>
> diff --git a/include/uapi/linux/sctp.h b/include/uapi/linux/sctp.h
> index ce70fe6..9fd31cf 100644
> --- a/include/uapi/linux/sctp.h
> +++ b/include/uapi/linux/sctp.h
> @@ -53,7 +53,11 @@
>   #ifndef _UAPI_SCTP_H
>   #define _UAPI_SCTP_H
>
> +#ifdef __KERNEL__
>   #include <linux/types.h>
> +#else
> +#include <stdint.h>
> +#endif
>   #include <linux/socket.h>

Looks ugly, but could also be resolved in user space.

Re: [PATCH 20/45] sctp.h: include stdint.h in userspace

[Mikko Rapeli]

On Wed, Feb 18, 2015 at 04:48:04PM +0100, Daniel Borkmann wrote:
> On 02/17/2015 12:05 AM, Mikko Rapeli wrote:
> >Fixes compilation error:
> >
> >linux/sctp.h:652:2: error: unknown type name ‘uint32_t’
> >
> >Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> 
> I have no idea where the rest of this series went (neither netdev, nor
> lkml, nor linux-sctp) and what user space application making use of the
> header you are trying to fix !?

There is a test written to test compilation of exported kernel headers
one by one in userspace. lkml thread is here:

https://lkml.org/lkml/2015/2/16/521

The series should have gone to the right addresses with:

$ git send-email --to linux-kernel@vger.kernel.org \
	--cc-cmd "scripts/get_maintainer.pl --norolestats"

-Mikko

[PATCH 21/45] scsi_bsg_fc.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

scsi/scsi_bsg_fc.h:83:2: error: unknown type name ‘uint8_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/scsi/scsi_bsg_fc.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/scsi/scsi_bsg_fc.h b/include/uapi/scsi/scsi_bsg_fc.h
index 3031b90..5cc25df 100644
--- a/include/uapi/scsi/scsi_bsg_fc.h
+++ b/include/uapi/scsi/scsi_bsg_fc.h
@@ -22,6 +22,12 @@
 #ifndef SCSI_BSG_FC_H
 #define SCSI_BSG_FC_H
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 /*
  * This file intended to be included by both kernel and user space
  */
-- 
2.1.4

[PATCH 22/45] scsi_netlink.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

scsi/scsi_netlink.h:43:2: error: unknown type name ‘uint8_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/scsi/scsi_netlink.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/include/uapi/scsi/scsi_netlink.h b/include/uapi/scsi/scsi_netlink.h
index 62b4eda..9d14592 100644
--- a/include/uapi/scsi/scsi_netlink.h
+++ b/include/uapi/scsi/scsi_netlink.h
@@ -22,8 +22,12 @@
 #ifndef SCSI_NETLINK_H
 #define SCSI_NETLINK_H
 
-#include <linux/netlink.h>
+#ifdef __KERNEL__
 #include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+#include <linux/netlink.h>
 
 /*
  * This file intended to be included by both kernel and user space
-- 
2.1.4

[PATCH 23/45] scsi_netlink_fc.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

scsi/scsi_netlink_fc.h:60:2: error: expected specifier-qualifier-list before ‘uint64_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/scsi/scsi_netlink_fc.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/uapi/scsi/scsi_netlink_fc.h b/include/uapi/scsi/scsi_netlink_fc.h
index cbf76e4..d969337 100644
--- a/include/uapi/scsi/scsi_netlink_fc.h
+++ b/include/uapi/scsi/scsi_netlink_fc.h
@@ -21,6 +21,11 @@
 #ifndef SCSI_NETLINK_FC_H
 #define SCSI_NETLINK_FC_H
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
 #include <scsi/scsi_netlink.h>
 
 /*
-- 
2.1.4

[PATCH 24/45] hdspm.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

sound/hdspm.h:43:2: error: unknown type name ‘uint32_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/sound/hdspm.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/sound/hdspm.h b/include/uapi/sound/hdspm.h
index d956c35..f799828 100644
--- a/include/uapi/sound/hdspm.h
+++ b/include/uapi/sound/hdspm.h
@@ -20,6 +20,12 @@
  *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  */
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 /* Maximum channels is 64 even on 56Mode you have 64playbacks to matrix */
 #define HDSPM_MAX_CHANNELS      64
 
-- 
2.1.4

Re: [PATCH 24/45] hdspm.h: include stdint.h in userspace

[Takashi Iwai]

At Tue, 17 Feb 2015 00:05:27 +0100,
Mikko Rapeli wrote:
> 
> Fixes compilation error:
> 
> sound/hdspm.h:43:2: error: unknown type name ‘uint32_t’
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>

Applied for 3.21, thanks.


Takashi

> ---
>  include/uapi/sound/hdspm.h | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/include/uapi/sound/hdspm.h b/include/uapi/sound/hdspm.h
> index d956c35..f799828 100644
> --- a/include/uapi/sound/hdspm.h
> +++ b/include/uapi/sound/hdspm.h
> @@ -20,6 +20,12 @@
>   *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
>   */
>  
> +#ifdef __KERNEL__
> +#include <linux/types.h>
> +#else
> +#include <stdint.h>
> +#endif
> +
>  /* Maximum channels is 64 even on 56Mode you have 64playbacks to matrix */
>  #define HDSPM_MAX_CHANNELS      64
>  
> -- 
> 2.1.4
> 

Re: [PATCH 24/45] hdspm.h: include stdint.h in userspace

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 07:46:02AM +0100, Takashi Iwai wrote:
> At Tue, 17 Feb 2015 00:05:27 +0100,
> Mikko Rapeli wrote:
> > 
> > Fixes compilation error:
> > 
> > sound/hdspm.h:43:2: error: unknown type name ‘uint32_t’
> > 
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> 
> Applied for 3.21, thanks.

Sorry, but this should maybe be dropped or reverted from the sound tree.

Arnd and others pointed out that kernel headers should be using __u32 etc
types from linux/types.h instead of including stdint.h from libc and using
uint32_t et al.

I'll post a new patch version for hdspm.h too in the next revision.

-Mikko

> 
> Takashi
> 
> > ---
> >  include/uapi/sound/hdspm.h | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/include/uapi/sound/hdspm.h b/include/uapi/sound/hdspm.h
> > index d956c35..f799828 100644
> > --- a/include/uapi/sound/hdspm.h
> > +++ b/include/uapi/sound/hdspm.h
> > @@ -20,6 +20,12 @@
> >   *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
> >   */
> >  
> > +#ifdef __KERNEL__
> > +#include <linux/types.h>
> > +#else
> > +#include <stdint.h>
> > +#endif
> > +
> >  /* Maximum channels is 64 even on 56Mode you have 64playbacks to matrix */
> >  #define HDSPM_MAX_CHANNELS      64
> >  
> > -- 
> > 2.1.4
> > 

Re: [PATCH 24/45] hdspm.h: include stdint.h in userspace

[Takashi Iwai]

At Wed, 11 Mar 2015 02:28:12 +0200,
Mikko Rapeli wrote:
> 
> On Tue, Feb 17, 2015 at 07:46:02AM +0100, Takashi Iwai wrote:
> > At Tue, 17 Feb 2015 00:05:27 +0100,
> > Mikko Rapeli wrote:
> > > 
> > > Fixes compilation error:
> > > 
> > > sound/hdspm.h:43:2: error: unknown type name ‘uint32_t’
> > > 
> > > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > 
> > Applied for 3.21, thanks.
> 
> Sorry, but this should maybe be dropped or reverted from the sound tree.
> 
> Arnd and others pointed out that kernel headers should be using __u32 etc
> types from linux/types.h instead of including stdint.h from libc and using
> uint32_t et al.
> 
> I'll post a new patch version for hdspm.h too in the next revision.

Then please post an incremental patch to for-next branch of sound git
tree (or the current linux-next tree).


Takashi

[PATCH 25/45] gntalloc.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

xen/gntalloc.h:22:2: error: unknown type name ‘uint16_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/xen/gntalloc.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/xen/gntalloc.h b/include/uapi/xen/gntalloc.h
index 76bd580..184df7e 100644
--- a/include/uapi/xen/gntalloc.h
+++ b/include/uapi/xen/gntalloc.h
@@ -11,6 +11,12 @@
 #ifndef __LINUX_PUBLIC_GNTALLOC_H__
 #define __LINUX_PUBLIC_GNTALLOC_H__
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 /*
  * Allocates a new page and creates a new grant reference.
  */
-- 
2.1.4

Re: [Xen-devel] [PATCH 25/45] gntalloc.h: include stdint.h in userspace

[David Vrabel]

On 16/02/15 23:05, Mikko Rapeli wrote:
> Fixes compilation error:
> 
> xen/gntalloc.h:22:2: error: unknown type name ‘uint16_t’
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  include/uapi/xen/gntalloc.h | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/include/uapi/xen/gntalloc.h b/include/uapi/xen/gntalloc.h
> index 76bd580..184df7e 100644
> --- a/include/uapi/xen/gntalloc.h
> +++ b/include/uapi/xen/gntalloc.h
> @@ -11,6 +11,12 @@
>  #ifndef __LINUX_PUBLIC_GNTALLOC_H__
>  #define __LINUX_PUBLIC_GNTALLOC_H__
>  
> +#ifdef __KERNEL__
> +#include <linux/types.h>
> +#else
> +#include <stdint.h>
> +#endif

I think it would be preferrable to #include <linux/types.h> only and
switch to using the __u32 etc. types (as others have suggested).

David

Re: [Xen-devel] [PATCH 25/45] gntalloc.h: include stdint.h in userspace

[Mikko Rapeli]

On Wed, Feb 18, 2015 at 11:53:22AM +0000, David Vrabel wrote:
> On 16/02/15 23:05, Mikko Rapeli wrote:
> > Fixes compilation error:
> > 
> > xen/gntalloc.h:22:2: error: unknown type name ‘uint16_t’
> > 
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > ---
> >  include/uapi/xen/gntalloc.h | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/include/uapi/xen/gntalloc.h b/include/uapi/xen/gntalloc.h
> > index 76bd580..184df7e 100644
> > --- a/include/uapi/xen/gntalloc.h
> > +++ b/include/uapi/xen/gntalloc.h
> > @@ -11,6 +11,12 @@
> >  #ifndef __LINUX_PUBLIC_GNTALLOC_H__
> >  #define __LINUX_PUBLIC_GNTALLOC_H__
> >  
> > +#ifdef __KERNEL__
> > +#include <linux/types.h>
> > +#else
> > +#include <stdint.h>
> > +#endif
> 
> I think it would be preferrable to #include <linux/types.h> only and
> switch to using the __u32 etc. types (as others have suggested).

Yes, I'm changes the patches to these.

-Mikko

[PATCH 26/45] gntdev.h: include stdint.h in userspace

[Mikko Rapeli]

Fixes compilation error:

xen/gntdev.h:38:2: error: unknown type name ‘uint32_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/xen/gntdev.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/xen/gntdev.h b/include/uapi/xen/gntdev.h
index 5304bd3..f724f75 100644
--- a/include/uapi/xen/gntdev.h
+++ b/include/uapi/xen/gntdev.h
@@ -33,6 +33,12 @@
 #ifndef __LINUX_PUBLIC_GNTDEV_H__
 #define __LINUX_PUBLIC_GNTDEV_H__
 
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
 struct ioctl_gntdev_grant_ref {
 	/* The domain ID of the grant to be mapped. */
 	uint32_t domid;
-- 
2.1.4

[PATCH 27/45] include/uapi/linux/sysctl.h: include stdlib.h in userspace

[Mikko Rapeli]

Fixes userspace compilation error:

error: unknown type name ‘size_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/sysctl.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/uapi/linux/sysctl.h b/include/uapi/linux/sysctl.h
index 43aaba1..b7b344f 100644
--- a/include/uapi/linux/sysctl.h
+++ b/include/uapi/linux/sysctl.h
@@ -26,6 +26,10 @@
 #include <linux/types.h>
 #include <linux/compiler.h>
 
+#ifndef __KERNEL__
+#include <stdlib.h>
+#endif
+
 struct completion;
 
 #define CTL_MAXNAME 10		/* how many path components do we allow in a
-- 
2.1.4

[PATCH 28/45] include/uapi/asm-generic/ipcbuf.h: include linux/posix_types.h

[Mikko Rapeli]

Fixes userspace compilation error:

error: unknown type name ‘__kernel_key_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/ipcbuf.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/asm-generic/ipcbuf.h b/include/uapi/asm-generic/ipcbuf.h
index 3dbcc1e..909f825 100644
--- a/include/uapi/asm-generic/ipcbuf.h
+++ b/include/uapi/asm-generic/ipcbuf.h
@@ -1,6 +1,8 @@
 #ifndef __ASM_GENERIC_IPCBUF_H
 #define __ASM_GENERIC_IPCBUF_H
 
+#include <linux/posix_types.h>
+
 /*
  * The generic ipc64_perm structure:
  * Note extra padding because this structure is passed back and forth
-- 
2.1.4

[PATCH 29/45] include/uapi/asm-generic/msgbuf.h: include asm/ipcbuf.h

[Mikko Rapeli]

Fixes userspace compilation errors like:

error: field ‘msg_perm’ has incomplete type
struct ipc64_perm msg_perm

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/msgbuf.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/asm-generic/msgbuf.h b/include/uapi/asm-generic/msgbuf.h
index f55ecc4..f5fbd8e 100644
--- a/include/uapi/asm-generic/msgbuf.h
+++ b/include/uapi/asm-generic/msgbuf.h
@@ -2,6 +2,8 @@
 #define __ASM_GENERIC_MSGBUF_H
 
 #include <asm/bitsperlong.h>
+#include <asm/ipcbuf.h>
+
 /*
  * generic msqid64_ds structure.
  *
-- 
2.1.4

[PATCH 30/45] include/uapi/asm-generic/shmbuf.h: include fixes

[Mikko Rapeli]

Fixes userspace compilation errors like:

error: field ‘shm_perm’ has incomplete type
struct ipc64_perm shm_perm; /* operation perms */
error: unknown type name ‘size_t’
error: unknown type name ‘__kernel_time_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/shmbuf.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/asm-generic/shmbuf.h b/include/uapi/asm-generic/shmbuf.h
index 7e9fb2f..24ef152 100644
--- a/include/uapi/asm-generic/shmbuf.h
+++ b/include/uapi/asm-generic/shmbuf.h
@@ -1,7 +1,13 @@
 #ifndef __ASM_GENERIC_SHMBUF_H
 #define __ASM_GENERIC_SHMBUF_H
 
+#include <linux/types.h>
 #include <asm/bitsperlong.h>
+#include <asm/msgbuf.h>
+
+#ifndef __KERNEL__
+#include <stdlib.h>
+#endif
 
 /*
  * The shmid64_ds structure for x86 architecture.
-- 
2.1.4

[PATCH 31/45] asm-generic/sembuf.h: include asm/posix_types.h and asm/ipcbuf.h

[Mikko Rapeli]

Fixes userspace compilation errors like:

error: field ‘msg_perm’ has incomplete type
struct ipc64_perm msg_perm;
error: unknown type name ‘__kernel_ulong_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/sembuf.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/asm-generic/sembuf.h b/include/uapi/asm-generic/sembuf.h
index 4cb2c13..17d523f 100644
--- a/include/uapi/asm-generic/sembuf.h
+++ b/include/uapi/asm-generic/sembuf.h
@@ -2,6 +2,8 @@
 #define __ASM_GENERIC_SEMBUF_H
 
 #include <asm/bitsperlong.h>
+#include <asm/posix_types.h>
+#include <asm/ipcbuf.h>
 
 /*
  * The semid64_ds structure for x86 architecture.
-- 
2.1.4

[PATCH 32/45] drm/i810_drm.h: include drm/drm.h

[Mikko Rapeli]

Fixes userspace compilation error:

error: array type has incomplete element type
struct drm_clip_rect boxes[I810_NR_SAREA_CLIPRECTS];

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/drm/i810_drm.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/uapi/drm/i810_drm.h b/include/uapi/drm/i810_drm.h
index 7a10bb6..34736ef 100644
--- a/include/uapi/drm/i810_drm.h
+++ b/include/uapi/drm/i810_drm.h
@@ -1,6 +1,8 @@
 #ifndef _I810_DRM_H_
 #define _I810_DRM_H_
 
+#include <drm/drm.h>
+
 /* WARNING: These defines must be the same as what the Xserver uses.
  * if you change them, you must change the defines in the Xserver.
  */
-- 
2.1.4

[PATCH 33/45] include/uapi/asm-generic/signal.h: include stdlib.h in userspace

[Mikko Rapeli]

Fixes compiler warning:
error: unknown type name ‘size_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/signal.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/uapi/asm-generic/signal.h b/include/uapi/asm-generic/signal.h
index 9df61f1..8a341a2 100644
--- a/include/uapi/asm-generic/signal.h
+++ b/include/uapi/asm-generic/signal.h
@@ -3,6 +3,10 @@
 
 #include <linux/types.h>
 
+#ifndef __KERNEL__
+#include <stdlib.h>
+#endif
+
 #define _NSIG		64
 #define _NSIG_BPW	__BITS_PER_LONG
 #define _NSIG_WORDS	(_NSIG / _NSIG_BPW)
-- 
2.1.4

Re: [PATCH 33/45] include/uapi/asm-generic/signal.h: include stdlib.h in userspace

[Arnd Bergmann]

On Tuesday 17 February 2015 00:05:36 Mikko Rapeli wrote:
> Fixes compiler warning:
> error: unknown type name ‘size_t’
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  include/uapi/asm-generic/signal.h | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/include/uapi/asm-generic/signal.h b/include/uapi/asm-generic/signal.h
> index 9df61f1..8a341a2 100644
> --- a/include/uapi/asm-generic/signal.h
> +++ b/include/uapi/asm-generic/signal.h
> @@ -3,6 +3,10 @@
>  
>  #include <linux/types.h>
>  
> +#ifndef __KERNEL__
> +#include <stdlib.h>
> +#endif
> +

Hmm, maybe we should use __kernel_size_t instead?

	Arnd

Re: [PATCH 33/45] include/uapi/asm-generic/signal.h: include stdlib.h in userspace

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 10:44:48AM +0100, Arnd Bergmann wrote:
> On Tuesday 17 February 2015 00:05:36 Mikko Rapeli wrote:
> > Fixes compiler warning:
> > error: unknown type name ‘size_t’
> > 
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > ---
> >  include/uapi/asm-generic/signal.h | 4 ++++
> >  1 file changed, 4 insertions(+)
> > 
> > diff --git a/include/uapi/asm-generic/signal.h b/include/uapi/asm-generic/signal.h
> > index 9df61f1..8a341a2 100644
> > --- a/include/uapi/asm-generic/signal.h
> > +++ b/include/uapi/asm-generic/signal.h
> > @@ -3,6 +3,10 @@
> >  
> >  #include <linux/types.h>
> >  
> > +#ifndef __KERNEL__
> > +#include <stdlib.h>
> > +#endif
> > +
> 
> Hmm, maybe we should use __kernel_size_t instead?

This problem is in several uapi header files on x86:

asm-generic/shmbuf.h
asm-generic/signal.h
asm/signal.h
drm/drm.h
linux/agpgart.h
linux/kexec.h
linux/ncp_fs.h
linux/nfc.h
linux/omapfb.h
linux/reiserfs_xattr.h
linux/sysctl.h
sound/asound.h

so I guess using __kernel_size_t is the correct approach for all of them,
right?

linux/kexec.h seems suspicious though:

#ifndef __KERNEL__
/*
 * This structure is used to hold the arguments that are used when
 * loading  kernel binaries.
 */
struct kexec_segment {
        const void *buf;
        size_t bufsz;
        const void *mem;
        size_t memsz;
};

#endif /* __KERNEL__ */

-Mikko

Re: [PATCH 33/45] include/uapi/asm-generic/signal.h: include stdlib.h in userspace

[Arnd Bergmann]

On Wednesday 11 March 2015 00:58:28 Mikko Rapeli wrote:
> linux/kexec.h seems suspicious though:
> 
> #ifndef __KERNEL__
> /*
>  * This structure is used to hold the arguments that are used when
>  * loading  kernel binaries.
>  */
> struct kexec_segment {
>         const void *buf;
>         size_t bufsz;
>         const void *mem;
>         size_t memsz;
> };
> 
> #endif /* __KERNEL__ */
> 

Hmm, maybe that just needs to be removed as well, after we
have removed the function declaration that used to follow it
as part of this commit:

http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=9dc5c05f45ca8101025046cd

Using __kernel_size_t would work here as well, but I suspect that
the entire structure is useless as it is not used by kernel or
user space, and kexec_tools has its own copy of the structure.

	Arnd

[PATCH 34/45] include/uapi/asm-generic/signal.h: hide sigset_t definition in userspace

[Mikko Rapeli]

It is already defined by libc headers in
/usr/include/i386-linux-gnu/sys/select.h

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/signal.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/uapi/asm-generic/signal.h b/include/uapi/asm-generic/signal.h
index 8a341a2..af4d83a 100644
--- a/include/uapi/asm-generic/signal.h
+++ b/include/uapi/asm-generic/signal.h
@@ -88,9 +88,12 @@
 #define SIGSTKSZ	8192
 
 #ifndef __ASSEMBLY__
+#ifdef __KERNEL__
+/* already defined in userspace via stdlib.h */
 typedef struct {
 	unsigned long sig[_NSIG_WORDS];
 } sigset_t;
+#endif /* __KERNEL__ */
 
 /* not actually used, but required for linux/syscalls.h */
 typedef unsigned long old_sigset_t;
-- 
2.1.4

Re: [PATCH 34/45] include/uapi/asm-generic/signal.h: hide sigset_t definition in userspace

[Arnd Bergmann]

On Tuesday 17 February 2015 00:05:37 Mikko Rapeli wrote:
> +#ifdef __KERNEL__
> +/* already defined in userspace via stdlib.h */
>  typedef struct {
>         unsigned long sig[_NSIG_WORDS];
>  } sigset_t;
> +#endif /* __KERNEL__ */

I'm not sure here: Is this structure always identical to the one
we see in user space?

If not, we might want to rename the typedef to __kernel_sigset_t
instead and use that inside of the uapi headers.

	Arnd

Re: [PATCH 34/45] include/uapi/asm-generic/signal.h: hide sigset_t definition in userspace

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 10:16:09AM +0100, Arnd Bergmann wrote:
> On Tuesday 17 February 2015 00:05:37 Mikko Rapeli wrote:
> > +#ifdef __KERNEL__
> > +/* already defined in userspace via stdlib.h */
> >  typedef struct {
> >         unsigned long sig[_NSIG_WORDS];
> >  } sigset_t;
> > +#endif /* __KERNEL__ */
> 
> I'm not sure here: Is this structure always identical to the one
> we see in user space?
> 
> If not, we might want to rename the typedef to __kernel_sigset_t
> instead and use that inside of the uapi headers.

I was copying this ifdef from the struct sigaction part in the same file:

#ifndef __KERNEL__
struct sigaction {
        __sighandler_t sa_handler;
        unsigned long sa_flags;
#ifdef SA_RESTORER
        __sigrestore_t sa_restorer;
#endif
        sigset_t sa_mask;               /* mask last for extensibility */
};
#endif

I assumed it had the same problems at some point.

-Mikko

[PATCH 35/45] include/uapi/sound/asound.h: include stdlib.h in userspace

[Mikko Rapeli]

Fixes compiler errors like:
error: field ‘trigger_tstamp’ has incomplete type
error: invalid application of ‘sizeof’ to incomplete t
ype ‘struct timespec’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/sound/asound.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/uapi/sound/asound.h b/include/uapi/sound/asound.h
index 941d32f..af156b0 100644
--- a/include/uapi/sound/asound.h
+++ b/include/uapi/sound/asound.h
@@ -25,6 +25,9 @@
 
 #include <linux/types.h>
 
+#ifndef __KERNEL__
+#include <stdlib.h>
+#endif
 
 /*
  *  protocol version
-- 
2.1.4

Re: [PATCH 35/45] include/uapi/sound/asound.h: include stdlib.h in userspace

[Takashi Iwai]

At Tue, 17 Feb 2015 00:05:38 +0100,
Mikko Rapeli wrote:
> 
> Fixes compiler errors like:
> error: field ‘trigger_tstamp’ has incomplete type
> error: invalid application of ‘sizeof’ to incomplete t
> ype ‘struct timespec’
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>

Applied for 3.21, thanks.


Takashi

> ---
>  include/uapi/sound/asound.h | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/include/uapi/sound/asound.h b/include/uapi/sound/asound.h
> index 941d32f..af156b0 100644
> --- a/include/uapi/sound/asound.h
> +++ b/include/uapi/sound/asound.h
> @@ -25,6 +25,9 @@
>  
>  #include <linux/types.h>
>  
> +#ifndef __KERNEL__
> +#include <stdlib.h>
> +#endif
>  
>  /*
>   *  protocol version
> -- 
> 2.1.4
> 

[PATCH 36/45] include/uapi/linux/netfilter.h: include in.h and in6.h

[Mikko Rapeli]

Fixes userspace compile errors like:
error: field ‘in’ has incomplete type
error: field ‘in6’ has incomplete type

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/netfilter.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/include/uapi/linux/netfilter.h b/include/uapi/linux/netfilter.h
index ef1b1f8..750c04e 100644
--- a/include/uapi/linux/netfilter.h
+++ b/include/uapi/linux/netfilter.h
@@ -4,7 +4,8 @@
 #include <linux/types.h>
 #include <linux/compiler.h>
 #include <linux/sysctl.h>
-
+#include <linux/in.h>
+#include <linux/in6.h>
 
 /* Responses from hook functions. */
 #define NF_DROP 0
-- 
2.1.4

[PATCH 37/45] include/uapi/linux/socket.h: include sys/socket.h in userspace

[Mikko Rapeli]

This libc header has sockaddr definition for userspace.
Fixes compilation errors like:
error: field ‘ifru_addr’ has incomplete type
struct sockaddr ifru_addr;

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/socket.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/uapi/linux/socket.h b/include/uapi/linux/socket.h
index 76ab0c6..8a81197 100644
--- a/include/uapi/linux/socket.h
+++ b/include/uapi/linux/socket.h
@@ -1,6 +1,10 @@
 #ifndef _UAPI_LINUX_SOCKET_H
 #define _UAPI_LINUX_SOCKET_H
 
+#ifndef __KERNEL__
+#include <sys/socket.h>
+#endif
+
 /*
  * Desired design of maximum size and alignment (see RFC2553)
  */
-- 
2.1.4

Re: [PATCH 37/45] include/uapi/linux/socket.h: include sys/socket.h in userspace

[Arnd Bergmann]

On Tuesday 17 February 2015 00:05:40 Mikko Rapeli wrote:
> This libc header has sockaddr definition for userspace.
> Fixes compilation errors like:
> error: field ‘ifru_addr’ has incomplete type
> struct sockaddr ifru_addr;
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  include/uapi/linux/socket.h | 4 ++++
>  1 file changed, 4 insertions(+)

This looks wrong: where do you see a use of ifru_addr
in uapi/linux/socket.h?

	Arnd

Re: [PATCH 37/45] include/uapi/linux/socket.h: include sys/socket.h in userspace

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 10:14:10AM +0100, Arnd Bergmann wrote:
> On Tuesday 17 February 2015 00:05:40 Mikko Rapeli wrote:
> > This libc header has sockaddr definition for userspace.
> > Fixes compilation errors like:
> > error: field ‘ifru_addr’ has incomplete type
> > struct sockaddr ifru_addr;
> > 
> > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > ---
> >  include/uapi/linux/socket.h | 4 ++++
> >  1 file changed, 4 insertions(+)
> 
> This looks wrong: where do you see a use of ifru_addr
> in uapi/linux/socket.h?

Indeed, this one seems completely wrong. Based on the test logs the
change should have gone to linux/if.h.

-Mikko

Re: [PATCH 37/45] include/uapi/linux/socket.h: include sys/socket.h in userspace

[Mikko Rapeli]

On Wed, Mar 11, 2015 at 02:38:15AM +0200, Mikko Rapeli wrote:
> On Tue, Feb 17, 2015 at 10:14:10AM +0100, Arnd Bergmann wrote:
> > On Tuesday 17 February 2015 00:05:40 Mikko Rapeli wrote:
> > > This libc header has sockaddr definition for userspace.
> > > Fixes compilation errors like:
> > > error: field ‘ifru_addr’ has incomplete type
> > > struct sockaddr ifru_addr;
> > > 
> > > Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> > > ---
> > >  include/uapi/linux/socket.h | 4 ++++
> > >  1 file changed, 4 insertions(+)
> > 
> > This looks wrong: where do you see a use of ifru_addr
> > in uapi/linux/socket.h?
> 
> Indeed, this one seems completely wrong. Based on the test logs the
> change should have gone to linux/if.h.

Hmm, turns out this was not so straight forward. If I drop this I get 20
more errors from headers which include linux/socket.h and expect sockaddr
to be defined which is not true since it comes from sys/socket.h.

One way to fix this for all linux/socket.h users is to include sys/socket.h
already in linux/socket.h. After this I get rid of all compile test errors from
missing sockaddr definitions except cases where headers fail to include even
linux/socket.h.

Other option would be to include sys/socket.h in all headers using sockaddr.

I would prefer to include sys/socket.h already in linux/socket.h but I will
improve the commit message with this detail.

-Mikko

[PATCH 38/45] include/uapi/linux/agpgart.h: include stdlib.h in userspace

[Mikko Rapeli]

Fixes userspace compiler error:
error: unknown type name ‘size_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/agpgart.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/include/uapi/linux/agpgart.h b/include/uapi/linux/agpgart.h
index 4e828cf..f525104 100644
--- a/include/uapi/linux/agpgart.h
+++ b/include/uapi/linux/agpgart.h
@@ -52,6 +52,7 @@
 
 #ifndef __KERNEL__
 #include <linux/types.h>
+#include <stdlib.h>
 
 struct agp_version {
 	__u16 major;
-- 
2.1.4

[PATCH 39/45] include/uapi/sound/asequencer.h: include sound/asound.h

[Mikko Rapeli]

Fixes userspace compilation error:
error: unknown type name ‘snd_seq_client_type_t’
snd_seq_client_type_t type; /* client type */

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/sound/asequencer.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/include/uapi/sound/asequencer.h b/include/uapi/sound/asequencer.h
index 09c8a00..5a5fa49 100644
--- a/include/uapi/sound/asequencer.h
+++ b/include/uapi/sound/asequencer.h
@@ -22,6 +22,7 @@
 #ifndef _UAPI__SOUND_ASEQUENCER_H
 #define _UAPI__SOUND_ASEQUENCER_H
 
+#include <sound/asound.h>
 
 /** version of the sequencer */
 #define SNDRV_SEQ_VERSION SNDRV_PROTOCOL_VERSION (1, 0, 1)
-- 
2.1.4

Re: [PATCH 39/45] include/uapi/sound/asequencer.h: include sound/asound.h

[Takashi Iwai]

At Tue, 17 Feb 2015 00:05:42 +0100,
Mikko Rapeli wrote:
> 
> Fixes userspace compilation error:
> error: unknown type name ‘snd_seq_client_type_t’
> snd_seq_client_type_t type; /* client type */
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>

Applied for 3.21, thanks.


Takashi

> ---
>  include/uapi/sound/asequencer.h | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/include/uapi/sound/asequencer.h b/include/uapi/sound/asequencer.h
> index 09c8a00..5a5fa49 100644
> --- a/include/uapi/sound/asequencer.h
> +++ b/include/uapi/sound/asequencer.h
> @@ -22,6 +22,7 @@
>  #ifndef _UAPI__SOUND_ASEQUENCER_H
>  #define _UAPI__SOUND_ASEQUENCER_H
>  
> +#include <sound/asound.h>
>  
>  /** version of the sequencer */
>  #define SNDRV_SEQ_VERSION SNDRV_PROTOCOL_VERSION (1, 0, 1)
> -- 
> 2.1.4
> 

[PATCH 40/45] include/uapi/sound/emu10k1.h: include sound/asound.h

[Mikko Rapeli]

Fixes userspace compilation errors like:
error: field ‘id’ has incomplete type
struct snd_ctl_elem_id id;  /* full control ID definition */

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/sound/emu10k1.h | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/include/uapi/sound/emu10k1.h b/include/uapi/sound/emu10k1.h
index d1bbaf7..ec1535b 100644
--- a/include/uapi/sound/emu10k1.h
+++ b/include/uapi/sound/emu10k1.h
@@ -23,8 +23,7 @@
 #define _UAPI__SOUND_EMU10K1_H
 
 #include <linux/types.h>
-
-
+#include <sound/asound.h>
 
 /*
  * ---- FX8010 ----
-- 
2.1.4

Re: [PATCH 40/45] include/uapi/sound/emu10k1.h: include sound/asound.h

[Takashi Iwai]

At Tue, 17 Feb 2015 00:05:43 +0100,
Mikko Rapeli wrote:
> 
> Fixes userspace compilation errors like:
> error: field ‘id’ has incomplete type
> struct snd_ctl_elem_id id;  /* full control ID definition */
> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>

Applied for 3.21, thanks.


Takashi

> ---
>  include/uapi/sound/emu10k1.h | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/include/uapi/sound/emu10k1.h b/include/uapi/sound/emu10k1.h
> index d1bbaf7..ec1535b 100644
> --- a/include/uapi/sound/emu10k1.h
> +++ b/include/uapi/sound/emu10k1.h
> @@ -23,8 +23,7 @@
>  #define _UAPI__SOUND_EMU10K1_H
>  
>  #include <linux/types.h>
> -
> -
> +#include <sound/asound.h>
>  
>  /*
>   * ---- FX8010 ----
> -- 
> 2.1.4
> 

[PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Mikko Rapeli]

The DECLARE_BITMAP macro is not available in userspace headers.
Fixes userspace compile error:
error: expected specifier-qualifier-list before ‘DECLARE_BITMAP’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/sound/emu10k1.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/uapi/sound/emu10k1.h b/include/uapi/sound/emu10k1.h
index ec1535b..f2fd870 100644
--- a/include/uapi/sound/emu10k1.h
+++ b/include/uapi/sound/emu10k1.h
@@ -300,7 +300,9 @@ struct snd_emu10k1_fx8010_control_old_gpr {
 struct snd_emu10k1_fx8010_code {
 	char name[128];
 
+#ifdef __KERNEL__
 	DECLARE_BITMAP(gpr_valid, 0x200); /* bitmask of valid initializers */
+#endif
 	__u32 __user *gpr_map;		/* initializers */
 
 	unsigned int gpr_add_control_count; /* count of GPR controls to add/replace */
@@ -313,11 +315,15 @@ struct snd_emu10k1_fx8010_code {
 	unsigned int gpr_list_control_total; /* total count of GPR controls */
 	struct snd_emu10k1_fx8010_control_gpr __user *gpr_list_controls; /* listed GPR controls */
 
+#ifdef __KERNEL__
 	DECLARE_BITMAP(tram_valid, 0x100); /* bitmask of valid initializers */
+#endif
 	__u32 __user *tram_data_map;	  /* data initializers */
 	__u32 __user *tram_addr_map;	  /* map initializers */
 
+#ifdef __KERNEL__
 	DECLARE_BITMAP(code_valid, 1024); /* bitmask of valid instructions */
+#endif
 	__u32 __user *code;		  /* one instruction - 64 bits */
 };
 
-- 
2.1.4

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Takashi Iwai]

At Tue, 17 Feb 2015 00:05:44 +0100,
Mikko Rapeli wrote:
> 
> The DECLARE_BITMAP macro is not available in userspace headers.
> Fixes userspace compile error:
> error: expected specifier-qualifier-list before ‘DECLARE_BITMAP’

It's nonsense.  This results in an incompatible structure, thus ABI
would be broken completely (actually this will break the compile of
ld10k1).


Takashi

> 
> Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
> ---
>  include/uapi/sound/emu10k1.h | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/include/uapi/sound/emu10k1.h b/include/uapi/sound/emu10k1.h
> index ec1535b..f2fd870 100644
> --- a/include/uapi/sound/emu10k1.h
> +++ b/include/uapi/sound/emu10k1.h
> @@ -300,7 +300,9 @@ struct snd_emu10k1_fx8010_control_old_gpr {
>  struct snd_emu10k1_fx8010_code {
>  	char name[128];
>  
> +#ifdef __KERNEL__
>  	DECLARE_BITMAP(gpr_valid, 0x200); /* bitmask of valid initializers */
> +#endif
>  	__u32 __user *gpr_map;		/* initializers */
>  
>  	unsigned int gpr_add_control_count; /* count of GPR controls to add/replace */
> @@ -313,11 +315,15 @@ struct snd_emu10k1_fx8010_code {
>  	unsigned int gpr_list_control_total; /* total count of GPR controls */
>  	struct snd_emu10k1_fx8010_control_gpr __user *gpr_list_controls; /* listed GPR controls */
>  
> +#ifdef __KERNEL__
>  	DECLARE_BITMAP(tram_valid, 0x100); /* bitmask of valid initializers */
> +#endif
>  	__u32 __user *tram_data_map;	  /* data initializers */
>  	__u32 __user *tram_addr_map;	  /* map initializers */
>  
> +#ifdef __KERNEL__
>  	DECLARE_BITMAP(code_valid, 1024); /* bitmask of valid instructions */
> +#endif
>  	__u32 __user *code;		  /* one instruction - 64 bits */
>  };
>  
> -- 
> 2.1.4
> 

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 07:27:38AM +0100, Takashi Iwai wrote:
> At Tue, 17 Feb 2015 00:05:44 +0100,
> Mikko Rapeli wrote:
> > 
> > The DECLARE_BITMAP macro is not available in userspace headers.
> > Fixes userspace compile error:
> > error: expected specifier-qualifier-list before ‘DECLARE_BITMAP’
> 
> It's nonsense.  This results in an incompatible structure, thus ABI
> would be broken completely (actually this will break the compile of
> ld10k1).

None of the exported headers after 'make headers_install' have definition
of DECLARE_BITMAP macro. It is defined in include/linux/types.h which is
different from include/uapi/linux/types.h and missing this definition and
a few other things.

One option would be add DECLARE_BITMAP macro to include/uapi/linux/types.h
and add include/linux/bitops.h to uapi.

Thoughts?

-Mikko

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Takashi Iwai]

At Wed, 11 Mar 2015 03:22:04 +0200,
Mikko Rapeli wrote:
> 
> On Tue, Feb 17, 2015 at 07:27:38AM +0100, Takashi Iwai wrote:
> > At Tue, 17 Feb 2015 00:05:44 +0100,
> > Mikko Rapeli wrote:
> > > 
> > > The DECLARE_BITMAP macro is not available in userspace headers.
> > > Fixes userspace compile error:
> > > error: expected specifier-qualifier-list before ‘DECLARE_BITMAP’
> > 
> > It's nonsense.  This results in an incompatible structure, thus ABI
> > would be broken completely (actually this will break the compile of
> > ld10k1).
> 
> None of the exported headers after 'make headers_install' have definition
> of DECLARE_BITMAP macro. It is defined in include/linux/types.h which is
> different from include/uapi/linux/types.h and missing this definition and
> a few other things.
> 
> One option would be add DECLARE_BITMAP macro to include/uapi/linux/types.h
> and add include/linux/bitops.h to uapi.
> 
> Thoughts?

Are there any other headers like that?  If this is the only one, leave
it as is.  The only program that reads this are some alsa-tools ones
and they have already own DECLARE_BITMAP() definition.  Adding the
extra definition here will even break the compilation out of sudden.


Takashi

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Arnd Bergmann]

On Wednesday 11 March 2015 07:11:18 Takashi Iwai wrote:
> At Wed, 11 Mar 2015 03:22:04 +0200,
> Mikko Rapeli wrote:
> > 
> > On Tue, Feb 17, 2015 at 07:27:38AM +0100, Takashi Iwai wrote:
> > > At Tue, 17 Feb 2015 00:05:44 +0100,
> > > Mikko Rapeli wrote:
> > > > 
> > > > The DECLARE_BITMAP macro is not available in userspace headers.
> > > > Fixes userspace compile error:
> > > > error: expected specifier-qualifier-list before ‘DECLARE_BITMAP’
> > > 
> > > It's nonsense.  This results in an incompatible structure, thus ABI
> > > would be broken completely (actually this will break the compile of
> > > ld10k1).
> > 
> > None of the exported headers after 'make headers_install' have definition
> > of DECLARE_BITMAP macro. It is defined in include/linux/types.h which is
> > different from include/uapi/linux/types.h and missing this definition and
> > a few other things.
> > 
> > One option would be add DECLARE_BITMAP macro to include/uapi/linux/types.h
> > and add include/linux/bitops.h to uapi.
> > 
> > Thoughts?
> 
> Are there any other headers like that?  If this is the only one, leave
> it as is.  The only program that reads this are some alsa-tools ones
> and they have already own DECLARE_BITMAP() definition.  Adding the
> extra definition here will even break the compilation out of sudden.

I think it's a worthy goal to have the header files be compilable
standalone, but I don't think we should make the DECLARE_BITMAP()
macro globally visible in user space, in particular because it will
clash with every instance in which user space has a macro of the
same name.

What we could do here is to add a private copy of the macro to emu10k1.h
under a different name, such as __EMU10K1_DECLARE_BITMAP().

	Arnd

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Takashi Iwai]

At Wed, 11 Mar 2015 10:46:29 +0100,
Arnd Bergmann wrote:
> 
> On Wednesday 11 March 2015 07:11:18 Takashi Iwai wrote:
> > At Wed, 11 Mar 2015 03:22:04 +0200,
> > Mikko Rapeli wrote:
> > > 
> > > On Tue, Feb 17, 2015 at 07:27:38AM +0100, Takashi Iwai wrote:
> > > > At Tue, 17 Feb 2015 00:05:44 +0100,
> > > > Mikko Rapeli wrote:
> > > > > 
> > > > > The DECLARE_BITMAP macro is not available in userspace headers.
> > > > > Fixes userspace compile error:
> > > > > error: expected specifier-qualifier-list before ‘DECLARE_BITMAP’
> > > > 
> > > > It's nonsense.  This results in an incompatible structure, thus ABI
> > > > would be broken completely (actually this will break the compile of
> > > > ld10k1).
> > > 
> > > None of the exported headers after 'make headers_install' have definition
> > > of DECLARE_BITMAP macro. It is defined in include/linux/types.h which is
> > > different from include/uapi/linux/types.h and missing this definition and
> > > a few other things.
> > > 
> > > One option would be add DECLARE_BITMAP macro to include/uapi/linux/types.h
> > > and add include/linux/bitops.h to uapi.
> > > 
> > > Thoughts?
> > 
> > Are there any other headers like that?  If this is the only one, leave
> > it as is.  The only program that reads this are some alsa-tools ones
> > and they have already own DECLARE_BITMAP() definition.  Adding the
> > extra definition here will even break the compilation out of sudden.
> 
> I think it's a worthy goal to have the header files be compilable
> standalone,

In general yes, but this case is very minor issue:
- the file in question is for a hardware device-specific data
  definition,
- there are only two programs read this file, both can be built
  properly,
- and the device and the programs are very old, modifying such need
  extra care.

> but I don't think we should make the DECLARE_BITMAP()
> macro globally visible in user space, in particular because it will
> clash with every instance in which user space has a macro of the
> same name.
> 
> What we could do here is to add a private copy of the macro to emu10k1.h
> under a different name, such as __EMU10K1_DECLARE_BITMAP().

Yes, it's a better option.


thanks,

Takashi

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Arnd Bergmann]

On Thursday 12 March 2015 07:11:48 Takashi Iwai wrote:
> At Wed, 11 Mar 2015 10:46:29 +0100,
> Arnd Bergmann wrote:
> > 
> > On Wednesday 11 March 2015 07:11:18 Takashi Iwai wrote:
> > > At Wed, 11 Mar 2015 03:22:04 +0200,
> > > 
> > > Are there any other headers like that?  If this is the only one, leave
> > > it as is.  The only program that reads this are some alsa-tools ones
> > > and they have already own DECLARE_BITMAP() definition.  Adding the
> > > extra definition here will even break the compilation out of sudden.
> > 
> > I think it's a worthy goal to have the header files be compilable
> > standalone,
> 
> In general yes, but this case is very minor issue:
> - the file in question is for a hardware device-specific data
>   definition,
> - there are only two programs read this file, both can be built
>   properly,
> - and the device and the programs are very old, modifying such need
>   extra care.

Right, we should only do it if the goal is to have all uapi headers
includable standalone. For a particular header file there is very
little benefit as you say, but it would be useful if we can automatically
test for regressions with new or modified headers.

	Arnd

Re: [PATCH 41/45] include/uapi/sound/emu10k1.h: hide gpr_valid, tram_valid and code_valid in userspace

[Takashi Iwai]

At Thu, 12 Mar 2015 09:45:42 +0100,
Arnd Bergmann wrote:
> 
> On Thursday 12 March 2015 07:11:48 Takashi Iwai wrote:
> > At Wed, 11 Mar 2015 10:46:29 +0100,
> > Arnd Bergmann wrote:
> > > 
> > > On Wednesday 11 March 2015 07:11:18 Takashi Iwai wrote:
> > > > At Wed, 11 Mar 2015 03:22:04 +0200,
> > > > 
> > > > Are there any other headers like that?  If this is the only one, leave
> > > > it as is.  The only program that reads this are some alsa-tools ones
> > > > and they have already own DECLARE_BITMAP() definition.  Adding the
> > > > extra definition here will even break the compilation out of sudden.
> > > 
> > > I think it's a worthy goal to have the header files be compilable
> > > standalone,
> > 
> > In general yes, but this case is very minor issue:
> > - the file in question is for a hardware device-specific data
> >   definition,
> > - there are only two programs read this file, both can be built
> >   properly,
> > - and the device and the programs are very old, modifying such need
> >   extra care.
> 
> Right, we should only do it if the goal is to have all uapi headers
> includable standalone. For a particular header file there is very
> little benefit as you say, but it would be useful if we can automatically
> test for regressions with new or modified headers.

True.  Yet another option would be just move this file back from
include/uapi/sound to include/sound.  Basically no user-space programs
care about this file, as they have already a copy of old header fils
in the package.

But maybe defining __EMU10K1_DECLARE_BITMAP() would be the easiest
solution, I suppose.


thanks,

Takashi

[PATCH 42/45] include/uapi/linux/rds.h: include linux/socket.h and linux/types.h always

[Mikko Rapeli]

Fixes compilation errors in userspace like:

error: unknown type name ‘__be32’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/rds.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/uapi/linux/rds.h b/include/uapi/linux/rds.h
index 7ff0c70..47b1ede 100644
--- a/include/uapi/linux/rds.h
+++ b/include/uapi/linux/rds.h
@@ -34,9 +34,9 @@
 #ifndef _LINUX_RDS_H
 #define _LINUX_RDS_H
 
-#ifdef __KERNEL__
 #include <linux/types.h>
-#else
+#include <linux/socket.h>
+#ifndef __KERNEL__
 #include <stdint.h>
 #endif
 
-- 
2.1.4

[PATCH 43/45] include/uapi/linux/netfilter_bridge.h: include if.h

[Mikko Rapeli]

Fixes userspace compilation errors like:

error: field ‘in’ has incomplete type
struct in_addr in;

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/netfilter_bridge.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/include/uapi/linux/netfilter_bridge.h b/include/uapi/linux/netfilter_bridge.h
index a5eda6d..5bb0528 100644
--- a/include/uapi/linux/netfilter_bridge.h
+++ b/include/uapi/linux/netfilter_bridge.h
@@ -4,6 +4,7 @@
 /* bridge-specific defines for netfilter. 
  */
 
+#include <linux/if.h>
 #include <linux/netfilter.h>
 #include <linux/if_ether.h>
 #include <linux/if_vlan.h>
-- 
2.1.4

Re: [PATCH 43/45] include/uapi/linux/netfilter_bridge.h: include if.h

[Jan Engelhardt]

On Tuesday 2015-02-17 00:05, Mikko Rapeli wrote:

>Fixes userspace compilation errors like:
>
>error: field ‘in’ has incomplete type
>struct in_addr in;
> 
>+#include <linux/if.h>

Patch 36/45 included linux/in.h instead of linux/if.h for addressing "in has incomplete
type". Should this be used here too?

Re: [PATCH 43/45] include/uapi/linux/netfilter_bridge.h: include if.h

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 01:02:50AM +0100, Jan Engelhardt wrote:
> On Tuesday 2015-02-17 00:05, Mikko Rapeli wrote:
> 
> >Fixes userspace compilation errors like:
> >
> >error: field ‘in’ has incomplete type
> >struct in_addr in;
> > 
> >+#include <linux/if.h>
> 
> Patch 36/45 included linux/in.h instead of linux/if.h for addressing "in has incomplete
> type". Should this be used here too?

It could, yes. It seems I cut some corners since this was hiding behind it:

In file included from ./linux/netfilter_bridge.h:11:0,
                 from ./linux/netfilter_bridge.c:1:
./linux/if_pppox.h:42:11: error: ‘IFNAMSIZ’ undeclared here (not in a function)
  char          dev[IFNAMSIZ];          /* Local device to use */

I'll fix these.

-Mikko

[PATCH 44/45] nf_conntrack_tuple_common.h: include linux/types.h and linux/netfilter.h

[Mikko Rapeli]

Fixes userspace compilation errors like:
error: unknown type name ‘__be16’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/linux/netfilter/nf_conntrack_tuple_common.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h b/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h
index 2f6bbc5..a9c3834 100644
--- a/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h
+++ b/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h
@@ -1,6 +1,9 @@
 #ifndef _NF_CONNTRACK_TUPLE_COMMON_H
 #define _NF_CONNTRACK_TUPLE_COMMON_H
 
+#include <linux/types.h>
+#include <linux/netfilter.h>
+
 enum ip_conntrack_dir {
 	IP_CT_DIR_ORIGINAL,
 	IP_CT_DIR_REPLY,
-- 
2.1.4

[PATCH 45/45] include/uapi/asm-generic/ucontext.h: include signal.h and sigcontext.h

[Mikko Rapeli]

Fixes userspace compiler errors:

error: unknown type name ‘stack_t’
error: field ‘uc_mcontext’ has incomplete type
struct sigcontext uc_mcontext;
error: unknown type name ‘sigset_t’

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
---
 include/uapi/asm-generic/ucontext.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/uapi/asm-generic/ucontext.h b/include/uapi/asm-generic/ucontext.h
index ad77343..31ece341 100644
--- a/include/uapi/asm-generic/ucontext.h
+++ b/include/uapi/asm-generic/ucontext.h
@@ -1,6 +1,9 @@
 #ifndef __ASM_GENERIC_UCONTEXT_H
 #define __ASM_GENERIC_UCONTEXT_H
 
+#include <asm-generic/signal.h>
+#include <asm/sigcontext.h>
+
 struct ucontext {
 	unsigned long	  uc_flags;
 	struct ucontext  *uc_link;
-- 
2.1.4

Re: [PATCH 45/45] include/uapi/asm-generic/ucontext.h: include signal.h and sigcontext.h

[Arnd Bergmann]

On Tuesday 17 February 2015 00:05:48 Mikko Rapeli wrote:
>  #ifndef __ASM_GENERIC_UCONTEXT_H
>  #define __ASM_GENERIC_UCONTEXT_H
>  
> +#include <asm-generic/signal.h>
> +#include <asm/sigcontext.h>
> +
>  struct ucontext {
> 

Including another asm-generic header here is a bad idea: it breaks
if an architecture overrides asm/signal.h with its own version
but wants to use the asm-generic/ucontext.h file.

It would be best to just use linux/signal.h here, which includes
the correct architecture specific files.

	Arnd

Re: [PATCH 45/45] include/uapi/asm-generic/ucontext.h: include signal.h and sigcontext.h

[Mikko Rapeli]

On Tue, Feb 17, 2015 at 10:10:42AM +0100, Arnd Bergmann wrote:
> On Tuesday 17 February 2015 00:05:48 Mikko Rapeli wrote:
> >  #ifndef __ASM_GENERIC_UCONTEXT_H
> >  #define __ASM_GENERIC_UCONTEXT_H
> >  
> > +#include <asm-generic/signal.h>
> > +#include <asm/sigcontext.h>
> > +
> >  struct ucontext {
> > 
> 
> Including another asm-generic header here is a bad idea: it breaks
> if an architecture overrides asm/signal.h with its own version
> but wants to use the asm-generic/ucontext.h file.
> 
> It would be best to just use linux/signal.h here, which includes
> the correct architecture specific files.

Thanks, I will use asm/signal.h instead.

-Mikko

Re: [PATCH RFC v02 00/45] Compile test and fixes for exported header files

[Mikko Rapeli]

Thanks for all the reviews!

I'm going through the review feedback and will try to address all findings
and find correct ways to fix things but this will take some time from my
side.

-Mikko