[PATCH 00/10] man2: SYNOPSIS: Use syscall, and fix includes.

[PATCH 00/10] man2: SYNOPSIS: Use syscall, and fix includes.

[Alejandro Colomar]

Hi Michael,

Almost finished these:
$ man_lsfunc man2 | sed -n '/socketcall/,$p' | wc -l
67
$ man_lsfunc man2 | wc -l
401
$

A remainder for myself of commands needed to do this:

.../man-pages$ man_section man2 SYNOPSIS 'CONFORMING TO' NOTES \
               | sed -n '/SOCKETCALL/,$p' \
               | sponge | less;


.../glibc$ man_lsfunc ../../linux/man-pages/man2 \
           | while read -r syscall; do \
                   echo "=============================  ${syscall}"; \
                   grep_glibc_prototype ${syscall}; \
           done \
           | sed -e 's/\bextern //' -e 's/\b_*//g' \
           | sed -n '/socketcall/,$p' \
           | sponge | less;

Kind regards,

Alex


Alejandro Colomar (10):
  open.2: Remove unused <sys/stat.h>
  rt_sigqueueinfo.2: Use syscall(SYS_...); for system calls without a
    wrapper
  seccomp.2: Document why each header is needed
	(not sending this one; already sent last week)
	<https://lore.kernel.org/linux-man/20210515182254.186607-1-alx.manpages@gmail.com/T/#u>
  seccomp.2: Use syscall(SYS_...); for system calls without a wrapper
	(depends on the previous one)
  set_thread_area.2: Use syscall(SYS_...); for system calls without a
    wrapper
  set_tid_address.2: Use syscall(SYS_...); for system calls without a
    wrapper
  sgetmask.2: Use syscall(SYS_...); for system calls without a wrapper
  shmop.2: Remove unused include
  sigprocmask.2: Use syscall(SYS_...); for raw system calls
  socketcall.2: Use syscall(SYS_...); for system calls without a wrapper

 man2/open.2            |  1 -
 man2/rt_sigqueueinfo.2 | 11 ++++++++---
 man2/seccomp.2         | 24 +++++++++++++-----------
 man2/set_thread_area.2 | 24 ++++++++++++------------
 man2/set_tid_address.2 | 13 +++++++------
 man2/sgetmask.2        | 15 ++++++++-------
 man2/shmop.2           |  1 -
 man2/sigprocmask.2     |  8 ++++++--
 man2/socketcall.2      | 15 ++++++++-------
 9 files changed, 62 insertions(+), 50 deletions(-)

-- 
2.31.1

[PATCH 01/10] open.2: Remove unused <sys/stat.h>

[Alejandro Colomar]

I can't see a reason to include it.  <fcntl.h> provides O_*
constants for 'flags', S_* constants for 'mode', and mode_t.

Probably a long time ago, some of those weren't defined in
<fcntl.h>, and both headers needed to be included, or maybe it's
a historical error.

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/open.2 | 1 -
 1 file changed, 1 deletion(-)

diff --git a/man2/open.2 b/man2/open.2
index bac1ee806..1bcd995f4 100644
--- a/man2/open.2
+++ b/man2/open.2
@@ -53,7 +53,6 @@
 open, openat, creat \- open and possibly create a file
 .SH SYNOPSIS
 .nf
-.B #include <sys/stat.h>
 .B #include <fcntl.h>
 .PP
 .BI "int open(const char *" pathname ", int " flags );
-- 
2.31.1

[PATCH 02/10] rt_sigqueueinfo.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/rt_sigqueueinfo.2 | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/man2/rt_sigqueueinfo.2 b/man2/rt_sigqueueinfo.2
index 2dd1a5877..d48ecbdcd 100644
--- a/man2/rt_sigqueueinfo.2
+++ b/man2/rt_sigqueueinfo.2
@@ -27,9 +27,14 @@
 rt_sigqueueinfo, rt_tgsigqueueinfo \- queue a signal and data
 .SH SYNOPSIS
 .nf
-.BI "int rt_sigqueueinfo(pid_t " tgid ", int " sig ", siginfo_t *" info );
-.BI "int rt_tgsigqueueinfo(pid_t " tgid ", pid_t " tid ", int " sig \
-", siginfo_t *" info );
+.BR "#include <linux/signal.h>" "     /* Definition of " SI_* " constants */"
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
+.PP
+.BI "int syscall(SYS_rt_sigqueueinfo, pid_t " tgid ,
+.BI "            int " sig ", siginfo_t *" info );
+.BI "int syscall(SYS_rt_tgsigqueueinfo, pid_t " tgid ", pid_t " tid ,
+.BI "            int " sig ", siginfo_t *" info );
 .fi
 .PP
 .IR Note :
-- 
2.31.1

[PATCH 04/10] seccomp.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/seccomp.2 | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/man2/seccomp.2 b/man2/seccomp.2
index 1be7d6766..476e5c2f0 100644
--- a/man2/seccomp.2
+++ b/man2/seccomp.2
@@ -37,13 +37,18 @@ seccomp \- operate on Secure Computing state of the process
 .BR "#include <sys/ptrace.h>" "     /* Definition of " PTRACE_* " constants */"
 .\" Kees Cook noted: Anything that uses SECCOMP_RET_TRACE returns will
 .\"                  need <sys/ptrace.h>
+.BR "#include <sys/syscall.h>" "    /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
 .PP
-.BI "int seccomp(unsigned int " operation ", unsigned int " flags \
-", void *" args );
+.BI "int syscall(SYS_seccomp, unsigned int " operation ", unsigned int " flags ,
+.BI "            void *" args );
 .fi
 .PP
 .IR Note :
-There is no glibc wrapper for this system call; see NOTES.
+glibc provides no wrapper for
+.BR seccomp (2),
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 The
 .BR seccomp ()
@@ -858,9 +863,6 @@ Tile (since Linux 4.3)
 PA-RISC (since Linux 4.6)
 .\" User mode Linux since Linux 4.6
 .PD
-.PP
-Glibc does not provide a wrapper for this system call; call it using
-.BR syscall (2).
 .\"
 .SS Caveats
 There are various subtleties to consider when applying seccomp filters
-- 
2.31.1

[PATCH 05/10] set_thread_area.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/set_thread_area.2 | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/man2/set_thread_area.2 b/man2/set_thread_area.2
index 284dea390..43f3009f7 100644
--- a/man2/set_thread_area.2
+++ b/man2/set_thread_area.2
@@ -11,28 +11,31 @@
 get_thread_area, set_thread_area \- manipulate thread-local storage information
 .SH SYNOPSIS
 .nf
-.B #include <linux/unistd.h>
+.BR "#include <sys/syscall.h>" "     /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
 .PP
 .B #if defined __i386__ || defined __x86_64__
-.B # include <asm/ldt.h>
+.BR "# include <asm/ldt.h>" "        /* Definition of " "struct user_desc" " */"
 .PP
-.BI "int get_thread_area(struct user_desc *" u_info );
-.BI "int set_thread_area(struct user_desc *" u_info );
+.BI "int syscall(SYS_get_thread_area, struct user_desc *" u_info );
+.BI "int syscall(SYS_set_thread_area, struct user_desc *" u_info );
 .PP
 .B #elif defined __m68k__
 .PP
-.B "int get_thread_area(void);"
-.BI "int set_thread_area(unsigned long " tp );
+.B "int syscall(SYS_get_thread_area);"
+.BI "int syscall(SYS_set_thread_area, unsigned long " tp );
 .PP
 .B #elif defined __mips__
 .PP
-.BI "int set_thread_area(unsigned long " addr );
+.BI "int syscall(SYS_set_thread_area, unsigned long " addr );
 .PP
 .B #endif
 .fi
 .PP
 .IR Note :
-There are no glibc wrappers for these system calls; see NOTES.
+glibc provides no wrappers for these system calls,
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 These calls provide architecture-specific support for a thread-local storage
 implementation.
@@ -172,10 +175,7 @@ and
 are Linux-specific and should not be used in programs that are intended
 to be portable.
 .SH NOTES
-Glibc does not provide wrappers for these system calls,
-since they are generally intended for use only by threading libraries.
-In the unlikely event that you want to call them directly, use
-.BR syscall (2).
+These system calls are generally intended for use only by threading libraries.
 .PP
 .BR arch_prctl (2)
 can interfere with
-- 
2.31.1

[PATCH 06/10] set_tid_address.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/set_tid_address.2 | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/man2/set_tid_address.2 b/man2/set_tid_address.2
index 258c9b9ef..4b81865dc 100644
--- a/man2/set_tid_address.2
+++ b/man2/set_tid_address.2
@@ -27,13 +27,17 @@
 set_tid_address \- set pointer to thread ID
 .SH SYNOPSIS
 .nf
-.B #include <linux/unistd.h>
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
 .PP
-.BI "pid_t set_tid_address(int *" tidptr );
+.BI "pid_t syscall(SYS_set_tid_address, int *" tidptr );
 .fi
 .PP
 .IR Note :
-There is no glibc wrapper for this system call; see NOTES.
+glibc provides no wrapper for
+.BR set_tid_address (2),
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 For each thread, the kernel maintains two attributes (addresses) called
 .I set_child_tid
@@ -99,9 +103,6 @@ This call is present since Linux 2.5.48.
 Details as given here are valid since Linux 2.5.49.
 .SH CONFORMING TO
 This system call is Linux-specific.
-.SH NOTES
-Glibc does not provide a wrapper for this system call; call it using
-.BR syscall (2).
 .SH SEE ALSO
 .BR clone (2),
 .BR futex (2),
-- 
2.31.1

[PATCH 07/10] sgetmask.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/sgetmask.2 | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/man2/sgetmask.2 b/man2/sgetmask.2
index 7563b57ff..59b58ceaf 100644
--- a/man2/sgetmask.2
+++ b/man2/sgetmask.2
@@ -27,12 +27,17 @@
 sgetmask, ssetmask \- manipulation of signal mask (obsolete)
 .SH SYNOPSIS
 .nf
-.B "long sgetmask(void);"
-.BI "long ssetmask(long " newmask );
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
+.PP
+.B "long syscall(SYS_sgetmask, void);"
+.BI "long syscall(SYS_ssetmask, long " newmask );
 .fi
 .PP
 .IR Note :
-There are no glibc wrappers for these system calls; see NOTES.
+glibc provides no wrappers for these functions,
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 These system calls are obsolete.
 .IR "Do not use them" ;
@@ -73,10 +78,6 @@ option.
 .SH CONFORMING TO
 These system calls are Linux-specific.
 .SH NOTES
-Glibc does not provide wrappers for these obsolete system calls;
-in the unlikely event that you want to call them, use
-.BR syscall (2).
-.PP
 These system calls are unaware of signal numbers greater than 31
 (i.e., real-time signals).
 .PP
-- 
2.31.1

[PATCH 08/10] shmop.2: Remove unused include

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/shmop.2 | 1 -
 1 file changed, 1 deletion(-)

diff --git a/man2/shmop.2 b/man2/shmop.2
index c7e66af0e..52baed96c 100644
--- a/man2/shmop.2
+++ b/man2/shmop.2
@@ -42,7 +42,6 @@
 shmat, shmdt \- System V shared memory operations
 .SH SYNOPSIS
 .nf
-.B #include <sys/types.h>
 .B #include <sys/shm.h>
 .PP
 .BI "void *shmat(int " shmid ", const void *" shmaddr ", int " shmflg );
-- 
2.31.1

[PATCH 09/10] sigprocmask.2: Use syscall(SYS_...); for raw system calls

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/sigprocmask.2 | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/man2/sigprocmask.2 b/man2/sigprocmask.2
index a622e4c18..b82386f14 100644
--- a/man2/sigprocmask.2
+++ b/man2/sigprocmask.2
@@ -37,12 +37,16 @@ sigprocmask, rt_sigprocmask \- examine and change blocked signals
 .BI "int sigprocmask(int " how ", const sigset_t *restrict " set ,
 .BI "                sigset_t *restrict " oldset );
 .PP
+.BR "#include <signal.h>" "           /* Definition of " SIG_* " constants */"
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
+.PP
 /* Prototype for the underlying system call */
-.BI "int rt_sigprocmask(int " how ", const kernel_sigset_t *" set ,
+.BI "int syscall(SYS_rt_sigprocmask, int " how ", const kernel_sigset_t *" set ,
 .BI "                kernel_sigset_t *" oldset ", size_t " sigsetsize );
 .PP
 /* Prototype for the legacy system call (deprecated) */
-.BI "int sigprocmask(int " how ", const old_kernel_sigset_t *" set ,
+.BI "int syscall(SYS_sigprocmask, int " how ", const old_kernel_sigset_t *" set ,
 .BI "                old_kernel_sigset_t *" oldset );
 .fi
 .PP
-- 
2.31.1

[PATCH 10/10] socketcall.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/socketcall.2 | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/man2/socketcall.2 b/man2/socketcall.2
index 8ef3d65d3..f632ba011 100644
--- a/man2/socketcall.2
+++ b/man2/socketcall.2
@@ -27,13 +27,18 @@
 socketcall \- socket system calls
 .SH SYNOPSIS
 .nf
-.B #include <linux/net.h>
+.BR "#include <linux/net.h>" "        /* Definition of " SYS_* " constants */"
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_socketcall " */"
+.B #include <unistd.h>
 .PP
-.BI "int socketcall(int " call ", unsigned long *" args );
+.BI "int syscall(SYS_socketcall, int " call ", unsigned long *" args );
 .fi
 .PP
 .IR Note :
-There is no glibc wrapper for this system call; see NOTES.
+glibc provides no wrapper for
+.BR socketcall (2),
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 .BR socketcall ()
 is a common kernel entry point for the socket system calls.
@@ -156,10 +161,6 @@ T}
 This call is specific to Linux, and should not be used in programs
 intended to be portable.
 .SH NOTES
-Glibc does not provide a wrapper for this system call;
-in the unlikely event that you want to call it directly, do so using
-.BR syscall (2).
-.PP
 On some architectures\(emfor example, x86-64 and ARM\(emthere is no
 .BR socketcall ()
 system call; instead
-- 
2.31.1

[PATCH v2 01/10] open.2: Remove unused <sys/stat.h>

[Alejandro Colomar]

I can't see a reason to include it.  <fcntl.h> provides O_*
constants for 'flags', S_* constants for 'mode', and mode_t.

Probably a long time ago, some of those weren't defined in
<fcntl.h>, and both headers needed to be included, or maybe it's
a historical error.

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/open.2 | 1 -
 1 file changed, 1 deletion(-)

diff --git a/man2/open.2 b/man2/open.2
index bac1ee806..1bcd995f4 100644
--- a/man2/open.2
+++ b/man2/open.2
@@ -53,7 +53,6 @@
 open, openat, creat \- open and possibly create a file
 .SH SYNOPSIS
 .nf
-.B #include <sys/stat.h>
 .B #include <fcntl.h>
 .PP
 .BI "int open(const char *" pathname ", int " flags );
-- 
2.31.1

[PATCH v2 02/10] rt_sigqueueinfo.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/rt_sigqueueinfo.2 | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/man2/rt_sigqueueinfo.2 b/man2/rt_sigqueueinfo.2
index 2dd1a5877..d48ecbdcd 100644
--- a/man2/rt_sigqueueinfo.2
+++ b/man2/rt_sigqueueinfo.2
@@ -27,9 +27,14 @@
 rt_sigqueueinfo, rt_tgsigqueueinfo \- queue a signal and data
 .SH SYNOPSIS
 .nf
-.BI "int rt_sigqueueinfo(pid_t " tgid ", int " sig ", siginfo_t *" info );
-.BI "int rt_tgsigqueueinfo(pid_t " tgid ", pid_t " tid ", int " sig \
-", siginfo_t *" info );
+.BR "#include <linux/signal.h>" "     /* Definition of " SI_* " constants */"
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
+.PP
+.BI "int syscall(SYS_rt_sigqueueinfo, pid_t " tgid ,
+.BI "            int " sig ", siginfo_t *" info );
+.BI "int syscall(SYS_rt_tgsigqueueinfo, pid_t " tgid ", pid_t " tid ,
+.BI "            int " sig ", siginfo_t *" info );
 .fi
 .PP
 .IR Note :
-- 
2.31.1

[PATCH v2 03/10] seccomp.2: Document why each header is needed

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Tyler Hicks <tyhicks@canonical.com>
Cc: Will Drewry <wad@chromium.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
---
 man2/seccomp.2 | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/man2/seccomp.2 b/man2/seccomp.2
index 584a16f22..1be7d6766 100644
--- a/man2/seccomp.2
+++ b/man2/seccomp.2
@@ -30,11 +30,11 @@
 seccomp \- operate on Secure Computing state of the process
 .SH SYNOPSIS
 .nf
-.B #include <linux/seccomp.h>
-.B #include <linux/filter.h>
-.B #include <linux/audit.h>
-.B #include <linux/signal.h>
-.B #include <sys/ptrace.h>
+.BR "#include <linux/seccomp.h>" "  /* Definition of " SECCOMP_* " constants */"
+.BR "#include <linux/filter.h>" "   /* Definition of " "struct sock_fprog" " */"
+.BR "#include <linux/audit.h>" "    /* Definition of " AUDIT_* " constants */"
+.BR "#include <linux/signal.h>" "   /* Definition of " SIG* " constants */"
+.BR "#include <sys/ptrace.h>" "     /* Definition of " PTRACE_* " constants */"
 .\" Kees Cook noted: Anything that uses SECCOMP_RET_TRACE returns will
 .\"                  need <sys/ptrace.h>
 .PP
-- 
2.31.1

[PATCH v2 04/10] seccomp.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/seccomp.2 | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/man2/seccomp.2 b/man2/seccomp.2
index 1be7d6766..0f3f30767 100644
--- a/man2/seccomp.2
+++ b/man2/seccomp.2
@@ -37,13 +37,18 @@ seccomp \- operate on Secure Computing state of the process
 .BR "#include <sys/ptrace.h>" "     /* Definition of " PTRACE_* " constants */"
 .\" Kees Cook noted: Anything that uses SECCOMP_RET_TRACE returns will
 .\"                  need <sys/ptrace.h>
+.BR "#include <sys/syscall.h>" "    /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
 .PP
-.BI "int seccomp(unsigned int " operation ", unsigned int " flags \
-", void *" args );
+.BI "int syscall(SYS_seccomp, unsigned int " operation ", unsigned int " flags ,
+.BI "            void *" args );
 .fi
 .PP
 .IR Note :
-There is no glibc wrapper for this system call; see NOTES.
+glibc provides no wrapper for
+.BR seccomp (),
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 The
 .BR seccomp ()
@@ -858,9 +863,6 @@ Tile (since Linux 4.3)
 PA-RISC (since Linux 4.6)
 .\" User mode Linux since Linux 4.6
 .PD
-.PP
-Glibc does not provide a wrapper for this system call; call it using
-.BR syscall (2).
 .\"
 .SS Caveats
 There are various subtleties to consider when applying seccomp filters
-- 
2.31.1

[PATCH v2 05/10] set_thread_area.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/set_thread_area.2 | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/man2/set_thread_area.2 b/man2/set_thread_area.2
index 284dea390..43f3009f7 100644
--- a/man2/set_thread_area.2
+++ b/man2/set_thread_area.2
@@ -11,28 +11,31 @@
 get_thread_area, set_thread_area \- manipulate thread-local storage information
 .SH SYNOPSIS
 .nf
-.B #include <linux/unistd.h>
+.BR "#include <sys/syscall.h>" "     /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
 .PP
 .B #if defined __i386__ || defined __x86_64__
-.B # include <asm/ldt.h>
+.BR "# include <asm/ldt.h>" "        /* Definition of " "struct user_desc" " */"
 .PP
-.BI "int get_thread_area(struct user_desc *" u_info );
-.BI "int set_thread_area(struct user_desc *" u_info );
+.BI "int syscall(SYS_get_thread_area, struct user_desc *" u_info );
+.BI "int syscall(SYS_set_thread_area, struct user_desc *" u_info );
 .PP
 .B #elif defined __m68k__
 .PP
-.B "int get_thread_area(void);"
-.BI "int set_thread_area(unsigned long " tp );
+.B "int syscall(SYS_get_thread_area);"
+.BI "int syscall(SYS_set_thread_area, unsigned long " tp );
 .PP
 .B #elif defined __mips__
 .PP
-.BI "int set_thread_area(unsigned long " addr );
+.BI "int syscall(SYS_set_thread_area, unsigned long " addr );
 .PP
 .B #endif
 .fi
 .PP
 .IR Note :
-There are no glibc wrappers for these system calls; see NOTES.
+glibc provides no wrappers for these system calls,
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 These calls provide architecture-specific support for a thread-local storage
 implementation.
@@ -172,10 +175,7 @@ and
 are Linux-specific and should not be used in programs that are intended
 to be portable.
 .SH NOTES
-Glibc does not provide wrappers for these system calls,
-since they are generally intended for use only by threading libraries.
-In the unlikely event that you want to call them directly, use
-.BR syscall (2).
+These system calls are generally intended for use only by threading libraries.
 .PP
 .BR arch_prctl (2)
 can interfere with
-- 
2.31.1

[PATCH v2 06/10] set_tid_address.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/set_tid_address.2 | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/man2/set_tid_address.2 b/man2/set_tid_address.2
index 258c9b9ef..65b4931eb 100644
--- a/man2/set_tid_address.2
+++ b/man2/set_tid_address.2
@@ -27,13 +27,17 @@
 set_tid_address \- set pointer to thread ID
 .SH SYNOPSIS
 .nf
-.B #include <linux/unistd.h>
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
 .PP
-.BI "pid_t set_tid_address(int *" tidptr );
+.BI "pid_t syscall(SYS_set_tid_address, int *" tidptr );
 .fi
 .PP
 .IR Note :
-There is no glibc wrapper for this system call; see NOTES.
+glibc provides no wrapper for
+.BR set_tid_address (),
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 For each thread, the kernel maintains two attributes (addresses) called
 .I set_child_tid
@@ -99,9 +103,6 @@ This call is present since Linux 2.5.48.
 Details as given here are valid since Linux 2.5.49.
 .SH CONFORMING TO
 This system call is Linux-specific.
-.SH NOTES
-Glibc does not provide a wrapper for this system call; call it using
-.BR syscall (2).
 .SH SEE ALSO
 .BR clone (2),
 .BR futex (2),
-- 
2.31.1

[PATCH v2 07/10] sgetmask.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/sgetmask.2 | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/man2/sgetmask.2 b/man2/sgetmask.2
index 7563b57ff..59b58ceaf 100644
--- a/man2/sgetmask.2
+++ b/man2/sgetmask.2
@@ -27,12 +27,17 @@
 sgetmask, ssetmask \- manipulation of signal mask (obsolete)
 .SH SYNOPSIS
 .nf
-.B "long sgetmask(void);"
-.BI "long ssetmask(long " newmask );
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
+.PP
+.B "long syscall(SYS_sgetmask, void);"
+.BI "long syscall(SYS_ssetmask, long " newmask );
 .fi
 .PP
 .IR Note :
-There are no glibc wrappers for these system calls; see NOTES.
+glibc provides no wrappers for these functions,
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 These system calls are obsolete.
 .IR "Do not use them" ;
@@ -73,10 +78,6 @@ option.
 .SH CONFORMING TO
 These system calls are Linux-specific.
 .SH NOTES
-Glibc does not provide wrappers for these obsolete system calls;
-in the unlikely event that you want to call them, use
-.BR syscall (2).
-.PP
 These system calls are unaware of signal numbers greater than 31
 (i.e., real-time signals).
 .PP
-- 
2.31.1

[PATCH v2 08/10] shmop.2: Remove unused include

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/shmop.2 | 1 -
 1 file changed, 1 deletion(-)

diff --git a/man2/shmop.2 b/man2/shmop.2
index c7e66af0e..52baed96c 100644
--- a/man2/shmop.2
+++ b/man2/shmop.2
@@ -42,7 +42,6 @@
 shmat, shmdt \- System V shared memory operations
 .SH SYNOPSIS
 .nf
-.B #include <sys/types.h>
 .B #include <sys/shm.h>
 .PP
 .BI "void *shmat(int " shmid ", const void *" shmaddr ", int " shmflg );
-- 
2.31.1

[PATCH v2 09/10] sigprocmask.2: Use syscall(SYS_...); for raw system calls

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/sigprocmask.2 | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/man2/sigprocmask.2 b/man2/sigprocmask.2
index a622e4c18..b82386f14 100644
--- a/man2/sigprocmask.2
+++ b/man2/sigprocmask.2
@@ -37,12 +37,16 @@ sigprocmask, rt_sigprocmask \- examine and change blocked signals
 .BI "int sigprocmask(int " how ", const sigset_t *restrict " set ,
 .BI "                sigset_t *restrict " oldset );
 .PP
+.BR "#include <signal.h>" "           /* Definition of " SIG_* " constants */"
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
+.B #include <unistd.h>
+.PP
 /* Prototype for the underlying system call */
-.BI "int rt_sigprocmask(int " how ", const kernel_sigset_t *" set ,
+.BI "int syscall(SYS_rt_sigprocmask, int " how ", const kernel_sigset_t *" set ,
 .BI "                kernel_sigset_t *" oldset ", size_t " sigsetsize );
 .PP
 /* Prototype for the legacy system call (deprecated) */
-.BI "int sigprocmask(int " how ", const old_kernel_sigset_t *" set ,
+.BI "int syscall(SYS_sigprocmask, int " how ", const old_kernel_sigset_t *" set ,
 .BI "                old_kernel_sigset_t *" oldset );
 .fi
 .PP
-- 
2.31.1

[PATCH v2 10/10] socketcall.2: Use syscall(SYS_...); for system calls without a wrapper

[Alejandro Colomar]

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
---
 man2/socketcall.2 | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/man2/socketcall.2 b/man2/socketcall.2
index 8ef3d65d3..883c2aafe 100644
--- a/man2/socketcall.2
+++ b/man2/socketcall.2
@@ -27,13 +27,18 @@
 socketcall \- socket system calls
 .SH SYNOPSIS
 .nf
-.B #include <linux/net.h>
+.BR "#include <linux/net.h>" "        /* Definition of " SYS_* " constants */"
+.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_socketcall " */"
+.B #include <unistd.h>
 .PP
-.BI "int socketcall(int " call ", unsigned long *" args );
+.BI "int syscall(SYS_socketcall, int " call ", unsigned long *" args );
 .fi
 .PP
 .IR Note :
-There is no glibc wrapper for this system call; see NOTES.
+glibc provides no wrapper for
+.BR socketcall (),
+necessitating the use of
+.BR syscall (2).
 .SH DESCRIPTION
 .BR socketcall ()
 is a common kernel entry point for the socket system calls.
@@ -156,10 +161,6 @@ T}
 This call is specific to Linux, and should not be used in programs
 intended to be portable.
 .SH NOTES
-Glibc does not provide a wrapper for this system call;
-in the unlikely event that you want to call it directly, do so using
-.BR syscall (2).
-.PP
 On some architectures\(emfor example, x86-64 and ARM\(emthere is no
 .BR socketcall ()
 system call; instead
@@ -198,3 +199,4 @@ and also provides a (very) small performance improvement.
 .BR shutdown (2),
 .BR socket (2),
 .BR socketpair (2)
+
-- 
2.31.1

Re: [PATCH 00/10] man2: SYNOPSIS: Use syscall, and fix includes.

[Alejandro Colomar (man-pages)]

Hi Michael,

On 5/22/21 1:25 AM, Alejandro Colomar wrote:
> Hi Michael,
> 
> Almost finished these:
> $ man_lsfunc man2 | sed -n '/socketcall/,$p' | wc -l
> 67
> $ man_lsfunc man2 | wc -l
> 401
> $

Right now I have around 20-30 more patches about this pending, and 
that's all; I've finished.

I'll hold them until you apply (or reject) the various patch sets that 
are pending to avoid creating conflicts.

After that, I think we should release v5.12 :).

And after the release, I have 4 branches to develop a few things for 
5.13/14 (some of them depending on what we decide about some RFCs):
  - library: Use LIBRARY section (similar to FreeBSD)
  - stdint: Use standard (<stdint.h>) types, as well as standard attributes.
  - system_data_types: Document more types (I was documenting those you 
mark in TLPI as important).
  - examples: Fix the includes in the EXAMPLES sections to use the 
minimum number of includes, by using iwyu(1).



Cheers,

Alex


-- 
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/

Re: [PATCH v2 03/10] seccomp.2: Document why each header is needed

[Alejandro Colomar (man-pages)]

Hi Michael,

On 5/24/21 8:19 PM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
> Cc: Kees Cook <keescook@chromium.org>
> Cc: Tyler Hicks <tyhicks@canonical.com>
> Cc: Will Drewry <wad@chromium.org>
> Cc: Michael Kerrisk <mtk.manpages@gmail.com>
> ---

Please discards this one, for the same reasons as in v1: repeated.
See 
<https://lore.kernel.org/linux-man/20210515182254.186607-1-alx.manpages@gmail.com/T/#u> 
instead.

That one has a better mail.

Thanks,

Alex


>   man2/seccomp.2 | 10 +++++-----
>   1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/man2/seccomp.2 b/man2/seccomp.2
> index 584a16f22..1be7d6766 100644
> --- a/man2/seccomp.2
> +++ b/man2/seccomp.2
> @@ -30,11 +30,11 @@
>   seccomp \- operate on Secure Computing state of the process
>   .SH SYNOPSIS
>   .nf
> -.B #include <linux/seccomp.h>
> -.B #include <linux/filter.h>
> -.B #include <linux/audit.h>
> -.B #include <linux/signal.h>
> -.B #include <sys/ptrace.h>
> +.BR "#include <linux/seccomp.h>" "  /* Definition of " SECCOMP_* " constants */"
> +.BR "#include <linux/filter.h>" "   /* Definition of " "struct sock_fprog" " */"
> +.BR "#include <linux/audit.h>" "    /* Definition of " AUDIT_* " constants */"
> +.BR "#include <linux/signal.h>" "   /* Definition of " SIG* " constants */"
> +.BR "#include <sys/ptrace.h>" "     /* Definition of " PTRACE_* " constants */"
>   .\" Kees Cook noted: Anything that uses SECCOMP_RET_TRACE returns will
>   .\"                  need <sys/ptrace.h>
>   .PP
> 


-- 
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/

Re: [PATCH v2 01/10] open.2: Remove unused <sys/stat.h>

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> I can't see a reason to include it.  <fcntl.h> provides O_*
> constants for 'flags', S_* constants for 'mode', and mode_t.
> 
> Probably a long time ago, some of those weren't defined in
> <fcntl.h>, and both headers needed to be included, or maybe it's
> a historical error.
> 

Applied.

Thanks,

Michael


> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
> ---
>  man2/open.2 | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/man2/open.2 b/man2/open.2
> index bac1ee806..1bcd995f4 100644
> --- a/man2/open.2
> +++ b/man2/open.2
> @@ -53,7 +53,6 @@
>  open, openat, creat \- open and possibly create a file
>  .SH SYNOPSIS
>  .nf
> -.B #include <sys/stat.h>
>  .B #include <fcntl.h>
>  .PP
>  .BI "int open(const char *" pathname ", int " flags );
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 02/10] rt_sigqueueinfo.2: Use syscall(SYS_...); for system calls without a wrapper

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

Applied.

Thanks,

Michael


> ---
>  man2/rt_sigqueueinfo.2 | 11 ++++++++---
>  1 file changed, 8 insertions(+), 3 deletions(-)
> 
> diff --git a/man2/rt_sigqueueinfo.2 b/man2/rt_sigqueueinfo.2
> index 2dd1a5877..d48ecbdcd 100644
> --- a/man2/rt_sigqueueinfo.2
> +++ b/man2/rt_sigqueueinfo.2
> @@ -27,9 +27,14 @@
>  rt_sigqueueinfo, rt_tgsigqueueinfo \- queue a signal and data
>  .SH SYNOPSIS
>  .nf
> -.BI "int rt_sigqueueinfo(pid_t " tgid ", int " sig ", siginfo_t *" info );
> -.BI "int rt_tgsigqueueinfo(pid_t " tgid ", pid_t " tid ", int " sig \
> -", siginfo_t *" info );
> +.BR "#include <linux/signal.h>" "     /* Definition of " SI_* " constants */"
> +.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
> +.B #include <unistd.h>
> +.PP
> +.BI "int syscall(SYS_rt_sigqueueinfo, pid_t " tgid ,
> +.BI "            int " sig ", siginfo_t *" info );
> +.BI "int syscall(SYS_rt_tgsigqueueinfo, pid_t " tgid ", pid_t " tid ,
> +.BI "            int " sig ", siginfo_t *" info );
>  .fi
>  .PP
>  .IR Note :
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 05/10] set_thread_area.2: Use syscall(SYS_...); for system calls without a wrapper

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
> ---
>  man2/set_thread_area.2 | 24 ++++++++++++------------
>  1 file changed, 12 insertions(+), 12 deletions(-)

Applied.

Thanks,

Michael


> diff --git a/man2/set_thread_area.2 b/man2/set_thread_area.2
> index 284dea390..43f3009f7 100644
> --- a/man2/set_thread_area.2
> +++ b/man2/set_thread_area.2
> @@ -11,28 +11,31 @@
>  get_thread_area, set_thread_area \- manipulate thread-local storage information
>  .SH SYNOPSIS
>  .nf
> -.B #include <linux/unistd.h>
> +.BR "#include <sys/syscall.h>" "     /* Definition of " SYS_* " constants */"
> +.B #include <unistd.h>
>  .PP
>  .B #if defined __i386__ || defined __x86_64__
> -.B # include <asm/ldt.h>
> +.BR "# include <asm/ldt.h>" "        /* Definition of " "struct user_desc" " */"
>  .PP
> -.BI "int get_thread_area(struct user_desc *" u_info );
> -.BI "int set_thread_area(struct user_desc *" u_info );
> +.BI "int syscall(SYS_get_thread_area, struct user_desc *" u_info );
> +.BI "int syscall(SYS_set_thread_area, struct user_desc *" u_info );
>  .PP
>  .B #elif defined __m68k__
>  .PP
> -.B "int get_thread_area(void);"
> -.BI "int set_thread_area(unsigned long " tp );
> +.B "int syscall(SYS_get_thread_area);"
> +.BI "int syscall(SYS_set_thread_area, unsigned long " tp );
>  .PP
>  .B #elif defined __mips__
>  .PP
> -.BI "int set_thread_area(unsigned long " addr );
> +.BI "int syscall(SYS_set_thread_area, unsigned long " addr );
>  .PP
>  .B #endif
>  .fi
>  .PP
>  .IR Note :
> -There are no glibc wrappers for these system calls; see NOTES.
> +glibc provides no wrappers for these system calls,
> +necessitating the use of
> +.BR syscall (2).
>  .SH DESCRIPTION
>  These calls provide architecture-specific support for a thread-local storage
>  implementation.
> @@ -172,10 +175,7 @@ and
>  are Linux-specific and should not be used in programs that are intended
>  to be portable.
>  .SH NOTES
> -Glibc does not provide wrappers for these system calls,
> -since they are generally intended for use only by threading libraries.
> -In the unlikely event that you want to call them directly, use
> -.BR syscall (2).
> +These system calls are generally intended for use only by threading libraries.
>  .PP
>  .BR arch_prctl (2)
>  can interfere with
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 06/10] set_tid_address.2: Use syscall(SYS_...); for system calls without a wrapper

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

Applied.

Thanks,

Michael


> ---
>  man2/set_tid_address.2 | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/man2/set_tid_address.2 b/man2/set_tid_address.2
> index 258c9b9ef..65b4931eb 100644
> --- a/man2/set_tid_address.2
> +++ b/man2/set_tid_address.2
> @@ -27,13 +27,17 @@
>  set_tid_address \- set pointer to thread ID
>  .SH SYNOPSIS
>  .nf
> -.B #include <linux/unistd.h>
> +.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
> +.B #include <unistd.h>
>  .PP
> -.BI "pid_t set_tid_address(int *" tidptr );
> +.BI "pid_t syscall(SYS_set_tid_address, int *" tidptr );
>  .fi
>  .PP
>  .IR Note :
> -There is no glibc wrapper for this system call; see NOTES.
> +glibc provides no wrapper for
> +.BR set_tid_address (),
> +necessitating the use of
> +.BR syscall (2).
>  .SH DESCRIPTION
>  For each thread, the kernel maintains two attributes (addresses) called
>  .I set_child_tid
> @@ -99,9 +103,6 @@ This call is present since Linux 2.5.48.
>  Details as given here are valid since Linux 2.5.49.
>  .SH CONFORMING TO
>  This system call is Linux-specific.
> -.SH NOTES
> -Glibc does not provide a wrapper for this system call; call it using
> -.BR syscall (2).
>  .SH SEE ALSO
>  .BR clone (2),
>  .BR futex (2),
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 07/10] sgetmask.2: Use syscall(SYS_...); for system calls without a wrapper

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

Applied.

Thanks,

Michael


> ---
>  man2/sgetmask.2 | 15 ++++++++-------
>  1 file changed, 8 insertions(+), 7 deletions(-)
> 
> diff --git a/man2/sgetmask.2 b/man2/sgetmask.2
> index 7563b57ff..59b58ceaf 100644
> --- a/man2/sgetmask.2
> +++ b/man2/sgetmask.2
> @@ -27,12 +27,17 @@
>  sgetmask, ssetmask \- manipulation of signal mask (obsolete)
>  .SH SYNOPSIS
>  .nf
> -.B "long sgetmask(void);"
> -.BI "long ssetmask(long " newmask );
> +.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
> +.B #include <unistd.h>
> +.PP
> +.B "long syscall(SYS_sgetmask, void);"
> +.BI "long syscall(SYS_ssetmask, long " newmask );
>  .fi
>  .PP
>  .IR Note :
> -There are no glibc wrappers for these system calls; see NOTES.
> +glibc provides no wrappers for these functions,
> +necessitating the use of
> +.BR syscall (2).
>  .SH DESCRIPTION
>  These system calls are obsolete.
>  .IR "Do not use them" ;
> @@ -73,10 +78,6 @@ option.
>  .SH CONFORMING TO
>  These system calls are Linux-specific.
>  .SH NOTES
> -Glibc does not provide wrappers for these obsolete system calls;
> -in the unlikely event that you want to call them, use
> -.BR syscall (2).
> -.PP
>  These system calls are unaware of signal numbers greater than 31
>  (i.e., real-time signals).
>  .PP
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 08/10] shmop.2: Remove unused include

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

Applied.

Thanks,

Michael


> ---
>  man2/shmop.2 | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/man2/shmop.2 b/man2/shmop.2
> index c7e66af0e..52baed96c 100644
> --- a/man2/shmop.2
> +++ b/man2/shmop.2
> @@ -42,7 +42,6 @@
>  shmat, shmdt \- System V shared memory operations
>  .SH SYNOPSIS
>  .nf
> -.B #include <sys/types.h>
>  .B #include <sys/shm.h>
>  .PP
>  .BI "void *shmat(int " shmid ", const void *" shmaddr ", int " shmflg );
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 09/10] sigprocmask.2: Use syscall(SYS_...); for raw system calls

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

Applied.

Thanks,

Michael


> ---
>  man2/sigprocmask.2 | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/man2/sigprocmask.2 b/man2/sigprocmask.2
> index a622e4c18..b82386f14 100644
> --- a/man2/sigprocmask.2
> +++ b/man2/sigprocmask.2
> @@ -37,12 +37,16 @@ sigprocmask, rt_sigprocmask \- examine and change blocked signals
>  .BI "int sigprocmask(int " how ", const sigset_t *restrict " set ,
>  .BI "                sigset_t *restrict " oldset );
>  .PP
> +.BR "#include <signal.h>" "           /* Definition of " SIG_* " constants */"
> +.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_* " constants */"
> +.B #include <unistd.h>
> +.PP
>  /* Prototype for the underlying system call */
> -.BI "int rt_sigprocmask(int " how ", const kernel_sigset_t *" set ,
> +.BI "int syscall(SYS_rt_sigprocmask, int " how ", const kernel_sigset_t *" set ,
>  .BI "                kernel_sigset_t *" oldset ", size_t " sigsetsize );
>  .PP
>  /* Prototype for the legacy system call (deprecated) */
> -.BI "int sigprocmask(int " how ", const old_kernel_sigset_t *" set ,
> +.BI "int syscall(SYS_sigprocmask, int " how ", const old_kernel_sigset_t *" set ,
>  .BI "                old_kernel_sigset_t *" oldset );
>  .fi
>  .PP
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 10/10] socketcall.2: Use syscall(SYS_...); for system calls without a wrapper

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

Applied.

Thanks,

Michael


> ---
>  man2/socketcall.2 | 16 +++++++++-------
>  1 file changed, 9 insertions(+), 7 deletions(-)
> 
> diff --git a/man2/socketcall.2 b/man2/socketcall.2
> index 8ef3d65d3..883c2aafe 100644
> --- a/man2/socketcall.2
> +++ b/man2/socketcall.2
> @@ -27,13 +27,18 @@
>  socketcall \- socket system calls
>  .SH SYNOPSIS
>  .nf
> -.B #include <linux/net.h>
> +.BR "#include <linux/net.h>" "        /* Definition of " SYS_* " constants */"
> +.BR "#include <sys/syscall.h>" "      /* Definition of " SYS_socketcall " */"
> +.B #include <unistd.h>
>  .PP
> -.BI "int socketcall(int " call ", unsigned long *" args );
> +.BI "int syscall(SYS_socketcall, int " call ", unsigned long *" args );
>  .fi
>  .PP
>  .IR Note :
> -There is no glibc wrapper for this system call; see NOTES.
> +glibc provides no wrapper for
> +.BR socketcall (),
> +necessitating the use of
> +.BR syscall (2).
>  .SH DESCRIPTION
>  .BR socketcall ()
>  is a common kernel entry point for the socket system calls.
> @@ -156,10 +161,6 @@ T}
>  This call is specific to Linux, and should not be used in programs
>  intended to be portable.
>  .SH NOTES
> -Glibc does not provide a wrapper for this system call;
> -in the unlikely event that you want to call it directly, do so using
> -.BR syscall (2).
> -.PP
>  On some architectures\(emfor example, x86-64 and ARM\(emthere is no
>  .BR socketcall ()
>  system call; instead
> @@ -198,3 +199,4 @@ and also provides a (very) small performance improvement.
>  .BR shutdown (2),
>  .BR socket (2),
>  .BR socketpair (2)
> +
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH 00/10] man2: SYNOPSIS: Use syscall, and fix includes.

[Michael Kerrisk (man-pages)]

Hi Alex,

On 5/25/21 7:09 AM, Alejandro Colomar (man-pages) wrote:
> Hi Michael,
> 
> On 5/22/21 1:25 AM, Alejandro Colomar wrote:
>> Hi Michael,
>>
>> Almost finished these:
>> $ man_lsfunc man2 | sed -n '/socketcall/,$p' | wc -l
>> 67
>> $ man_lsfunc man2 | wc -l
>> 401
>> $

From v2, I applied all except 3+4. Could you please resend
whatever should be applied to seccomp.2.

Thanks,

Michael

> Right now I have around 20-30 more patches about this pending, and 
> that's all; I've finished.
> 
> I'll hold them until you apply (or reject) the various patch sets that 
> are pending to avoid creating conflicts.
> 
> After that, I think we should release v5.12 :).
> 
> And after the release, I have 4 branches to develop a few things for 
> 5.13/14 (some of them depending on what we decide about some RFCs):
>   - library: Use LIBRARY section (similar to FreeBSD)
>   - stdint: Use standard (<stdint.h>) types, as well as standard attributes.
>   - system_data_types: Document more types (I was documenting those you 
> mark in TLPI as important).
>   - examples: Fix the includes in the EXAMPLES sections to use the 
> minimum number of includes, by using iwyu(1).
> 
> 
> 
> Cheers,
> 
> Alex
> 
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH 00/10] man2: SYNOPSIS: Use syscall, and fix includes.

[Michael Kerrisk (man-pages)]

[I hit send too soon....]

On 5/25/21 7:09 AM, Alejandro Colomar (man-pages) wrote:
> Hi Michael,
> 
> On 5/22/21 1:25 AM, Alejandro Colomar wrote:
>> Hi Michael,
>>
>> Almost finished these:
>> $ man_lsfunc man2 | sed -n '/socketcall/,$p' | wc -l
>> 67
>> $ man_lsfunc man2 | wc -l
>> 401
>> $
> 
> Right now I have around 20-30 more patches about this pending, and 
> that's all; I've finished.

Good!

> I'll hold them until you apply (or reject) the various patch sets that 
> are pending to avoid creating conflicts.
> 
> After that, I think we should release v5.12 :).
> 
> And after the release, I have 4 branches to develop a few things for 
> 5.13/14 (some of them depending on what we decide about some RFCs):
>   - library: Use LIBRARY section (similar to FreeBSD)
>   - stdint: Use standard (<stdint.h>) types, as well as standard attributes.
>   - system_data_types: Document more types (I was documenting those you 
> mark in TLPI as important).
>   - examples: Fix the includes in the EXAMPLES sections to use the 
> minimum number of includes, by using iwyu(1).

Okay.

Thanks,

Michael


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 03/10] seccomp.2: Document why each header is needed

[Michael Kerrisk (man-pages)]

On 5/25/21 7:34 AM, Alejandro Colomar (man-pages) wrote:
> Hi Michael,
> 
> On 5/24/21 8:19 PM, Alejandro Colomar wrote:
>> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
>> Cc: Kees Cook <keescook@chromium.org>
>> Cc: Tyler Hicks <tyhicks@canonical.com>
>> Cc: Will Drewry <wad@chromium.org>
>> Cc: Michael Kerrisk <mtk.manpages@gmail.com>
>> ---
> 
> Please discards this one, for the same reasons as in v1: repeated.
> See 
> <https://lore.kernel.org/linux-man/20210515182254.186607-1-alx.manpages@gmail.com/T/#u> 
> instead.
> 
> That one has a better mail.

Dropped.

Thanks,

Michael

>>   man2/seccomp.2 | 10 +++++-----
>>   1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/man2/seccomp.2 b/man2/seccomp.2
>> index 584a16f22..1be7d6766 100644
>> --- a/man2/seccomp.2
>> +++ b/man2/seccomp.2
>> @@ -30,11 +30,11 @@
>>   seccomp \- operate on Secure Computing state of the process
>>   .SH SYNOPSIS
>>   .nf
>> -.B #include <linux/seccomp.h>
>> -.B #include <linux/filter.h>
>> -.B #include <linux/audit.h>
>> -.B #include <linux/signal.h>
>> -.B #include <sys/ptrace.h>
>> +.BR "#include <linux/seccomp.h>" "  /* Definition of " SECCOMP_* " constants */"
>> +.BR "#include <linux/filter.h>" "   /* Definition of " "struct sock_fprog" " */"
>> +.BR "#include <linux/audit.h>" "    /* Definition of " AUDIT_* " constants */"
>> +.BR "#include <linux/signal.h>" "   /* Definition of " SIG* " constants */"
>> +.BR "#include <sys/ptrace.h>" "     /* Definition of " PTRACE_* " constants */"
>>   .\" Kees Cook noted: Anything that uses SECCOMP_RET_TRACE returns will
>>   .\"                  need <sys/ptrace.h>
>>   .PP
>>
> 
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

Re: [PATCH v2 04/10] seccomp.2: Use syscall(SYS_...); for system calls without a wrapper

[Michael Kerrisk (man-pages)]

On 5/25/21 6:19 AM, Alejandro Colomar wrote:
> Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>

This patch doesn't apply. Probably because I dropped 3/10?

Thanks,

Michael

> ---
>  man2/seccomp.2 | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)
> 
> diff --git a/man2/seccomp.2 b/man2/seccomp.2
> index 1be7d6766..0f3f30767 100644
> --- a/man2/seccomp.2
> +++ b/man2/seccomp.2
> @@ -37,13 +37,18 @@ seccomp \- operate on Secure Computing state of the process
>  .BR "#include <sys/ptrace.h>" "     /* Definition of " PTRACE_* " constants */"
>  .\" Kees Cook noted: Anything that uses SECCOMP_RET_TRACE returns will
>  .\"                  need <sys/ptrace.h>
> +.BR "#include <sys/syscall.h>" "    /* Definition of " SYS_* " constants */"
> +.B #include <unistd.h>
>  .PP
> -.BI "int seccomp(unsigned int " operation ", unsigned int " flags \
> -", void *" args );
> +.BI "int syscall(SYS_seccomp, unsigned int " operation ", unsigned int " flags ,
> +.BI "            void *" args );
>  .fi
>  .PP
>  .IR Note :
> -There is no glibc wrapper for this system call; see NOTES.
> +glibc provides no wrapper for
> +.BR seccomp (),
> +necessitating the use of
> +.BR syscall (2).
>  .SH DESCRIPTION
>  The
>  .BR seccomp ()
> @@ -858,9 +863,6 @@ Tile (since Linux 4.3)
>  PA-RISC (since Linux 4.6)
>  .\" User mode Linux since Linux 4.6
>  .PD
> -.PP
> -Glibc does not provide a wrapper for this system call; call it using
> -.BR syscall (2).
>  .\"
>  .SS Caveats
>  There are various subtleties to consider when applying seccomp filters
> 


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/