ioctl_userfaultfd.2: Add UFFD_FEATURE_WP_ASYNC

Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ld.so.8: Clarify LD_PROFILE in secure-execution mode

Different than specified, glibc 2.2.5 does not ignore LD_PROFILE in
secure-execution mode.  Instead, it uses the default LD_PROFILE_OUTPUT
(/var/profile) folder.

Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ld.so.8: Describe the --list-diagnostics loader option

It was added on glibc 2.33 as a way to print system diagnostic
information in a machine-readable format.

Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

etc/clang-tidy/config.yaml: Ignore 'modernize-macro-to-enum'

Signed-off-by: Alejandro Colomar <alx@kernel.org>

ld.so.8: Add filtered-out environment variable in secure-execution mode

They are defined by unservars.h file in glibc code [1].

[1] https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/generic/unsecvars.h;h=8278c50a84ffda18e28c178d390f1899d8306e26;hb=HEAD

Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ld.so.8: Update environment variables for recent glibc versions

Recent glibc versions removed support for LD_TRACE_PRELINKING,
LD_ASSUME_KERNEL, and LD_USE_LOAD_BIAS.

Also add a note for LD_HWCAP_MASK that since glibc 2.26 the environment
variable might be ignored if glibc does not support tunables.

Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

man*/: ffix: Follow rules for lists from man-pages(7)

Including in commented out lists, so that grep(1) doesn't show odd cases.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: Document new UFFDIO_POISON ioctl

This is a new feature recently added to the kernel.  So, document the
new ioctl the same way we do other UFFDIO_* ioctls.

Also note the corresponding new ioctl flag we can return in response to
a UFFDIO_REGISTER call.

Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>
Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: srcfix

Fixes: a252b3345f5b ("ioctl_userfaultfd.2: Describe two-step feature handshake")
Reported-by: groff(1) (`make build-catman`)
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: Clarify the state of the uffdio_api structure on error

The old FIXME noted that the zeroing was done to differentiate the two
EINVAL cases.  It's possible something like this was true historically,
but in current Linux we zero it in *both* EINVAL cases, so this is at
least no longer true.

After reading the code, I can't determine any clear reason why we zero
it in some cases but not in others.  So, some simple advice we can give
userspace is: if an error occurs, treat the contents of the structure as
unspecified. Just re-initialize it before retrying UFFDIO_API again.

Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>
Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: srcfix

Reported-by: groff(1) (`make build-catman`)
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: ffix

Signed-off-by: Alejandro Colomar <alx@kernel.org>

CONTRIBUTING: Clarify that subscribing to the mailing list is not necessary

Signed-off-by: Alejandro Colomar <alx@kernel.org>

set_thread_area.2: Document C-SKY arch

C-SKY only needs set_thread_area(), no need for get_thread_area(), the
same as MIPS.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
Link: <https://lore.kernel.org/linux-man/014e670b-2a11-3deb-4e7e-bfe9defd8597@gmail.com/T/>
Cc: Arnd Bergmann <arnd@arndb.de>
[alx: ffix; tweak commit message too]
Signed-off-by: Alejandro Colomar <alx@kernel.org>

fanotify.7: Fix field name in struct fanotify_event_info_fid

The final field of this struct as defined in include/uapi/linux/fanotify.h
(and as used in the example program) is called handle, not file_handle.
Fix this, and also replace the zero-length array with a flexible array
member to match the definition in the header.

Fixes: 0a4db6dc742d ("fanotify.7, fanotify_init.2, fanotify_mark.2: Document FAN_REPORT_FID and directory modification events")
Signed-off-by: Matthew House <mattlloydhouse@gmail.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: Add host attributes

Add in some host level sysfs entries:
enable_stream_detection - Send RAID 5/6 writes down the normal I/O path.
                          Allow for full-stripe write requests.
ssd_smart_path_enabled  - Indicates that our Accelerated I/O (AIO) path
                          is enabled (AIO).
enable_r5_writes        - Allow RAID 5 write operations down the AIO path.
enable_r6_writes        - Allow RAID 6 write operations down the AIO path.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: ffix

Signed-off-by: Alejandro Colomar <alx@kernel.org>

prctl.2: Document PR_SET_MDWE and PR_GET_MDWE

Memory-Deny-Write-Execute is a W^X process control originally introduced
by Joey Gouly.  I'm the author of the PR_MDWE_NO_INHERIT flag.

Signed-off-by: Florent Revest <revest@chromium.org>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

execve.2: execve() also returns E2BIG if a string is too long

The execve syscall returns -E2BIG in 3 cases:
-  The total length of the command line arguments and environment is too
   large.
-  An argument or environment string (including the NUL byte) is longer
   than MAX_ARG_STRLEN.
-  The full path to the executable (including the NUL byte) exceeds
   MAX_ARG_STRLEN.

Spell out all 3 cases in the -E2BIG section.

Discovered by moving a too large commandline parameter to an environment
variable, and finding that things still did not work. Examined the code
in fs/exec.c to get the details.

This simple shell script starts failing at 2^17 on a system with 4kB
page size:

#!/bin/sh
STRING="a"

for loop in `seq 20`; do
STRING="$STRING$STRING"
export STRING
if /bin/true ; then
: # still under the limit
else
echo "fork failed at loop $loop"
fi
done

./exec2big.sh: line 10: /bin/true: Argument list too long
fork failed at loop 17

Suggested-by: Matthew House <mattlloydhouse@gmail.com>
Signed-off-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

string_copying.7: Clarify 'ustr' and 'zustr'

Suggested-by: "Serge E. Hallyn" <serge@hallyn.com>
Acked-by: "Serge E. Hallyn" <serge@hallyn.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: wfix

Fixes: 1f0f451c78fd ("ioctl_userfaultfd.2: Describe missing UFFDIO_API feature flags")
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Peter Xu <peterx@redhat.com>
Suggested-by: Mike Rapoport (IBM) <rppt@kernel.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

userfaultfd.2: ffix

Signed-off-by: "G. Branden Robinson" <branden@debian.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: FILES: wfix: Change "Logical drives" to "Disk drives" for /dev/sd*.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: Add module parameter ctrl_ready_timeout

Allow user to change the length of time that the driver
will wait for the controller to become ready.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: srcfix: Use one line per identifier

Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: ffix: Use \[lq] and \[rq]

Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: Correct and update UFFDIO_API ioctl error codes

First, it is not correct that repeated UFFDIO_API calls result in
EINVAL.  This is true *if both calls enable features*, but in the case
where we're doing a two-step feature detection handshake, the kernel
explicitly expects 2 calls (one with no features set).  So, correct this
description.

Then, some new error cases have been added to the kernel recently, and
the man page wasn't updated to note these.  So, add in descriptions of
these new error cases.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Cc: Peter Xu <peterx@redhat.com>
Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: Describe two-step feature handshake

Fully describe how UFFDIO_API can be used to perform a two-step feature
handshake, and also note the case where this isn't necessary (programs
which don't depend on any extra features).

This lets us clean up an old FIXME asking for this to be described.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Cc: Peter Xu <peterx@redhat.com>
Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_key_create.3: Mention glibc instead of LinuxThreads

glibc defines PTHREAD_DESTRUCTOR_ITERATIONS in <limits.h>.  This means
it must not loop forever.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_*.3: ffix (semantic newlines)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_*.3: ffix (paragraphing)

Replace blank lines by either a no-op '.', the dummy character '\&', or
an appropriate paragraphing macro.  I've experimented with two '.' lines
before section headings, which may increase readability.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_*.3: Remove AUTHOR section; add copyright; adapt TH

Xavier asked that his name be removed from the AUTHOR section, and since
we don't use that section in this project, we can just wipe it.

Add copyright notices to the pages, with Xavier as the original author,
and me, who will maintain them from now on.  Xavier ACKed the
Linux-man-pages-copyleft license.

Adapt the TH line:

-  Specify the date as (date).
-  Specify the project name in the 4th arg.
-  The first arg, the page name, must be the same as the file name.

Suggested-by: Xavier Leroy <Xavier.Leroy@inria.fr>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_cond_init.3, pthread_condattr_init.3, pthread_key_create.3, pthread_mutex_init.3, pthread_mutexattr_setkind_np.3, pthread_once.3: Update the glibc pages with the debian/glibc version of them

Signed-off-by: Alejandro Colomar <alx@kernel.org>

Import debian/local/manpages/pthread_*.3 git history from debian/glibc

I extracted the same pages that we imported from glibc, from the Debian
package history.  They were imported from glibc to Debian, so we will be
able to pick those updates too.  Below is the process.

$ git clone git@salsa.debian.org:glibc-team/glibc.git
$ cd glibc/
$ git filter-repo --path debian/local/manpages/
$ git filter-repo \
--path debian/local/manpages/pthread_cond_init.3 \
--path debian/local/manpages/pthread_condattr_init.3 \
--path debian/local/manpages/pthread_key_create.3 \
--path debian/local/manpages/pthread_mutex_init.3 \
--path debian/local/manpages/pthread_mutexattr_setkind_np.3 \
--path debian/local/manpages/pthread_once.3
$ git tag | xargs git tag -d
$ git branch | xargs git branch -D
$ git remote add man file:///home/alx/src/linux/man-pages/man-pages/.bare.git
$ git fetch man
$ git rebase man/glibc

Subsequent commits will move the pages to the appropriate directory.

Link: <https://lore.kernel.org/linux-man/93aba0c9-ad1d-d488-1471-be30b7a8bb0c@gmail.com/T/>
Acked-by: Xavier Leroy <Xavier.Leroy@inria.fr>
Cc: Debian GNU Libc Maintainers <debian-glibc@lists.debian.org>
Cc: glibc <libc-alpha@sourceware.org>
Cc: Carlos O'Donell <carlos@redhat.com>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Reported-by: Arnav Rawat <rawat.arnav@gmail.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_cond_init.3, pthread_condattr_init.3, pthread_key_create.3, pthread_mutex_init.3, pthread_mutexattr_setkind_np.3, pthread_once.3: Import pages from glibc

Signed-off-by: Alejandro Colomar <alx@kernel.org>

  * Remove linuxthreads from the tarball:
    - rules.d/tarball.mk: don't fetech linuxthreads and linuxthreads_db.
    - rules.d/build.mk: don't build linuxthreads manpages.
    - rules: don't run make clean in linuxthreads directory.
    - patches/any/local-sysctl.diff: drop the linuxthreads part.
    - patches/all/local-pthread-manpages.diff: remove.
    - local/manpages/pthread_*.3: import the few remaining linuxthreads
      manpages.
    - debhelper.in/glibc-doc.manpages: update manpage locations.

svn path=/glibc-package/branches/eglibc-2.18/; revision=5804

Revert "linuxthreads, linuxthreads_db: Directories removed (preserved in ports repository)."

This reverts commit a3db24d46ceccad34e8e70729596e1b2a2d22f5e.

Import linuxthreads/man/ git history from glibc

I extracted the linuxthreads/man/ subdirectory from the glibc git
repository.  I also removed from that subdir the pages that we already
have in this project.  Then rebased on top of the Linux man-pages HEAD.
Below is the process.

$ git clone https://sourceware.org/git/glibc.git/
$ cd glibc/
$ git filter-repo --path linuxthreads/man/
$ git filter-repo --invert-paths \
--path linuxthreads/man/Makefile \
--path linuxthreads/man/pthread_atfork.man \
--path linuxthreads/man/pthread_attr_init.man \
--path linuxthreads/man/pthread_cancel.man \
--path linuxthreads/man/pthread_cleanup_push.man \
--path linuxthreads/man/pthread_create.man \
--path linuxthreads/man/pthread_detach.man \
--path linuxthreads/man/pthread_equal.man \
--path linuxthreads/man/pthread_exit.man \
--path linuxthreads/man/pthread_join.man \
--path linuxthreads/man/pthread_kill_other_threads_np.man \
--path linuxthreads/man/pthread_mutexattr_init.man \
--path linuxthreads/man/pthread_self.man \
--path linuxthreads/man/pthread_setschedparam.man \
--path linuxthreads/man/pthread_sigmask.man \
--path linuxthreads/man/sem_init.man \
--path linuxthreads/man/troffprepro
$ git tag | xargs git tag -d
$ git branch | xargs git branch -D
$ git remote add man file:///home/alx/src/linux/man-pages/man-pages/.bare.git
$ git fetch man
$ git rebase man/master

Subsequent commits will revert the commit that removed the pages, and
will move them to the appropriate directories.

Link: <https://lore.kernel.org/linux-man/93aba0c9-ad1d-d488-1471-be30b7a8bb0c@gmail.com/T/>
Acked-by: Xavier Leroy <Xavier.Leroy@inria.fr>
Cc: Debian GNU Libc Maintainers <debian-glibc@lists.debian.org>
Cc: glibc <libc-alpha@sourceware.org>
Cc: Carlos O'Donell <carlos@redhat.com>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Reported-by: Arnav Rawat <rawat.arnav@gmail.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

linuxthreads, linuxthreads_db: Directories removed (preserved in ports repository).

(CFLAGS-tst-align.c): Add -mpreferred-stack-boundary=4.

2.5-18.1

Update.

2002-07-28  Ulrich Drepper  <drepper@redhat.com>

* po/fi.po: New file.

Update.

2002-03-23  Jakub Jelinek  <jakub@redhat.com>

* sysdeps/generic/brk.c (__curbrk): Declare.
* sysdeps/generic/dl-brk.c: Add attribute_hidden to __curbrk.
* sysdeps/generic/dl-sbrk.c: Likewise.
* sysdeps/unix/arm/dl-brk.S: New file.
* sysdeps/unix/bsd/hp/m68k/dl-brk.S: New file.
* sysdeps/unix/bsd/osf/alpha/dl-brk.S: New file.
* sysdeps/unix/bsd/sun/m68k/dl-brk.S: New file.
* sysdeps/unix/bsd/vax/dl-brk.S: New file.
* sysdeps/unix/i386/dl-brk.S: New file.
* sysdeps/unix/mips/dl-brk.S: New file.
* sysdeps/unix/sparc/dl-brk.S: New file.
* sysdeps/unix/sysv/linux/alpha/dl-brk.S: New file.
* sysdeps/unix/sysv/linux/ia64/dl-brk.S: New file.
* sysdeps/unix/sysv/linux/powerpc/dl-brk.S: New file.
* sysdeps/unix/sysv/linux/sparc/sparc64/dl-brk.S: New file.
* sysdeps/unix/sysv/linux/i386/dl-brk.c: Remove.
* sysdeps/unix/sysv/linux/i386/dl-sbrk.c: Remove.

2002-04-03  Andreas Schwab  <schwab@suse.de>

* Makefile (headers): Add gnu/lib-names.h here instead of
install-others.
($(inst_includedir)/gnu/lib-names.h): Remove explicit
installation rule.
(install-headers): Add dependency on install-headers-nosubdir.
* stdio-common/Makefile (headers): Add bits/stdio_lim.h here
instead of install-others.
($(inst_includedir)/bits/stdio_lim.h): Remove explicit
installation rule.

2002-04-05  Ulrich Drepper  <drepper@redhat.com>

* manual/users.tex (Enable/Disable Setuid): Fix typo in example.
Reported by Sam Roberts <sroberts@uniserve.com>.

2002-04-03  Jakub Jelinek  <jakub@redhat.com>

* elf/do-rel.h (elf_dynamic_do_rel): Skip relative relocs if
l_addr == 0 and ELF_MACHINE_REL_RELATIVE.
* sysdeps/alpha/dl-machine.h (ELF_MACHINE_REL_RELATIVE): Define.
* sysdeps/ia64/dl-machine.h (ELF_MACHINE_REL_RELATIVE): Define.

2002-04-03  David Mosberger  <davidm@hpl.hp.com>

* sysdeps/ia64/dl-machine.h (TRAMPOLINE_TEMPLATE): Add unwind info.
(RTLD_START): Ditto.
(__ia64_init_bootstrap_fdesc_table): Insert stop bit to avoid RAW
dependency violation.

Correct example.

LinuxThreads library.

1998-03-11 00:42  Wolfram Gloger  <wmglo@dent.med.uni-muenchen.de>
* linuxthreads/manager.c: Enable resetting of the thread
scheduling policy to SCHED_OTHER when the parent thread
has a different one.
1998-02-01 13:51  Ulrich Drepper  <drepper@cygnus.com>
* sysdeps/unix/sysv/linux/bits/posix_opt.h: Define
_POSIX_ASYNCHRONOUS_IO.
* sysdeps/pthread/pthread.h: Define bits for Unix98 variants of
mutexes.
* mutex.c: Implement new mutex types.
* internals.h: Include <signal.h>.
* libpthread.map: Add __erno_location and __h_errno_location.
* errno.c: Return pointer to variable actually in use.  This might
not be the one in the thread structure.
* internals.h (struct _pthread_descr_struct): Add new fields p_errnop
and p_h_errnop.
* manager.c (__pthread_manager): Set p_errnop and p_h_errnop member
of manager thread structure.
(pthread_handle_create): Set p_errnop and p_h_errnop members for new
thread.
* pthread.c: Adapt initializer for thread structures.
(__pthread_initial_thread): Set p_errnop and p_h_errnop member.
(__pthread_reset_main_thread): Reset p_errnop and p_h_errnop of
current thread to global variables.
1998-01-31 17:27  Ulrich Drepper  <drepper@cygnus.com>
* rwlock.c: New file.
* Makefile (libpthread-routines): Add rwlock.
* sysdeps/pthread/pthread.h: Define data structures and declare
functions.
* libpthread.map: Add new functions.
1997-12-18 13:50  Philip Blundell  <pb@nexus.co.uk>
* sysdeps/arm/pt-machine.h: New file; add ARM support.
* sysdeps/arm/Implies: likewise.
* README: Document it.
1997-12-13  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>
* signals.c: Remove unneeded initializer for sigwaited, saving a
1997-04-11 01:18  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>
* semaphore.c (sem_init): Set sem_spinlock only if available.
1997-12-04 01:48  Ulrich Drepper  <drepper@cygnus.com>
* mutex.c: Implement PTHREAD_MUTEX_CHECKERROR.
* sysdeps/pthread/pthread.h: Define PTHREAD_MUTEX_CHECKERROR.
* Makefile: Update from LinuxThreads 0.7.
* internals.h. Likewise.
* manager.c: Likewise.
* mutex.c: Likewise.
* pthread.c: Likewise.
* signals.c: Likewise.
* specific.c: Likewise.
* Examples/ex3.c: Likewise.
1997-11-20 18:13  Ulrich Drepper  <drepper@cygnus.com>
* pthread.c (__pthread_reset_main_thread): Close pipe only if still
open.
1997-10-29 05:38  Ulrich Drepper  <drepper@cygnus.com>

* wrapsyscall.c: Add socket functions which are also cancelation
points.

1997-10-19 21:40  Wolfram Gloger  <wg@wolfram.dent.med.uni-muenchen.de>

* specific.c (__libc_internal_tsd_set, __libc_internal_tsd_get):
New functions for fast thread specific data within libc.

* internals.h: Add new array p_libc_specific to struct
_pthread_descr_struct.

* sysdeps/pthread/bits/libc-lock.h: Declare new functions.

1997-10-13 05:39  Ulrich Drepper  <drepper@cygnus.com>

* semaphore.h: Add __BEGIN_DECLS/__END_DECLS.
Reported by Ralf Corsepius <corsepiu@faw.uni-ulm.de>.

1997-08-29 03:05  Ulrich Drepper  <drepper@cygnus.com>

* internals.h (struct _pthread_descr_struct): Add definitions for
two-level specific key handling.
* manager.c (pthread_handle_create): Initialize specific memory array.
* specific.c: Implement two-level key handling.
* weaks.c: Don't provide dummy key handling.
* sysdeps/pthread/bits/libc-lock.h: Typedef __libc_lock_t (no #define).
Add definition of __libc_key_t.
* sysdeps/unix/sysv/linux/bits/local_lim.h: Define PTHREAD_KEYS_MAX
as 1024.
Add definition of _POSIX_THREAD_DESTRUCTOR_ITERATIONS and
PTHREAD_DESTRUCTOR_ITERATIONS.

* manager.c (pthread_handle_create): Compare mmap result with
MAP_FAILED.

* ptfork.c: Rename to __pthread_atfork and make old name a weak alias.
* sysdeps/pthread/bits/pthread.h: Add prototype for __pthread_atfork.

1997-08-22 19:04  Richard Henderson  <rth@cygnus.com>

sysdeps/sparc -> sysdeps/sparc/sparc32
sysdeps/sparc64 -> sysdeps/sparc/sparc64

* internals.h: Change definition of THREAD_SELF to be an expression,
not a statement that did a return.
* sysdeps/alpha/pt-machine.h (THREAD_SELF): Update accordingly.
* sysdeps/sparc/sparc32/pt-machine.h (THREAD_SELF, INIT_THREAD_SELF):
Follow Solaris and use a "system reserved" register (%g6) to hold
the thread descriptor.
* sysdeps/sparc/sparc64/pt-machine.h: Likewise.

1997-08-03 00:09  Ulrich Drepper  <drepper@cygnus.com>

* mutex.c: Correct pthread_once.  Patch by Xavier Leroy.
* sysdeps/pthread/pthread.h: Add prototype for __pthread_once.
* sysdeps/pthread/bits/pthread.h: Add macros for __libc_once.

* semaphore.c: Include spinlock.h only when needed.

* specific.c (__pthread_setsepcific, __pthread_getspecific): Reject
keys for entries not in use.

* weaks.c: Implement key handling functions for real.

1997-06-29  01:04  Richard Henderson  <richard@gnu.ai.mit.edu>

Initial sparc64-linux support:
* linuxthreads/sysdeps/sparc64/Implies: New file.
* linuxthreads/sysdeps/sparc64/pt-machine.h: Likewise.

1997-06-29 00:48  Ulrich Drepper  <drepper@cygnus.com>

* semaphore.c: Include spinlock.h at correct place.
Patch by HJ Lu.

1997-06-13 10:06  Richard Henderson  <rth@tamu.edu>

The Great Bit File Move:
* sysdeps/alpha/semaphorebits.h: -> .../bits/semaphore.h.
* sysdeps/powerpc/semaphorebits.h: Likewise.
* sysdeps/pthread/cmpxchg/semaphorebits.h: Likewise.
* sysdeps/pthread/no-cmpxchg/semaphorebits.h: Likewise.
* sysdeps/pthread/libc-lock.h: -> bits/
* sysdeps/pthread/stdio-lock.h: Likewise.
* sysdeps/unix/sysv/linux/local_lim.h: Likewise.
* sysdeps/unix/sysv/linux/posix_opt.h: Likewise.
* semaphore.h: Likewise.
* sysdeps/pthread/pthread.h: Likewise.

* lockfile.c: <foo.h> -> <bits/foo.h>.
* semaphore.h: Likewise.

* Makefile: (headers): foo.h -> bits/foo.h.
* sysdeps/pthread/Makefile: Likewise.

1997-04-11 01:18  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

* semaphore.c (sem_init): Set sem_spinlock only if available.

* sysdeps/m68k/pt-machine.h (testandset, __compare_and_swap): Fix
asm constraints.

1997-04-09 03:00  Ulrich Drepper  <drepper@cygnus.com>

Update from LinuxThreads 0.6.

* attr.c (pthread_attr_getdetachstate): Use __sched_get_priority_max
and __sched_get_priority_min instead of names without `__'.

* manager.c: Rewrite large parts to implement opaque pthread_t.

* cancel.c: Adapt for opaque pthread_t type.
* condvar.c: Likewise.
* errno.c: Likewise.
* join.c: Likewise.
* mutex.c: Likewise.
* pthread.c: Likewise.
* signals.c: Likewise.
* specific.c: Likewise.
* restart.h: Likewise.
* queue.h: Likewise.
* Examples/ex3.c: Likewise.
* Examples/ex4.c: Likewise.
* sysdeps/pthread/pthread.h: Likewise.

* pthread.c: Accumulate time for all threads in thread manager.

* semaphore.c: Implement fallback implementation for architectures
sometimes missing compare-exchange operations.

* cancel.c (pthread_cancel): Validate handle argument.
* join.c (pthread_join): Likewise.
(pthread_detach): Likewise.
* signals.c (pthread_kill): Likewise.

* spinlock.h (acquire): Use __sched_yield not sched_yield.

* queue.h (enqueue): Enqueue thread according to priority.

* internals.c (struct pthread_start_args): New struct for passing
args to cloning function.
(struct _pthread): Rename to _pthread_descr_struct and adapt for
opaque pthread_t.

* Examples/Makefile (clean): Pass -f option to rm.

* sysdeps/i386/pt-machine.h: Add check for compare-exchange instruction
and define TEST_FOR_COMPARE_AND_SWAP.
* sysdeps/i386/i486/pt-machine.h: Removed.

* sysdeps/unix/sysv/linux/local_lim.h (PTHREAD_THREADS_MAX): Increase
to 1024.

1997-04-04 16:38  Ulrich Drepper  <drepper@cygnus.com>

* restart.h (suspend): Clear p_signal before suspending.
(suspend_with_cancellation): Likewise.
Patch by Xavier Leroy <Xavier.Leroy@inria.fr>.

* weaks.c: Make __pthread_key_create return 1.
* sysdeps/pthread/libc-lock.h: Define __libc_key_create,
__libc_getspecific, __libc_setspecific, and __libc_key_t.
* sysdeps/pthread/stdio-lock.h: Don't care for implementation not
using libio.

1997-03-19 15:13  Miguel de Icaza  <miguel@nuclecu.unam.mx>

* sysdeps/sparc/pt-machine (RELEASE): Fix.

1997-03-01 07:55  Geoff Keating  <geoffk@ozemail.com.au>

* sysdeps/powerpc/Implies: Added.
* sysdeps/powerpc/pt-machine.h: Added.
* sysdeps/powerpc/semaphorebits.h: Added.

1997-01-22 01:22  Ulrich Drepper  <drepper@cygnus.com>

* linuxtheads/pthread.c (__pthread_initial_thread): Correct
initializer.
(__pthread_manager_thread): Likewise.
Reported by Andreas Jaeger.

1997-01-18 22:15  Richard Henderson  <rth@tamu.edu>

Since sigset_t no longer fits in a register, we can't pass in the
thread's initial mask so easily.  Take this opportunity to simplify
the clone implementation by only accepting a single void* argument.

* linuxthreads/manager.c (__pthread_manager): Put thread vitals
in the thread struct instead of as arguments through clone.
(pthread_start_thread): Look for them there.
* linuxthreads/internals.h (struct _pthread): Add p_initial_fn,
p_initial_fn_arg, p_initial_mask.  Fix __pthread_manager proto.
* linuxthreads/pthread.c (pthread_initialize_manager): Revise
clone invocation.

proc_sys.5, proc_sys_vm.5: Split /proc/sys/vm/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_user.5: Split /proc/sys/user/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_sunrpc.5: Split /proc/sys/proc/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_proc.5: Split /proc/sys/proc/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_net.5: Split /proc/sys/net/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_kernel.5: Split /proc/sys/kernel/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_fs.5: Split /proc/sys/fs/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_dev.5: Split /proc/sys/dev/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_debug.5: Split /proc/sys/debug/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_sys.5, proc_sys_abi.5: Split /proc/sys/abi/ from proc_sys(5)

Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: sas_address contains the SAS address

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: wfix

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: SEE ALSO: Add sg(4) entry

The sg driver can interface with smartpqi.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: Add module parameter disable_managed_interrupts

Add in a description for module parameter disable_managed_interrupts.

This module parameter allows users to disable automatic irq affinity
setting by turning off the PCI_IRQ_AFFINITY flag when requesting
interrupts.

When the flag PCI_IRQ_AFFINITY is set, users cannot change the interrupt
affinity through /proc/irq/<irq#>/*affinity*

Some customers would like to change interrupt affinity based on their
requirements.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

printf.3: Fix reference to CAVEATS (previously NOTES)

Fixes: 4131356cd ("man*/, man-pages.7: VERSIONS, STANDARDS, HISTORY: Reorganize sections")
Reported-by: Дилян Палаузов <dilyan.palauzov@aegee.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: Update module parameter descriptions

Minor changes to the descriptions of the following module parameters:
disable_device_id_wildcards
disable_heartbeat
disable_ctrl_shutdown
expose_ld_first
hide_vsep

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

smartpqi.4: Update copyright, email addresses, and URLs

Add Copyright entry for Microchip.
Change e-mail addresses from Microsemi to Microchip.
Change URL to Microchip.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

tmpfs.5: Extend with new noswap documentation

Linux commit 2c6efe9cf2d7 ("shmem: add support to ignore swap")
merged as of v6.4 added support to disable swap for tmpfs mounts.

This extends the man page to document that.

Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

system_data_types.7: Clarify si_addr field.

This patch reuses the clearer text from sigaction.2 instead.

I've also filed https://www.austingroupbugs.net/view.php?id=1775
to improve POSIX's wording here to not be SIGILL-specific.

Signed-off-by: Elliott Hughes <enh@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: Describe missing UFFDIO_API feature flags

Several new features have been added to the kernel recently, and the man
page wasn't updated to describe these new features.  So, add in
descriptions of any missing features.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_userfaultfd.2: ffix + wfix

-  Fix missing paragraph tag.  The lack of this tag yielded no blank
   line in the rendere dpage, which is inconsistent with style
   elsewhere.

-  The description of UFFDIO_WRITEPROTECT was a sentence fragment; the
   last half of the sentence was left out by mistake.  Add it in to fix
   the issue.

-  Move UFFDIO_WRITEPROTECT 'since' to its own line.  All other ioctls
   note the kernel version introduced on a separate line from the ioctl
   name.  Update UFFDIO_WRITEPROTECT to match the existing style.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

userfaultfd.2: Comment on feature detection in the example program

The example program doesn't depend on any extra features, so it does not
make use of the two-step feature handshake process. This is fine, but it
might set a bad example for programs which *do* depend on specific
features (e.g. they may conclude they don't need to do anything to
enable / detect them).

No need to make the example program more complicated: let's just add a
comment indicating why we do it the way we do it in the example, and
describing briefly what a more complicated program would need to do
instead.

The comment is kept rather brief; a full description of this feature
will be included in ioctl_userfaultfd.2 instead.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

userfaultfd.2: Reword to account for new fault resolution ioctls

Basically, reword the sentence to clarify that this isn't a complete
list.  I don't believe it's worth maintaining a fully complete list
here, instead ioctl_userfaultfd.2 is the place to do that. Let this just
be an example.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

userfaultfd.2: Briefly mention two-step feature handshake process

This process is critical for programs which depend on extra features, so
it's worth mentioning here.

Future commits will much more fully describe it in ioctl_userfaultfd.2.

Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

move_pages.2: Relicense to Linux-man-pages-copyleft

Christoph Lameter gave me explicit permission to relicense it for
consistency with the rest of the project.

Michael repeatedly stated that he wanted to use Linux-man-pages-copyleft
to reduce the number of licenses we have in the project.

Cc: Christopher Lameter <cl@os.amperecomputing.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

man*/: Update references to sigevent(7 => 3type)

Acked-by: Elliott Hughes <enh@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

sigevent.3type, sigevent.7: Merge sigevent(7) into sigevent(3type)

Keep the old page, sigevent(7), as a link page to the one in 3type.

Acked-by: Elliott Hughes <enh@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

sigevent.3type, sigval.3type, system_data_types.7: Move sigval to sigevent.3type

Acked-by: Elliott Hughes <enh@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

sigevent.3type, system_data_types.7: Move sigevent to its own page

Acked-by: Elliott Hughes <enh@google.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl_tty.2: Document CAP_SYS_ADMIN requirement for TIOCSTI

Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Günther Noack <gnoack@google.com>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ioctl.2: Note "int request" form, HISTORYise a bit

Bit me in <https://github.com/thecoshman/http/issues/155>:
musl (and, thus, bionic) takes an int!

Kill the "POSIX says 'request' is int" comment;
prior to Issue 7, ioctl() is shaded STREAMS,
it's obsolescent in Issue 7,
and gone completely in Issue 8.
The POSIX interface has never been supported by Linux,
and the common ioctl() interface has never existed in POSIX or the SUS.

Note how the interface evolved in HISTORY. The NetBSD CVS has
  revision 1.1
  date: 1993-03-21 10:45:37 +0100;  author: cgd;  state: Exp;
  branches:  1.1.1;
  Initial revision
  ----------------------------
  revision 1.1.1.2
  date: 1995-02-27 11:49:43 +0100;  author: cgd;  state: Exp;  lines: +6
  -6;
  from Lite
  ----------------------------
  revision 1.1.1.1
  date: 1993-03-21 10:45:37 +0100;  author: cgd;  state: Exp;  lines: +0
  -0;
  initial import of 386bsd-0.1 sources
  ----------------------------
and it's char * in 1.1.1.1 and 1.1.1.2 but ... in 1.1, so hell knows.

Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Cc: Jakub Wilk <jwilk@jwilk.net>
[alx: ffix]
Signed-off-by: Alejandro Colomar <alx@kernel.org>

popen.3: wfix

Signed-off-by: Tom Schwindl <schwindl@posteo.de>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

fmod.3: Add example section

Signed-off-by: Jan Engelhardt <jengelh@inai.de>
[alx: ffix]
Signed-off-by: Alejandro Colomar <alx@kernel.org>

fmod.3: Add note how to obtain modulus/LPR

Signed-off-by: Jan Engelhardt <jengelh@inai.de>
[alx: ffix]
Signed-off-by: Alejandro Colomar <alx@kernel.org>

ptrdiff_t.3type: tfix: Fix limits of ptrdiff_t

Signed-off-by: Alejandro Colomar <alx@kernel.org>

name_to_handle_at.2,fanotify_mark.2: Document the AT_HANDLE_FID flag

A flag to indicate that the requested file_handle is not intended
to be used for open_by_handle_at(2) and may be needed to identify
filesystem objects reported in fanotify events.

Reviewed-by: Jan Kara <jack@suse.cz>
Acked-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Cc: Tom Schwindl <schwindl@posteo.de>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

mmap.2: Fix missing text

Fixes: 090fdddb43 ("memfd_create.2, mmap.2, shmget.2: Document the EPERM for huge page allocations")
Signed-off-by: Avinesh Kumar <akumar@suse.de>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

inode.7: tfix: Correct the field name for number of blocks

Fixes: b48c75727c64 ("inode.7: New page with information about inodes")
Signed-off-by: Avinesh Kumar <akumar@suse.de>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

inode.7: Remove duplicate field name

Fixes: b48c75727c64 ("inode.7: New page with information about inodes")
Signed-off-by: Avinesh Kumar <akumar@suse.de>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

clang-tidy/config.yaml: Silence some warnings

Signed-off-by: Alejandro Colomar <alx@kernel.org>

man5/: srcfix: '\" t comment

Signed-off-by: Alejandro Colomar <alx@kernel.org>

splice.2: srcfix

Signed-off-by: Alejandro Colomar <alx@kernel.org>

pthread_setschedparam.3: EXAMPLES: Add missing [[noreturn]]

Signed-off-by: Alejandro Colomar <alx@kernel.org>

getutent.3: EXAMPLES: Fix use of time(3)

time(ptr) was wrong here, because struct utmp::ut_time is a int32_t, not
a time_t.  Also, the use of time(ptr) is obsolete, and time(NULL) is
always preferred.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

getaddrinfo_a.3: EXAMPLES: Use a safer wrapper around *alloc*(3) functions

These wrappers result in cleaner code, and prevent many mistakes that
can happen when using malloc(3), calloc(3), realloc(3), and functions of
that family.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

Revert "*.mk: DEFAULT_CPPFLAGS: -D_GNU_SOURCE"

This reverts commit bb42ee0d7793cf0fe2d2ec36c99891227797902c.

Programs should specify _GNU_SOURCE if they need it.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: tfix

Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: Have the makefiles as dependencies of all files

If any makefile is modified, all files should be remade.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

proc_self.5: Fix link page

Fixes: 68553d3f22ff ("proc.5, proc_pid.5, proc_self.5: Split /proc/PID/ (and /proc/self/) from proc(5)")
Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: SHELL, .SHELLFLAGS: Fix the SHELL

-  Arguments to the shell need to go in .SHELLFLAGS.
-  env(1) needs -S to split arguments to it.
   Maybe it's not necessary in make(1), though; not sure.
-  The scripts are passed to the shell as arguments, not via stdin;
   thus, we need -c.

Cc: Paul Smith <psmith@gnu.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: dist-*: Use a function to define similar rules

Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: GZIPFLAGS: Add -n

It was being hardocded in a recipe, but we shouldn't do that.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

strtol.3: EXAMPLES: Remove unused include

Fixes: 93f369892aea ("strtol.3: EXAMPLES: Simplify errno checking")
Reported-by: iwyu(1) (`make lint-c-iwyu`)
Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: lint-c-iwyu: Fail iff iwyu(1) finds problems

That is, ignore any warnings from clang(1).  We have a different target
for that.

Signed-off-by: Alejandro Colomar <alx@kernel.org>

*.mk: DEFAULT_CPPFLAGS: -D_GNU_SOURCE

To get some definitions that are essential in our programs.

Signed-off-by: Alejandro Colomar <alx@kernel.org>