Re: [PATCH 20/21] fdopendir: port to OS/2 kLIBC
KO Myung-Hun wrote:
>
>
> Paul Eggert wrote:
>> KO Myung-Hun wrote:
>>> I want to do later separately. No ?
>>
>> You could put it in a later patch, I suppose, but I wouldn't want to
>> install this patch without also installing the later patch.
>
> Ok. I'll try later.
>
Done. Review, please...
--
KO Myung-Hun
Using Mozilla SeaMonkey 2.7.2
Under OS/2 Warp 4 for Korean with FixPak #15
In VirtualBox v4.1.32 on Intel Core i7-3615QM 2.30GHz with 8GB RAM
Korean OS/2 User Community : http://www.ecomstation.co.kr
From 262a473f232c1565ff8c84dae94a9b8ef958a941 Mon Sep 17 00:00:00 2001
From: KO Myung-Hun
Date: Fri, 28 Nov 2014 16:47:12 +0900
Subject: [PATCH] opendir, closedir, dirfd, fdopendir: port to OS/2 kLIBC
* lib/closedir.c (closedir): Unregister fd if closedir() succeeds.
* lib/dirent.h.in (_gl_register_dirp_fd, _gl_unregister_dirp_fd):
Declare on kLIBC.
* lib/dirfd.c (struct dirp_fd_list): New. Structures to keep track of
fd associated with dirp.
(_gl_register_dirp_fd): New. Register fd associated with dirp to
dirp_fd_list.
(_gl_unregister_dirp_fd): New. Unregister fd with closing it.
(dirfd): Implemented for kLIBC.
* lib/fdopendir.c (fdopendir): Implemented for kLIBC.
* lib/opendir.c (opendir): New. Register fd and dirp pair if open()
succeeds.
* m4/closedir.m4 (gl_FUNC_CLOSEDIR): Check OS/2 as well.
* m4/dirfd.m4 (gl_FUNC_DIRFD): Likewise.
* m4/opendir.m4 (gl_FUNC_OPENDIR): Likewise.
* modules/closedir (Depends-on): Add dirfd.
* modules/dirfd (Depends-on): Add 'test $REPLACE_DIRFD = 1' to errno
condition.
(configure.ac): Add dirfd to LIBOBJS if $REPLACE_DIRFD = 1 as well.
* modules/fdopendir (Depends-on): Add dirfd.
* modules/opendir (Depends-on): Add dirfd.
---
lib/closedir.c| 6 -
lib/dirent.in.h | 7 ++
lib/dirfd.c | 68 +++
lib/fdopendir.c | 36 +
lib/opendir.c | 21 +
m4/closedir.m4| 2 +-
m4/dirfd.m4 | 3 ++-
m4/opendir.m4 | 2 +-
modules/closedir | 1 +
modules/dirfd | 5 ++--
modules/fdopendir | 1 +
modules/opendir | 1 +
12 files changed, 147 insertions(+), 6 deletions(-)
diff --git a/lib/closedir.c b/lib/closedir.c
index 940c6f9..b40181c 100644
--- a/lib/closedir.c
+++ b/lib/closedir.c
@@ -39,7 +39,7 @@
int
closedir (DIR *dirp)
{
-# if REPLACE_FCHDIR
+# if REPLACE_FCHDIR || defined __KLIBC__
int fd = dirfd (dirp);
# endif
int retval;
@@ -49,6 +49,10 @@ closedir (DIR *dirp)
retval = closedir (dirp);
+# ifdef __KLIBC__
+ if (!retval)
+_gl_unregister_dirp_fd (fd);
+# endif
#else
if (dirp->current != INVALID_HANDLE_VALUE)
diff --git a/lib/dirent.in.h b/lib/dirent.in.h
index 4822d6b..cce86d9 100644
--- a/lib/dirent.in.h
+++ b/lib/dirent.in.h
@@ -156,6 +156,13 @@ _GL_WARN_ON_USE (closedir, "closedir is not portable - "
# endif
_GL_FUNCDECL_RPL (dirfd, int, (DIR *) _GL_ARG_NONNULL ((1)));
_GL_CXXALIAS_RPL (dirfd, int, (DIR *));
+
+# ifdef __KLIBC__
+/* Gnulib internal hooks needed to maintain the dirfd metadata. */
+_GL_EXTERN_C int _gl_register_dirp_fd (int fd, DIR *dirp)
+ _GL_ARG_NONNULL ((2));
+_GL_EXTERN_C void _gl_unregister_dirp_fd (int fd);
+# endif
# else
# if defined __cplusplus && defined GNULIB_NAMESPACE && defined dirfd
/* dirfd is defined as a macro and not as a function.
diff --git a/lib/dirfd.c b/lib/dirfd.c
index 4d37928..aa6514f 100644
--- a/lib/dirfd.c
+++ b/lib/dirfd.c
@@ -22,11 +22,79 @@
#include
#include
+#ifdef __KLIBC__
+# include
+# include
+
+static struct dirp_fd_list
+{
+ DIR *dirp;
+ int fd;
+ struct dirp_fd_list *next;
+} *dirp_fd_start = NULL;
+
+/* Register fd associated with dirp to dirp_fd_list. */
+int
+_gl_register_dirp_fd (int fd, DIR *dirp)
+{
+ struct dirp_fd_list *new_dirp_fd;
+
+ new_dirp_fd = malloc (sizeof (*new_dirp_fd));
+ if (!new_dirp_fd)
+return -1;
+
+ new_dirp_fd->dirp = dirp;
+ new_dirp_fd->fd = fd;
+ new_dirp_fd->next = dirp_fd_start;
+
+ dirp_fd_start = new_dirp_fd;
+
+ return 0;
+}
+
+/* Unregister fd from dirp_fd_list with closing it */
+void
+_gl_unregister_dirp_fd (int fd)
+{
+ struct dirp_fd_list *dirp_fd;
+ struct dirp_fd_list *dirp_fd_prev;
+
+ for (dirp_fd_prev = NULL, dirp_fd = dirp_fd_start; dirp_fd;
+ dirp_fd_prev = dirp_fd, dirp_fd = dirp_fd->next)
+{
+ if (dirp_fd->fd == fd)
+{
+ if (dirp_fd_prev)
+dirp_fd_prev->next = dirp_fd->next;
+ else /* dirp_fd == dirp_fd_start */
+dirp_fd_start = dirp_fd_start->next;
+
+ close (fd);
+ free (dirp_fd);
+ break;
+}
+}
+}
+#endif
+
int
dirfd (DIR *dir_p)
{
int fd = DIR_TO_FD (dir_p);
if (fd == -1)
+#ifndef __KLIBC__
errno = ENOTSUP;
+#else
+{
+ struct dirp_fd_list *dirp_fd;
+
+ for (dirp_fd = dirp_fd_start; dirp_fd; dirp_fd = dirp_fd->next)
+if (dirp_fd->dirp == dir_p)
+ return dirp_fd->fd;
+
+ errno = EINVAL;
+
Re: [PATCH 15/21] freopen: workaround freopen() on OS/2 kLIBC
KO Myung-Hun wrote: > I also don't want to break the convention. > > Fixed. Thanks. This looks perfect now. Bruno
Re: [PATCH] printf: fix configure check on big endian systems
Pádraig Brady wrote on 2014-11-27:
> + printf: fix configure check on big endian systems
> + * m4/printf.m4 (gl_PRINTF_INFINITE_LONG_DOUBLE): Add missing bracket.
The same typo occurs in a few other places as well. I'm fixing them in the
same way.
2014-12-05 Bruno Haible
Fix LDBL80_WORDS macro on big endian platforms.
* m4/isfinite.m4 (gl_ISFINITEL_WORKS): Add missing parenthesis in
LDBL80_WORDS macro.
* m4/isinf.m4 (gl_ISINFL_WORKS): Likewise.
* m4/isnanl.m4 (gl_FUNC_ISNANL_WORKS): Likewise.
* tests/test-isfinite.c (test_isfinitel): Likewise.
* tests/test-isinf.c (test_isinfl): Likewise.
* tests/test-isnan.c (test_long_double): Likewise.
* tests/test-isnanl.h (main): Likewise.
* tests/test-snprintf-posix.h (LDBL80_WORDS): Add missing parenthesis.
* tests/test-sprintf-posix.h (LDBL80_WORDS): Likewise.
* tests/test-vasnprintf-posix.c (LDBL80_WORDS): Likewise.
* tests/test-vasprintf-posix.c (LDBL80_WORDS): Likewise.
Reported by Pádraig Brady.
diff --git a/ChangeLog b/ChangeLog
index 7379fa6..fb6723b 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,20 @@
+2014-12-05 Bruno Haible
+
+ Fix LDBL80_WORDS macro on big endian platforms.
+ * m4/isfinite.m4 (gl_ISFINITEL_WORKS): Add missing parenthesis in
+ LDBL80_WORDS macro.
+ * m4/isinf.m4 (gl_ISINFL_WORKS): Likewise.
+ * m4/isnanl.m4 (gl_FUNC_ISNANL_WORKS): Likewise.
+ * tests/test-isfinite.c (test_isfinitel): Likewise.
+ * tests/test-isinf.c (test_isinfl): Likewise.
+ * tests/test-isnan.c (test_long_double): Likewise.
+ * tests/test-isnanl.h (main): Likewise.
+ * tests/test-snprintf-posix.h (LDBL80_WORDS): Add missing parenthesis.
+ * tests/test-sprintf-posix.h (LDBL80_WORDS): Likewise.
+ * tests/test-vasnprintf-posix.c (LDBL80_WORDS): Likewise.
+ * tests/test-vasprintf-posix.c (LDBL80_WORDS): Likewise.
+ Reported by Pádraig Brady.
+
2014-12-02 KO Myung-Hun
git-version-gen: do not print new line characters
diff --git a/m4/isfinite.m4 b/m4/isfinite.m4
index 53ad909..bfd1ea3 100644
--- a/m4/isfinite.m4
+++ b/m4/isfinite.m4
@@ -1,4 +1,4 @@
-# isfinite.m4 serial 13
+# isfinite.m4 serial 14
dnl Copyright (C) 2007-2014 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
@@ -94,7 +94,7 @@ int main ()
# ifdef WORDS_BIGENDIAN
# define LDBL80_WORDS(exponent,manthi,mantlo) \
{ ((unsigned int) (exponent) << 16) | ((unsigned int) (manthi) >> 16), \
- ((unsigned int) (manthi) << 16) | (unsigned int) (mantlo) >> 16),\
+ ((unsigned int) (manthi) << 16) | ((unsigned int) (mantlo) >> 16), \
(unsigned int) (mantlo) << 16\
}
# else
diff --git a/m4/isinf.m4 b/m4/isinf.m4
index 7174ace..4d844b0 100644
--- a/m4/isinf.m4
+++ b/m4/isinf.m4
@@ -1,4 +1,4 @@
-# isinf.m4 serial 9
+# isinf.m4 serial 10
dnl Copyright (C) 2007-2014 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
@@ -101,7 +101,7 @@ int main ()
# ifdef WORDS_BIGENDIAN
# define LDBL80_WORDS(exponent,manthi,mantlo) \
{ ((unsigned int) (exponent) << 16) | ((unsigned int) (manthi) >> 16), \
- ((unsigned int) (manthi) << 16) | (unsigned int) (mantlo) >> 16),\
+ ((unsigned int) (manthi) << 16) | ((unsigned int) (mantlo) >> 16), \
(unsigned int) (mantlo) << 16\
}
# else
diff --git a/m4/isnanl.m4 b/m4/isnanl.m4
index 98b2b69..a26cc93 100644
--- a/m4/isnanl.m4
+++ b/m4/isnanl.m4
@@ -1,4 +1,4 @@
-# isnanl.m4 serial 17
+# isnanl.m4 serial 18
dnl Copyright (C) 2007-2014 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
@@ -177,7 +177,7 @@ int main ()
# ifdef WORDS_BIGENDIAN
# define LDBL80_WORDS(exponent,manthi,mantlo) \
{ ((unsigned int) (exponent) << 16) | ((unsigned int) (manthi) >> 16), \
- ((unsigned int) (manthi) << 16) | (unsigned int) (mantlo) >> 16),\
+ ((unsigned int) (manthi) << 16) | ((unsigned int) (mantlo) >> 16), \
(unsigned int) (mantlo) << 16\
}
# else
diff --git a/tests/test-isfinite.c b/tests/test-isfinite.c
index 5d320ea..5b3f1dd 100644
--- a/tests/test-isfinite.c
+++ b/tests/test-isfinite.c
@@ -181,7 +181,7 @@ test_isfinitel ()
# ifdef WORDS_BIGENDIAN
# define LDBL80_WORDS(exponent,manthi,mantlo) \
{ ((unsigned int) (exponent) << 16) | ((unsigned int) (manthi) >> 16), \
- ((unsigned int) (manthi) << 16) | (unsigned int) (mantlo) >> 16),\
+ ((unsigned int) (manthi) << 16) | ((unsigned int) (mantlo) >> 16), \
[PATCH] filevercmp, posixtm: avoid compiler warnings with -O3
* lib/filevercmp.h (filevercmp): Tag with _GL_ATTRIBUTE_PURE
* lib/posixtm.c: (IF_LINT): Define.
(posix_time_parse): Use it to void a "may be used uninitialized"
warning, seen only with -O3.
---
lib/filevercmp.h | 2 +-
lib/posixtm.c| 9 +
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/lib/filevercmp.h b/lib/filevercmp.h
index 72785cd..ad359eb 100644
--- a/lib/filevercmp.h
+++ b/lib/filevercmp.h
@@ -37,6 +37,6 @@
PREFIX VER2 SUFFIX) < 0.
This function is intended to be a replacement for strverscmp. */
-int filevercmp (const char *s1, const char *s2);
+int filevercmp (const char *s1, const char *s2) _GL_ATTRIBUTE_PURE;
#endif /* FILEVERCMP_H */
diff --git a/lib/posixtm.c b/lib/posixtm.c
index 2fe81ab..3023932 100644
--- a/lib/posixtm.c
+++ b/lib/posixtm.c
@@ -27,6 +27,13 @@
#include
#include
+/* Use this to suppress gcc's "...may be used uninitialized" warnings. */
+#ifdef lint
+# define IF_LINT(Code) Code
+#else
+# define IF_LINT(Code) /* empty */
+#endif
+
#if USE_UNLOCKED_IO
# include "unlocked-io.h"
#endif
@@ -118,6 +125,8 @@ posix_time_parse (struct tm *tm, const char *s, unsigned
int syntax_bits)
if (len != 8 && len != 10 && len != 12)
return 1;
+ IF_LINT(if (len < 8) return 1;)
+
if (dot)
{
if (!(syntax_bits & PDS_SECONDS))
--
2.1.0
Re: ARGP: Documentation of
Robert Hairgrove roberthairgrove.com> writes: > Robert Hairgrove roberthairgrove.com> writes: > > I am writing some little C++ wrapper classes for the argp functions. ... One more question: If argp_parse calls exit() and terminates the program, will the destructors of my C++ objects be called? Thank you.
Re: "test-lock" stuck on FreeBSD
On 4 December 2014 at 21:33, Daiki Ueno wrote: > Ed Maste writes: > >> OK, I can reproduce it in a single-vCPU bhyve VM (the default is 2). >> Do you mind checking if it's no longer reproducible in KVM with 2 or >> more vCPUs? > > Thanks, it indeed seems to fix the issue: Thanks for confirming. Upon further investigation this issue is fixed already in FreeBSD 10.1; further details are in FreeBSD PR 195098[1] for anyone interested. Hopefully we can get a 10.1 VM image for the pretest site - I'm happy to help if needed. [1] https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=195098
Re: [PATCH] printf: fix configure check on big endian systems
On Fri, Dec 5, 2014 at 2:37 AM, Bruno Haible wrote: > Pádraig Brady wrote on 2014-11-27: >> + printf: fix configure check on big endian systems >> + * m4/printf.m4 (gl_PRINTF_INFINITE_LONG_DOUBLE): Add missing bracket. > > The same typo occurs in a few other places as well. I'm fixing them in the > same way. > > > 2014-12-05 Bruno Haible > > Fix LDBL80_WORDS macro on big endian platforms. > * m4/isfinite.m4 (gl_ISFINITEL_WORKS): Add missing parenthesis in > LDBL80_WORDS macro. > * m4/isinf.m4 (gl_ISINFL_WORKS): Likewise. > * m4/isnanl.m4 (gl_FUNC_ISNANL_WORKS): Likewise. > * tests/test-isfinite.c (test_isfinitel): Likewise. > * tests/test-isinf.c (test_isinfl): Likewise. > * tests/test-isnan.c (test_long_double): Likewise. > * tests/test-isnanl.h (main): Likewise. > * tests/test-snprintf-posix.h (LDBL80_WORDS): Add missing parenthesis. > * tests/test-sprintf-posix.h (LDBL80_WORDS): Likewise. > * tests/test-vasnprintf-posix.c (LDBL80_WORDS): Likewise. > * tests/test-vasprintf-posix.c (LDBL80_WORDS): Likewise. > Reported by Pádraig Brady. Very nice!
Re: ARGP: Documentation of
Robert Hairgrove roberthairgrove.com> writes: > If argp_parse calls exit() and terminates the program, will the destructors > of my C++ objects be called? Silly question ... now I realize that they won't be called unless I take appropriate precautions, such as keeping a global vector of unique_ptr's or something along those lines.
recent changes to obstacks cause many out of memory failures in M4 master testsuite
Hi, To pick up bug fixes in maint.mk required for 'make public-submodule-commit' and dependers to work on Mac OS X 10.10, I just tried to update GNU M4's master branch to gnulib master HEAD, but that causes a raft of test failures where m4 bails out with a 'memory exhausted' error before completing the tests. Bisecting my way to the the first gnulib changeset that causes these failures, I landed at: bb2ab7e 2014-10-29 14:02 Alan Modra obstack: 64-bit obstack support, part 2 Which is not surprising as GNU M4 uses GNU obstack extensively throughout. Any revisions prior to that allow m4 to pass it's tests, but I'm afraid I don't understand the change well enough to point at the actual root of the issue. Hopefully someone here more knowledgable than I will be able to advise. Cheers, -- Gary V. Vaughan (gary AT gnu DOT org)
Re: recent changes to obstacks cause many out of memory failures in M4 master testsuite
On 12/05/2014 12:03 PM, Gary V. Vaughan wrote: > Bisecting my way to the the first gnulib changeset that causes these > failures, I landed at: > > bb2ab7e 2014-10-29 14:02 Alan Modra obstack: 64-bit obstack support, part > 2 > > Which is not surprising as GNU M4 uses GNU obstack extensively throughout. I bet it has to do with this: src/builtin.c: obstack_blank (data->obs, sizeof (symbol *)); src/builtin.c: obstack_blank_fast (obs, len); src/builtin.c: obstack_blank (obs, -1); src/macro.c: obstack_blank is documented to take a negative size to reduce the src/macro.c: obstack_blank (&argv_stack, -argc * sizeof (token_data *)); vs. the discussion on gnulib that we intentionally changed upstream obstack_blank to no longer do this, and that you HAVE to use obstack_blank_fast to get that effect. -- Eric Blake eblake redhat com+1-919-301-3266 Libvirt virtualization library http://libvirt.org signature.asc Description: OpenPGP digital signature
Re: [PATCH 0/5] obstacks again
On 10/29/2014 09:35 PM, Paul Eggert wrote: > Alan Modra wrote: > >> One thing though, I didn't put the ChangeLog diffs in the patch as I >> usually add them when committing. > > Oh, I missed that. I added them now. For Gnulib it's better to put > them into the patch. > >> It is no longer possible to shrink an obstack with obstack_blank (but >> you can still do that with obstack_blank_fast). > > Ouch, I hadn't noticed that. That's an incompatible change and I expect > it will break real-world usage for no particularly good reason, so we > really need to fix this. How about making the 2nd argument to > obstack_blank and obstack_blank_fast be of type ptrdiff_t rather than > size_t? It breaks GNU M4, for a starter :) But at least we predicted that it would happen, and I'm hoping the fallback of obstack_blank_fast does the job. -- Eric Blake eblake redhat com+1-919-301-3266 Libvirt virtualization library http://libvirt.org signature.asc Description: OpenPGP digital signature
Re: recent changes to obstacks cause many out of memory failures in M4 master testsuite
Hi Eric, > On Dec 5, 2014, at 7:11 PM, Eric Blake wrote: > > On 12/05/2014 12:03 PM, Gary V. Vaughan wrote: > >> Bisecting my way to the the first gnulib changeset that causes these >> failures, I landed at: >> >> bb2ab7e 2014-10-29 14:02 Alan Modra obstack: 64-bit obstack support, part >> 2 >> >> Which is not surprising as GNU M4 uses GNU obstack extensively throughout. > > I bet it has to do with this: > > src/builtin.c: obstack_blank (data->obs, sizeof (symbol *)); > src/builtin.c: obstack_blank_fast (obs, len); > src/builtin.c: obstack_blank (obs, -1); > src/macro.c: obstack_blank is documented to take a negative size to > reduce the > src/macro.c: obstack_blank (&argv_stack, -argc * sizeof (token_data *)); ACK. Only, on master I get: $ grep obstack_blank `git ls-files` grep: build-aux/gnulib: Is a directory m4/macro.c: obstack_blank (&context->trace_messages, start - len); modules/gnu.c: obstack_blank_fast (obs, len); modules/m4.c: obstack_blank (symbol_data->obs, sizeof *symbol_data->base); modules/m4.c: obstack_blank_fast (symbol_data->obs, sizeof *symbol_data->base); modules/m4.c: obstack_blank (obs, -1); modules/m4.c: obstack_blank (obs, min - uvalue); modules/m4.c: obstack_blank (obs, uvalue); modules/m4.c: obstack_blank (obs, min); Changing just the call in macro.c to obstack_blank_fast does indeed allow the tests to pass again. For correctness, I'll also change the `obstack_blank (obs, -1)` call, and audit the other call sites for possible changes along the same lines. > vs. the discussion on gnulib that we intentionally changed upstream > obstack_blank to no longer do this, and that you HAVE to use > obstack_blank_fast to get that effect. Bah, I didn't notice the nuances of that conversation at the time :-( Thanks for the pointer. Cheers, -- Gary V. Vaughan (gary AT gnu DOT org)
Re: [PATCH 0/5] obstacks again
On Dec 5, 2014, at 7:13 PM, Eric Blake wrote: > On 10/29/2014 09:35 PM, Paul Eggert wrote: >> Alan Modra wrote: >> >>> One thing though, I didn't put the ChangeLog diffs in the patch as I >>> usually add them when committing. >> >> Oh, I missed that. I added them now. For Gnulib it's better to put >> them into the patch. >> >>> It is no longer possible to shrink an obstack with obstack_blank (but >>> you can still do that with obstack_blank_fast). >> >> Ouch, I hadn't noticed that. That's an incompatible change and I expect >> it will break real-world usage for no particularly good reason, so we >> really need to fix this. How about making the 2nd argument to >> obstack_blank and obstack_blank_fast be of type ptrdiff_t rather than >> size_t? > > It breaks GNU M4, for a starter :) But at least we predicted that it > would happen, and I'm hoping the fallback of obstack_blank_fast does the > job. It does, thanks. Although for someone not following bug-gnulib fairly carefully, it is far from obvious why bumping gnulib makes your application suddenly use up all the available memory. Which is a shame, because, other than that M4 would have been perfectly functional after the gnulib bump without any special client code changes. If there's a way to at least diagnose negative arguments rather than silently change behavior, that would save other projects some migration headaches... Cheers, -- Gary V. Vaughan (gary AT gnu DOT org)
[PATCH] vasnprintf: fix potential use after free
* lib/vasnprintf.c (VASNPRINTF): Fix free-memory read, flagged by clang-analyzer 3.4.2. --- ChangeLog| 6 ++ lib/vasnprintf.c | 2 +- 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/ChangeLog b/ChangeLog index 1c4a7ab..41207ea 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,9 @@ +2014-12-06 Pádraig Brady + + vasnprintf: fix potential use after free + * lib/vasnprintf.c (VASNPRINTF): Fix free-memory read, + flagged by clang-analyzer 3.4.2. + 2014-12-05 Pádraig Brady filevercmp, posixtm: avoid compiler warnings with -O3 diff --git a/lib/vasnprintf.c b/lib/vasnprintf.c index 0484c4e..223be07 100644 --- a/lib/vasnprintf.c +++ b/lib/vasnprintf.c @@ -5184,13 +5184,13 @@ VASNPRINTF (DCHAR_T *resultbuf, size_t *lengthp, free (result); if (buf_malloced != NULL) free (buf_malloced); -CLEANUP (); errno = (saved_errno != 0 ? saved_errno : (dp->conversion == 'c' || dp->conversion == 's' ? EILSEQ : EINVAL)); +CLEANUP (); return NULL; } -- 2.1.0
[PATCH] apply _GL_ATTRIBUTE_PURE to some inline functions
clang 3.4.2 flagged these inline functions as pure
* lib/savewd.h (savewd_errno): Set _GL_ATTRIBUTE_PURE.
* lib/sig-handler.h (get_handler): Likewise.
* lib/stat-time.h (get_stat_{a,c,m,birth}time{,_ns}): Likewise.
* lib/timespec.h (timespec_cmp, timespec_sign): Likewize.
---
lib/savewd.h | 2 +-
lib/sig-handler.h | 2 +-
lib/stat-time.h | 16
lib/timespec.h| 4 ++--
4 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/lib/savewd.h b/lib/savewd.h
index cd0817b..4d2d7d6 100644
--- a/lib/savewd.h
+++ b/lib/savewd.h
@@ -125,7 +125,7 @@ int savewd_chdir (struct savewd *wd, char const *dir, int
options,
int savewd_restore (struct savewd *wd, int status);
/* Return WD's error number, or 0 if WD is not in an error state. */
-SAVEWD_INLINE int
+SAVEWD_INLINE int _GL_ATTRIBUTE_PURE
savewd_errno (struct savewd const *wd)
{
return (wd->state == ERROR_STATE ? wd->val.errnum : 0);
diff --git a/lib/sig-handler.h b/lib/sig-handler.h
index f4fe7ec..3af1a91 100644
--- a/lib/sig-handler.h
+++ b/lib/sig-handler.h
@@ -34,7 +34,7 @@ typedef void (*sa_handler_t) (int);
/* Return the handler of a signal, as a sa_handler_t value regardless
of its true type. The resulting function can be compared to
special values like SIG_IGN but it is not portable to call it. */
-SIG_HANDLER_INLINE sa_handler_t
+SIG_HANDLER_INLINE sa_handler_t _GL_ATTRIBUTE_PURE
get_handler (struct sigaction const *a)
{
#ifdef SA_SIGINFO
diff --git a/lib/stat-time.h b/lib/stat-time.h
index b3df6eb..9cbf8e0 100644
--- a/lib/stat-time.h
+++ b/lib/stat-time.h
@@ -54,7 +54,7 @@ _GL_INLINE_HEADER_BEGIN
#endif
/* Return the nanosecond component of *ST's access time. */
-_GL_STAT_TIME_INLINE long int
+_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE
get_stat_atime_ns (struct stat const *st)
{
# if defined STAT_TIMESPEC
@@ -67,7 +67,7 @@ get_stat_atime_ns (struct stat const *st)
}
/* Return the nanosecond component of *ST's status change time. */
-_GL_STAT_TIME_INLINE long int
+_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE
get_stat_ctime_ns (struct stat const *st)
{
# if defined STAT_TIMESPEC
@@ -80,7 +80,7 @@ get_stat_ctime_ns (struct stat const *st)
}
/* Return the nanosecond component of *ST's data modification time. */
-_GL_STAT_TIME_INLINE long int
+_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE
get_stat_mtime_ns (struct stat const *st)
{
# if defined STAT_TIMESPEC
@@ -93,7 +93,7 @@ get_stat_mtime_ns (struct stat const *st)
}
/* Return the nanosecond component of *ST's birth time. */
-_GL_STAT_TIME_INLINE long int
+_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE
get_stat_birthtime_ns (struct stat const *st)
{
# if defined HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC
@@ -108,7 +108,7 @@ get_stat_birthtime_ns (struct stat const *st)
}
/* Return *ST's access time. */
-_GL_STAT_TIME_INLINE struct timespec
+_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE
get_stat_atime (struct stat const *st)
{
#ifdef STAT_TIMESPEC
@@ -122,7 +122,7 @@ get_stat_atime (struct stat const *st)
}
/* Return *ST's status change time. */
-_GL_STAT_TIME_INLINE struct timespec
+_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE
get_stat_ctime (struct stat const *st)
{
#ifdef STAT_TIMESPEC
@@ -136,7 +136,7 @@ get_stat_ctime (struct stat const *st)
}
/* Return *ST's data modification time. */
-_GL_STAT_TIME_INLINE struct timespec
+_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE
get_stat_mtime (struct stat const *st)
{
#ifdef STAT_TIMESPEC
@@ -151,7 +151,7 @@ get_stat_mtime (struct stat const *st)
/* Return *ST's birth time, if available; otherwise return a value
with tv_sec and tv_nsec both equal to -1. */
-_GL_STAT_TIME_INLINE struct timespec
+_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE
get_stat_birthtime (struct stat const *st)
{
struct timespec t;
diff --git a/lib/timespec.h b/lib/timespec.h
index 872cbb7..dfc1277 100644
--- a/lib/timespec.h
+++ b/lib/timespec.h
@@ -74,7 +74,7 @@ make_timespec (time_t s, long int ns)
The (int) cast avoids a gcc -Wconversion warning. */
-_GL_TIMESPEC_INLINE int
+_GL_TIMESPEC_INLINE int _GL_ATTRIBUTE_PURE
timespec_cmp (struct timespec a, struct timespec b)
{
return (a.tv_sec < b.tv_sec ? -1
@@ -84,7 +84,7 @@ timespec_cmp (struct timespec a, struct timespec b)
/* Return -1, 0, 1, depending on the sign of A. A.tv_nsec must be
nonnegative. */
-_GL_TIMESPEC_INLINE int
+_GL_TIMESPEC_INLINE int _GL_ATTRIBUTE_PURE
timespec_sign (struct timespec a)
{
return a.tv_sec < 0 ? -1 : a.tv_sec || a.tv_nsec;
--
2.1.0
Re: [PATCH 0/5] obstacks again
On Fri, Dec 05, 2014 at 09:02:46PM +, Gary V. Vaughan wrote: > If there's a way to at least diagnose negative arguments rather than silently > change behavior, that would save other projects some migration headaches... We could diagnose one of the m4 uses of obstack_blank at compile time, but not the one in m4/macro.c:trace_flush which has a non-constant shrink value.. Yes, I agree this can cause some pain, but at least dying with out-of-memory is a loud symptom. The alternative is to do as Paul suggested and make obstack_blank accept a negative length argument, but that would - kill >2G obstacks on 32-bit targets, - lose the nice symmetry with other obstack functions, and obstack_blank_fast is the right interface to use for shrinking. -- Alan Modra Australia Development Lab, IBM
Re: [PATCH] vasnprintf: fix potential use after free
On 12/05/2014 06:23 PM, Pádraig Brady wrote: > * lib/vasnprintf.c (VASNPRINTF): Fix free-memory read, > flagged by clang-analyzer 3.4.2. > --- > ChangeLog| 6 ++ > lib/vasnprintf.c | 2 +- > 2 files changed, 7 insertions(+), 1 deletion(-) > > +++ b/lib/vasnprintf.c > @@ -5184,13 +5184,13 @@ VASNPRINTF (DCHAR_T *resultbuf, size_t *lengthp, >free (result); > if (buf_malloced != NULL) >free (buf_malloced); > -CLEANUP (); > errno = >(saved_errno != 0 > ? saved_errno > : (dp->conversion == 'c' || dp->conversion == 's' >? EILSEQ >: EINVAL)); > +CLEANUP (); Ouch. This is a bug. The whole point of assigning to errno after CLEANUP() is that CLEANUP() may invalidate the value stored in errno. I suggest doing something like: if (saved_errno == 0) saved_errno = dp->conversion == 'c' || dp->conversion == 's' ? EILSEQ : EINVAL; CLEANUP(); errno = saved_errno; -- Eric Blake eblake redhat com+1-919-301-3266 Libvirt virtualization library http://libvirt.org signature.asc Description: OpenPGP digital signature
