On Wed, 2019-06-05 at 07:14 -0300, Mauro Carvalho Chehab wrote:
> Em Tue, 04 Jun 2019 22:22:05 -0700
> Joe Perches <j...@perches.com> escreveu:
>
> > On Wed, 2019-06-05 at 07:10 +0200, Greg KH wrote:
> > > On Wed, Jun 05, 2019 at 01:10:41PM +0900, Masahiro Yamada wrote:
> > > > On Wed, Jun 5, 2019 at 3:21 AM Arnd Bergmann <a...@arndb.de> wrote:
> > []
> > > > This means we cannot reliably use uint{8,16,32,64}_t in UAPI headers.
> > >
> > > We should not be doing that as they are in the userspace "namespace" of
> > > variables, not in the kernel namespace. We've been over this many times
> > > in the past :(
> >
> > Just not very successfully...
> >
> > $ git grep -w -P 'u?_?int(?:8|16|32|64)_t' include/uapi | wc -l
> > 342
> > $ git grep -w -P --name-only 'u?_?int(?:8|16|32|64)_t' include/uapi | wc -l
> > 13
>
> Out of curiosity, I decided to check those occurrences...
>
> About half of those 13 files are false-positives:
>
> - bpf.h, pps.h and amdgpu_drm.h use those int types only inside comments;
> - drm.h and coda.h have their own typedefs for those int types;
> - vmwgfx_drm.h includes drm.h (which has the typedefs).
>
> So, only 6 headers actually use posix types in a way that it seems that
> it would require including stdint.h:
>
> - include/uapi/linux/fuse.h
>
> This one explicitly includes stdint.h if !__KERNEL__
>
> - include/uapi/linux/netfilter_bridge/ebtables.h,
> include/uapi/linux/sctp.h,
> include/uapi/scsi/scsi_netlink.h and
> include/uapi/scsi/scsi_netlink_fc.h
>
> They include linux/types.h unconditionally, relying on
> scripts/headers_install.sh to remove it;
>
> - include/uapi/scsi/scsi_bsg_fc.h
>
> It doesn't include anything. In other words, it assumes that the c file
> would include either linux/types.h or stdint.h.
>
> ---
>
> Not saying that this is a good idea, but, as we have already a script that
> it is meant to sanitize uAPI header files when installing them
> (scripts/headers_install.sh), one could modify it (or convert to perl/python)
> in a way that it would be doing something like[1]:
>
> sed -E
> ...
> -e 's,//(.*),/* \1 */,'
>
> [1] the actual rule should be more complex than that, in order to avoid
> replacing // inside /**/ comments.
Perhaps a checkpatch change too:
The first block updates unsigned only bitfields
The second tests uapi definitions and suggests "__<kernel_types"
---
scripts/checkpatch.pl | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
index c33c5002f190..afc4bb05a987 100755
--- a/scripts/checkpatch.pl
+++ b/scripts/checkpatch.pl
@@ -3747,7 +3747,7 @@ sub process {
}
# check for declarations of signed or unsigned without int
- while ($line =~
m{\b($Declare)\s*(?!char\b|short\b|int\b|long\b)\s*($Ident)?\s*[=,;\[\)\(]}g) {
+ while ($line =~
m{\b($Declare)\s*(?!char\b|short\b|int\b|long\b)\s*($Ident)?\s*[=,;:\[\)\(]}g) {
my $type = $1;
my $var = $2;
$var = "" if (!defined $var);
@@ -5905,10 +5905,10 @@ sub process {
"Using weak declarations can have unintended link
defects\n" . $herecurr);
}
-# check for c99 types like uint8_t used outside of uapi/ and tools/
- if ($realfile !~ m@\binclude/uapi/@ &&
- $realfile !~ m@\btools/@ &&
- $line =~ /\b($Declare)\s*$Ident\s*[=;,\[]/) {
+# check for c99 types like uint8_t used outside of tools/
+# for uapi, suggest using __<types>, otherwise use <types> like s8, u32, etc...
+ if ($realfile !~ m@\btools/@ &&
+ $line =~ /\b($Declare)\s*$Ident\s*[=,;:\[]/) {
my $type = $1;
if ($type =~ /\b($typeC99Typedefs)\b/) {
$type = $1;
@@ -5916,6 +5916,9 @@ sub process {
$kernel_type = 's' if ($type =~ /^_*[si]/);
$type =~ /(\d+)/;
$kernel_type .= $1;
+ if ($realfile =~ m@\binclude/uapi/@) {
+ $kernel_type = "__" . $kernel_type;
+ }
if (CHK("PREFER_KERNEL_TYPES",
"Prefer kernel type '$kernel_type' over
'$type'\n" . $herecurr) &&
$fix) {
