There are some very frequently used tags that checkpatch can look
for as an ending to the commit text rather than using a pattern,
which can generate false positives in the "no commit text" rule.
Change-Id: I5b4400017b8273bcd9f5a59b3e28965c0062bef4
Signed-off-by: Steve Muckle <smuckle@codeaurora.org>
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Steve Muckleauthored and Trilok Sonicommited 2016-08-08 15:22:58 -07:00
Commit text is almost always necessary to explain why a change is
needed. Exceptions are rare enough that these can be granted through
manual checkpatch overrides.
Change-Id: I926b2276f717940c2fec77a6709fa3088b1bf0c3
Signed-off-by: Steve Muckle <smuckle@codeaurora.org>
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Steve Muckleauthored and Trilok Sonicommited 2016-08-08 15:22:57 -07:00
When patches contain extended character sets, patches will contain
MIME headers after the subject line, which should not be confused
for a too-long summary line.
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit 8e6b9d3790595198a34320f1c3f4504cd258fed1)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:56 -07:00
Fixed case when no warning generated for long subject line that is
wrapped to more than one line, and all lines are less than line limit.
New warning message added:
"non-blank line after summary line"
Now there are two warnings possible for the subject line, the original
line over limit and the new one. Depending on the error(s) any
combination of the two warnings are possible.
Commit text requirements now:
1) Must be less than 75 characters
2) Must be followed by blank line.
Signed-off-by: David Keitel <dkeitel@codeaurora.org>
David Keitelauthored and Trilok Sonicommited 2016-08-08 15:22:54 -07:00
Warn on summary or commit text lines greater than 75 characters.
The summary and commit text are indented and may wrap on a terminal
if they are longer than 75 characters.
Signed-off-by: David Keitel <dkeitel@codeaurora.org>
David Keitelauthored and Trilok Sonicommited 2016-08-08 15:22:54 -07:00
Currently checkpatch.pl throws error for any use of "dsb"
keyword. Add exceptions when it is used in a variable or a
function name. Expectations are when 'dsb' keyword is used in a
variable or a function name, it is prefixed with class [-_>*\.]
and/or suffixed with class [-_\.;].
Change-Id: I06cd21e42135334bb41cf31f13088b118fcf3ac9
Signed-off-by: Sarangdhar Joshi <spjoshi@codeaurora.org>
Sarangdhar Joshiauthored and Trilok Sonicommited 2016-08-08 15:22:52 -07:00
Commit "scripts: checkpatch.pl: warn on invalid credentials" added
to checkpatch.pl logic to unconditionally check for invalid author
credentials. Add a flag to allow override of this behavior.
Change-Id: Ib254cc6f73c4de857ba923e6170750262cde01ad
Signed-off-by: Jeff Johnson <jjohnson@codeaurora.org>
Jeff Johnsonauthored and Trilok Sonicommited 2016-08-08 15:22:52 -07:00
@quicinc.com identities are not allowed. Check the "From:" field in the
patch, equivalent to the author in the git commit, and the Signed-off-by
field.
Signed-off-by: Bryan Huntsman <bryanh@codeaurora.org>
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
[imaund@codeaurora.org: Resolved context conflicts]
Signed-off-by: Ian Maund <imaund@codeaurora.org>
Bryan Huntsmanauthored and Trilok Sonicommited 2016-08-08 15:22:51 -07:00
Now that mb() does what we want, dsb() should be discouraged.
Change-Id: Ib8fe8f44f669753c3d91fac3c6e598e117d6d90e
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit 9c0619be7b93ad114d6f33a749d905ddff93df7d)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:50 -07:00
filp_open allows people to get around the ban on sys_open.
Close the loophole.
Change-Id: I6e2be62e848cbc064e07008d0886c0d003c8be4b
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit bb181a18a813a70176f71a0c64aa572fcfbef0f0)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:49 -07:00
Improve parsing of multiline macros which run beyond the available
diff context. These beyond-the-horizon macros previously caused
two distinct naughty behaviors:
- The scanner, confused by the trailing backslash, would grab
the header of the next context hunk and treat it as the last
line of the preceding macro.
- The analyzer, unable to fully reduce the macro, would blame
the patch for submitting an unbalanced or unprotected macro.
Change-Id: I6b7dd3d577c524d30b59dff7b20393bb5135f16d
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit ddd028c47b4d91aa9c0e97445eb584b2de367769)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:47 -07:00
Unbounded string functions are overflow risks. The 'n'
versions of those functions should be used instead.
Change-Id: Ice0fb3ebdae9aa88cc7e764ffdf68cbed857febf
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit 15e1e97d66dd6a6039c1ec2bd549a632fe361128)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:46 -07:00
Forbid read[bwl], write[bwl], in[bwl], and out[bwl], as they
contain a 'stealth barrier' which can harm performance.
Developers are expected to call appropriate __raw_* or *_relaxed
APIs and manage barriers explicitly.
Change-Id: Ie4da221c91a0505917199db9e2fdb704c3e47d44
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit 032fd4ba09e195d9913c08f460130da9905936ef)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:46 -07:00
This commit is a cherry pick and squash of commits:
85c2ee62d37c19456c6dc83db262123956f010ac
4bc7c6001daba7d4037f54f67bae7fa90f759402
checkpatch: Handle continuation headers.
Continuation headers baffle checkpatch, as it can only operate
on one line of context at a time. When continuation headers are found,
put them up with the header they're continuing so the whole thing
can be parsed in a single line of context.
checkpatch: Don't treat diffs as patches.
The patch-header cleanup code assumed that it would only ever
see patches, which was of course hogwash. This lead to crazy
results as it tried to wrap what it thought were continuation
lines at the beginnings of raw diffs.
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit 57d50ae730a057ee1099a94a397475bfd147d97b)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:45 -07:00
Use of the sys_open/close/read/write system calls from within
kernel code is inappropriate, and now triggers errors.
Change-Id: I98e20513c257d0664684b7144585853f617d771a
Signed-off-by: Gregory Bean <gbean@codeaurora.org>
(cherry picked from commit ee62f2afcac1bcb180b2f0dddf2c8f5cda54bc5b)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
[abhimany: resolve trivial merge conflicts]
Signed-off-by: Abhimanyu Kapur <abhimany@codeaurora.org>
Gregory Beanauthored and Trilok Sonicommited 2016-08-08 15:22:43 -07:00
Display context so users can see where there error was
OLD:
ERROR: illegal return value, please use an error code
NEW:
ERROR: illegal return value, please use an error code
#152: test.c:5:
+ return -1;
Change-Id: I098004d9a5dbeb6c39b35b84ac94fd7a861849d7
Signed-off-by: Patrick Pannuto <ppannuto@codeaurora.org>
(cherry picked from commit 8d09803de5d629f3ebc9cb357a40b5937a5d171e)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Patrick Pannutoauthored and Trilok Sonicommited 2016-08-08 15:22:43 -07:00
The only legal integer return is 0, anything else
following "return" should be -ERRCODE or a function.
http://lkml.org/lkml/2010/7/23/318
There's lots of "return -1;" statements in this patch - it's obscene
that this is used to indicate "some error occurred" in kernel space
rather than a real errno value - even when an existing function
(eg, request_irq) gave you an error code already.
Please note this for the future - and please review patches on this
point internally first.
Change-Id: I16268b2ee034f0b3b899115e45c28acfa734ddec
Signed-off-by: Patrick Pannuto <ppannuto@codeaurora.org>
(cherry picked from commit 39531a47164294315b5a7256b520fe22d6e87013)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Patrick Pannutoauthored and Trilok Sonicommited 2016-08-08 15:22:42 -07:00
Currently checkpatch is enforcing an 80 character column
limit. This should not be applied for patches modifying
checkpatch.
Change-Id: I8e8935e633d79a7ee535ce6a90e54447a22d9a91
Signed-off-by: Israel Schlesinger <israels@codeaurora.org>
(cherry picked from commit 4aa1b864119e2d0e49efefcd6dba7349f13c3939)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Israel Schlesingerauthored and Trilok Sonicommited 2016-08-08 15:22:41 -07:00
Add CLK_* macros used in MSM clock drivers to the list of exceptions
for the "Macros with complex values should be enclosed in parenthesis"
test.
Also, allow spaces following open branckets for macros where the
macro is the first token on the line, the space preceds a decimal
number, and the line ends with ")," or ")". Such arrangements can
be useful for aligning numerical columns of tables when the rows
are described by macros.
Change-Id: I7701119ada2ea8fd646e5448eae51786bbf1e8fa
Signed-off-by: Matt Wagantall <mattw@codeaurora.org>
(cherry picked from commit ed6d6ed1c6b8f6016ea5676d075331e31b7ac1f8)
Signed-off-by: Stepan Moskovchenko <stepanm@codeaurora.org>
Matt Wagantallauthored and Trilok Sonicommited 2016-08-08 15:22:38 -07:00
If no filenames are given, then read the patch from stdin.
Link: http://lkml.kernel.org/r/a8784f291ccb5067361992bf5d41ff6cfb0ce5cb.1469830917.git.allenbh@gmail.com
Signed-off-by: Allen Hubbe <allenbh@gmail.com>
Acked-by: Joe Perches <joe@perches.com>
Cc: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Allen Hubbeauthored and Linus Torvaldscommited 2016-08-02 19:35:13 -04:00
Signoff was not checked if the filename is '-', indicating reading the
patch from stdin. Commands such as the below would not warn about a
missing signoff, because the patch filename is '-'. This change allows
checkpatch to warn about a missing signoff, even if the input filename
is '-', but only if the patch has a commit message.
git show --pretty=email | scripts/checkpatch.pl -
A more common use of checkpatch with stdin is for piping git diff
through checkpatch. The diff output would not contain a commit message,
and therefore it would not contain a signoff line. For this common use
case, a warning should not be printed about the missing signoff. With
this change we will only warn about a missing signoff if the input
contains a commit message.
git diff | scripts/checkpatch.pl -
Before this patch, a workaround for the first command was to refer to
stdin by a name other than '-'. The workaround is not an elegant
solution, because elsewhere checkpatch uses the fact that filename
equals '-', such as in setting '$vname' to 'Your patch' for stdin. The
command below would report "/dev/stdin has style problems" instead of
"Your patch has style problems."
git show --pretty=email | scripts/checkpatch.pl /dev/stdin
Link: http://lkml.kernel.org/r/48be31e414bddc65bccfa6b1322359be9ba032eb.1469670589.git.allenbh@gmail.com
Signed-off-by: Allen Hubbe <allenbh@gmail.com>
Acked-by: Joe Perches <joe@perches.com>
Cc: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Allen Hubbeauthored and Linus Torvaldscommited 2016-08-02 19:35:13 -04:00
Fix false positive warning of identifiers ending in signed with an =
assignment of WARNING: Prefer 'signed int' to bare use of 'signed'.
Link: http://lkml.kernel.org/r/6a0e24c3e9102337528ecfcbbe91a0eb5b4820ed.1469529497.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reported-by: Alan Douglas <alanjhd@gmail.com>
Acked-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-08-02 19:35:12 -04:00
BIT macro cannot be exported to UAPI, don't complain about it.
Link: http://lkml.kernel.org/r/1468707033-16173-1-git-send-email-tomas.winkler@intel.com
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Acked-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Tomas Winklerauthored and Linus Torvaldscommited 2016-08-02 19:35:12 -04:00
Using \b isn't good enough to isolate what appears to be a commit id in
a commit message.
Make sure there is a space or a quote like character after a continuous
run of hexadecimal characters that could be a commit id.
Link: http://lkml.kernel.org/r/fdd22b47463a21c21132edbb8aa35e372950a1e6.1468869915.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Cc: "Zhuo, Qiuxu" <qiuxu.zhuo@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-08-02 19:35:11 -04:00
The __pmem address space was meant to annotate codepaths that touch
persistent memory and need to coordinate a call to wmb_pmem(). Now that
wmb_pmem() is gone, there is little need to keep this annotation.
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The --git <commit-count> shortcut can be confused by a tag with a dash
like v4.4-rc1.
Improve the test to verify the <commit-count> expression ends with a
dash followed by a numeric value.
Improve the git log result to verify the "<sha1> <subject>" output
as well.
Link: http://lkml.kernel.org/r/c4a3f759291d967641860c3a54bb81177f34325f.1462711962.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
checkpatch currently calls git log multiple times to first get the
<revision range> sha1 values and again to get the subject for each
individual sha1 commit.
Always get the sha1 and subject at the same time instead. Store the
subject in a sha1 hash to avoid the second git log exec.
Link: http://lkml.kernel.org/r/274efab2332ad2308ab5de85a95d255f6e2de5f3.1462711962.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
It's sometimes useful to scan already committed patches.
Add --git <revision range> to scan specific or multiple commits.
Single commits are scanned with
--git <rev>
Multiple commits are scanned with
--git <range>
--git <commit>-<count>
[joe@perches.com:
o Don't exec git for each <commit>-<count>,
use a single "git log -<count> <commit>"
o Consolidate the git exec for the <range> and <commit>-<count> variants
o Output 12 character commit hash ids
o Don't scan git commit merges
o Use -M to reduce the size of rename commits]
Signed-off-by: "Du, Changbin" <changbin.du@intel.com>
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Du, Changbinauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
The message types are not currently knowable without reading the code.
Add a mechanism to see what they are.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
The --fix option is relatively unknown and underutilized.
Add some text to show that it's available when style defects are found.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
Add a test for use of ACCESS_ONCE that could be written using READ_ONCE or
WRITE_ONCE.
--fix it too if desired.
The WRITE_ONCE fixes are less correct than the coccinelle script below as
checkpatch cannot have a completely correct "expression" mechanism because
checkpatch works on patches and not complete files.
$ cat access_once.cocci
@@
expression e1;
expression e2;
@@
- ACCESS_ONCE(e1) = e2
+ WRITE_ONCE(e1, e2)
@@
expression e1;
@@
- ACCESS_ONCE(e1)
+ READ_ONCE(e1)
Signed-off-by: Joe Perches <joe@perches.com>
Cc: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
It's somewhat common and in general a defect for c90 keywords to
not start on a tabstop.
Add a test for this condition and warn when it occurs.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
A "." dereference to an all uppercase structure member can be
incorrectly reported as a CONSTANT_COMPARISON.
ie: "if (table[i].PANELID == tempdx)"
Fix it by checking for "." before the constant test.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
Using #if defined CONFIG_<FOO> || defined CONFIG_<FOO>_MODULE is
more verbose than necessary and IS_ENABLED(CONFIG_<FOO>) is preferred.
So add a test and a message for it.
--fix it to if desired.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-05-20 17:58:30 -07:00
Merge first patch-bomb from Andrew Morton:
- some misc things
- ofs2 updates
- about half of MM
- checkpatch updates
- autofs4 update
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (120 commits)
autofs4: fix string.h include in auto_dev-ioctl.h
autofs4: use pr_xxx() macros directly for logging
autofs4: change log print macros to not insert newline
autofs4: make autofs log prints consistent
autofs4: fix some white space errors
autofs4: fix invalid ioctl return in autofs4_root_ioctl_unlocked()
autofs4: fix coding style line length in autofs4_wait()
autofs4: fix coding style problem in autofs4_get_set_timeout()
autofs4: coding style fixes
autofs: show pipe inode in mount options
kallsyms: add support for relative offsets in kallsyms address table
kallsyms: don't overload absolute symbol type for percpu symbols
x86: kallsyms: disable absolute percpu symbols on !SMP
checkpatch: fix another left brace warning
checkpatch: improve UNSPECIFIED_INT test for bare signed/unsigned uses
checkpatch: warn on bare unsigned or signed declarations without int
checkpatch: exclude asm volatile from complex macro check
mm: memcontrol: drop unnecessary lru locking from mem_cgroup_migrate()
mm: migrate: consolidate mem_cgroup_migrate() calls
mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous
...
This patch escapes a regex that uses left brace.
Using checkpatch.pl with Perl 5.22.0 generates the warning: "Unescaped
left brace in regex is deprecated, passed through in regex;"
Comment from regcomp.c in Perl source: "Currently we don't warn when the
lbrace is at the start of a construct. This catches it in the middle of
a literal string, or when it's the first thing after something like
"\b"."
This works as a complement to 4e5d56bd ("checkpatch: fix left brace
warning").
Signed-off-by: Geyslan G. Bem <geyslan@gmail.com>
Signed-off-by: Joe Perches <joe@perches.com>
Suggested-by: Peter Senna Tschudin <peter.senna@gmail.com>
Cc: Eddie Kovsky <ewk@edkovsky.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Geyslan G. Bemauthored and Linus Torvaldscommited 2016-03-15 16:55:16 -07:00
Improve the test to allow casts to (unsigned) or (signed) to be found
and fixed if desired.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-03-15 16:55:16 -07:00
Kernel style prefers "unsigned int <foo>" over "unsigned <foo>" and
"signed int <foo>" over "signed <foo>".
Emit a warning for these simple signed/unsigned <foo> declarations. Fix
it too if desired.
Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-03-15 16:55:16 -07:00
asm volatile and all its variants like __asm__ __volatile__ ("<foo>")
are reported as errors with "Macros with with complex values should be
enclosed in parentheses".
Make an exception for these asm volatile macro definitions by converting
the "asm volatile" to "asm_volatile" so it appears as a single function
call and the error isn't reported.
Signed-off-by: Joe Perches <joe@perches.com>
Reported-by: Jeff Merkey <linux.mdb@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-03-15 16:55:16 -07:00
In C programming language, we don't have a easy way to privatize a
member of a structure. However in kernel, sometimes there is a need to
privatize a member in case of potential bugs or misuses.
Fortunately, the noderef attribute of sparse is a way to privatize a
member, as by defining a member as noderef, the address-of operator on
the member will produce a noderef pointer to that member, and if anyone
wants to dereference that kind of pointers to read or modify the member,
sparse will yell.
Based on this, __private modifier and related operation ACCESS_PRIVATE()
are introduced, which could help detect undesigned public uses of
private members of structs. Here is an example of sparse's output if it
detect an undersigned public use:
| kernel/rcu/tree.c:4453:25: warning: incorrect type in argument 1 (different modifiers)
| kernel/rcu/tree.c:4453:25: expected struct raw_spinlock [usertype] *lock
| kernel/rcu/tree.c:4453:25: got struct raw_spinlock [noderef] *<noident>
Also, this patch improves compiler.h a little bit by adding comments for
"#else" and "#endif".
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Boqun Fengauthored and Paul E. McKenneycommited 2016-02-23 19:59:53 -08:00
A simple search over the kernel souce displays a number of correctly
defined multiline macro, which generally are used as an array element
initializer:
% find ../linux -type f | xargs grep -B1 -H '^[:space]*\[.*\\$'
However checkpatch.pl unexpectedly complains about all these macro
definitions:
% ./scripts/checkpatch.pl --types COMPLEX_MACRO -f include/linux/perf/arm_pmu.h
ERROR: Macros with complex values should be enclosed in parentheses
+#define PERF_MAP_ALL_UNSUPPORTED \
+ [0 ... PERF_COUNT_HW_MAX - 1] = HW_OP_UNSUPPORTED
The change intends to fix this type of false positives by flattening
only array members and skipping array element designators.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Acked-by: Joe Perches <joe@perches.com>
Cc: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Vladimir Zapolskiyauthored and Linus Torvaldscommited 2016-01-20 17:09:18 -08:00
The current test excludes any macro with ## concatenation from being
reported with hidden flow control.
Some macros are used with return or goto statements along with ##args or
##__VA_ARGS__. A somewhat common case is a logging macro like
pr_info(fmt, ...) then a return or goto statement.
Check the concatenated variable for args or __VA_ARGS__ and allow those
macros to also be reported when they contain a return or goto.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-01-20 17:09:18 -08:00
Linus Torvalds wrote:
> I can't but help to react that this:
> #define IOMMU_ERROR_CODE (~(unsigned long) 0)
> Not that this *matters*, but it's a bit odd to have to cast constants
> to perfectly regular C types.
So add a test that looks for constants that are cast to
standard C90 int or longer types and suggest using C90
"6.4.4.1 Integer constants" integer-suffixes instead.
Miscellanea:
o Add a --fix option too
Signed-off-by: Joe Perches <joe@perches.com>
Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Joe Perchesauthored and Linus Torvaldscommited 2016-01-20 17:09:18 -08:00
Add virt_ barriers to list of barriers to check for
presence of a comment.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Julian Calaby <julian.calaby@gmail.com>
Acked-by: Joe Perches <joe@perches.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Michael S. Tsirkincommited 2016-01-12 20:47:08 +02:00
Introduction of __smp barriers cleans up a bunch of duplicate code, but
it gives people an additional handle onto a "new" set of barriers - just
because they're prefixed with __* unfortunately doesn't stop anyone from
using it (as happened with other arch stuff before.)
Add a checkpatch test so it will trigger a warning.
Reported-by: Russell King <linux@arm.linux.org.uk>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Julian Calaby <julian.calaby@gmail.com>
Acked-by: Joe Perches <joe@perches.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Michael S. Tsirkincommited 2016-01-12 20:47:07 +02:00
Steve Muckle
authored and
Trilok Soni
commited 
Gregory Bean
authored and
David Keitel
authored and
Sarangdhar Joshi
authored and
Jeff Johnson
authored and
Bryan Huntsman
authored and
Patrick Pannuto
authored and
Israel Schlesinger
authored and
Matt Wagantall
authored and
Abhijeet Dharmapurikar
authored and
Allen Hubbe
authored and
Linus Torvalds
commited 
Joe Perches
authored and
Tomas Winkler
authored and
Dan Williams
commited 
Du, Changbin
authored and
Geyslan G. Bem
authored and
Boqun Feng
authored and
Paul E. McKenney
commited 
Vladimir Zapolskiy
authored and
Michael S. Tsirkin
commited