bug#64954: GNU 'uptime' on OpenBSD always prints "0 users"

[Paul Eggert]

Thanks, I propagated that into Coreutils and installed the simplified 
patch I mentioned yesterday. Closing the coreutils bug report.

bug#64954: GNU 'uptime' on OpenBSD always prints "0 users"

[Bruno Haible]

I wrote in :
>   - Gnulib modules should better provide .h files that can be #included
> on any platform. Thus, it's Gnulib's task to provide a readutmp.h
> and a read_utmp() function that can also be used on native Windows.

Done as follows:


2023-07-31  Bruno Haible  

readutmp: Make the header file and function usable on all platforms.
* lib/readutmp.h (struct gl_utmp, UTMP_STRUCT_NAME, UT_TIME_MEMBER,
UT_EXIT_E_TERMINATION, UT_EXIT_E_EXIT, UT_USER): Provide fallback
definitions.
(READ_UTMP_SUPPORTED): New macro.
* lib/readutmp.c (read_utmp) [!READ_UTMP_SUPPORTED]: Provide a dummy
definition.
* modules/readutmp (Depends-on): Add sys_time.
(configure.ac): Remove conditional.
(Makefile.am): Compile readutmp.c on all platforms.
(Include): Include readutmp.h on all platforms.
* tests/test-readutmp.c: Include readutmp.h on all platforms.
(main): Invoke read_utmp on all platforms.

diff --git a/lib/readutmp.c b/lib/readutmp.c
index acffe1000e..af43d1ad6b 100644
--- a/lib/readutmp.c
+++ b/lib/readutmp.c
@@ -61,6 +61,8 @@ extract_trimmed_name (const STRUCT_UTMP *ut)
   return trimmed_name;
 }
 
+#if READ_UTMP_SUPPORTED
+
 /* Is the utmp entry U desired by the user who asked for OPTIONS?  */
 
 static bool
@@ -77,12 +79,12 @@ desirable_utmp_entry (STRUCT_UTMP const *u, int options)
   return true;
 }
 
-#ifdef UTMP_NAME_FUNCTION
+# if defined UTMP_NAME_FUNCTION
 
 static void
 copy_utmp_entry (STRUCT_UTMP *dst, STRUCT_UTMP *src)
 {
-#if __GLIBC__ && _TIME_BITS == 64
+#  if __GLIBC__ && _TIME_BITS == 64
   /* Convert from external form in SRC to internal form in DST.
  It is OK to convert now, rather than earlier, before
  desirable_utmp_entry was invoked, because desirable_utmp_entry
@@ -119,9 +121,9 @@ copy_utmp_entry (STRUCT_UTMP *dst, STRUCT_UTMP *src)
   dst->ut_tv.tv_sec = s->ut_tv.tv_sec;
   dst->ut_tv.tv_usec = s->ut_tv.tv_usec;
   memcpy (&dst->ut_addr_v6, s->ut_addr_v6, sizeof dst->ut_addr_v6);
-#else
+#  else
   *dst = *src;
-#endif
+#  endif
 }
 
 int
@@ -158,7 +160,7 @@ read_utmp (char const *file, size_t *n_entries, STRUCT_UTMP 
**utmp_buf,
   return 0;
 }
 
-#else
+# else
 
 int
 read_utmp (char const *file, size_t *n_entries, STRUCT_UTMP **utmp_buf,
@@ -198,4 +200,16 @@ read_utmp (char const *file, size_t *n_entries, 
STRUCT_UTMP **utmp_buf,
   return 0;
 }
 
+# endif
+
+#else /* dummy fallback */
+
+int
+read_utmp (char const *file, size_t *n_entries, STRUCT_UTMP **utmp_buf,
+   int options)
+{
+  errno = ENOSYS;
+  return -1;
+}
+
 #endif
diff --git a/lib/readutmp.h b/lib/readutmp.h
index 7d2a628135..d710699525 100644
--- a/lib/readutmp.h
+++ b/lib/readutmp.h
@@ -38,6 +38,7 @@
 # endif
 
 # if HAVE_UTMPX_H
+
 #  if HAVE_UTMP_H
 /* HPUX 10.20 needs utmp.h, for the definition of e.g., UTMP_FILE.  */
 #   include 
@@ -114,6 +115,24 @@
 #   endif
 #  endif
 
+# else
+
+/* Provide a dummy fallback.  */
+
+/* Get 'struct timeval'.  */
+#  include 
+
+struct gl_utmp
+{
+  char ut_user[1];
+  char ut_line[1];
+  struct timeval ut_tv;
+};
+#  define UTMP_STRUCT_NAME gl_utmp
+#  define UT_TIME_MEMBER(UT_PTR) ((UT_PTR)->ut_tv.tv_sec)
+#  define UT_EXIT_E_TERMINATION(U) 0
+#  define UT_EXIT_E_EXIT(U) 0
+
 # endif
 
 /* Accessor macro for the member named ut_user or ut_name.  */
@@ -137,6 +156,10 @@
 #   define UT_USER(Utmp) ((Utmp)->ut_name)
 #  endif
 
+# else /* dummy fallback */
+
+#  define UT_USER(Utmp) ((Utmp)->ut_user)
+
 # endif
 
 # define HAVE_STRUCT_XTMP_UT_EXIT \
@@ -151,10 +174,14 @@
 (HAVE_STRUCT_UTMP_UT_PID \
  || HAVE_STRUCT_UTMPX_UT_PID)
 
+/* Type of entry returned by read_utmp().  */
 typedef struct UTMP_STRUCT_NAME STRUCT_UTMP;
 
+/* Size of the UT_USER (ut) member.  */
 enum { UT_USER_SIZE = sizeof UT_USER ((STRUCT_UTMP *) 0) };
 
+/* Definition of UTMP_FILE and WTMP_FILE.  */
+
 # if !defined UTMP_FILE && defined _PATH_UTMP
 #  define UTMP_FILE _PATH_UTMP
 # endif
@@ -181,12 +208,15 @@ enum { UT_USER_SIZE = sizeof UT_USER ((STRUCT_UTMP *) 0) 
};
 #  define WTMP_FILE "/etc/wtmp"
 # endif
 
+/* Accessor macro for the member named ut_pid.  */
 # if HAVE_STRUCT_XTMP_UT_PID
 #  define UT_PID(U) ((U)->ut_pid)
 # else
 #  define UT_PID(U) 0
 # endif
 
+/* Accessor macros for the member named ut_type.  */
+
 # if HAVE_STRUCT_UTMP_UT_TYPE || HAVE_STRUCT_UTMPX_UT_TYPE
 #  define UT_TYPE_EQ(U, V) ((U)->ut_type == (V))
 #  define UT_TYPE_NOT_DEFINED 0
@@ -207,11 +237,17 @@ enum { UT_USER_SIZE = sizeof UT_USER ((STRUCT_UTMP *) 0) 
};
 #  define UT_TYPE_USER_PROCESS(U) 0
 # endif
 
+/* Determines whether an entry *U corresponds to a user process.  */
 # define IS_USER_PROCESS(U) \
(UT_USER (U)[0]  \
 && (UT_TYPE_USER_PROCESS (U)\
 || (UT_TYPE_NOT_DEFINED && UT_TIME_MEMBER (U) != 0)))

bug#64937: "who" reports funny dates

[Thorsten Kukuk via GNU coreutils Bug Reports]

Hi,

On Sun, Jul 30, Paul Eggert wrote:

> Thorsten's draft coreutils patches 
>  are a bit 
> long, and I'm hoping we can simplify this by packaging the fix inside 
> Gnulib (much as we already packaged the fix for  not working on 
> 32-bit GNU/Linux x86 and ARM when _TIME_BITS is 64[1]), so that we 
> needn't patch Coreutils other than to upgrade Gnulib version.

If you haven't seen yet, I made some time ago a mapping
between utmp struct entries and libsystemd functions: 
https://github.com/thkukuk/utmpx/blob/main/utmp-to-logind.md

But be aware:
1. you need to extend ut_tv from 32bit time_t to 64bit time_t, or your
patch will not survive 2038.
2. the string entries are size limited in struct utmp, but are not with
libsystemd. So could be that you need to truncate device names, user
names and host names.

  Thorsten

-- 
Thorsten Kukuk, Distinguished Engineer, Senior Architect, Future Technologies
SUSE Software Solutions Germany GmbH, Frankenstraße 146, 90461 Nuernberg, 
Germany
Managing Director: Ivo Totev, Andrew McDonald, Werner Knoblich
(HRB 36809, AG Nürnberg)

bug#64954: GNU 'uptime' on OpenBSD always prints "0 users"

[Bruno Haible]

Paul Eggert wrote:
> > I'm fine with the change, but we'll also need to adjust
> > the sc_prohibit_always_true_header_tests syntax check in gnulib
> 
> I looked into that but it's such a hassle that I came up with the 
> attached simpler patch to Coreutils. How about installing it instead?

Yes, that's better than what I proposed, on two accounts:
  - Gnulib modules should better provide .h files that can be #included
on any platform. Thus, it's Gnulib's task to provide a readutmp.h
and a read_utmp() function that can also be used on native Windows.
  - It gets rid of the horrible hack to pass the values of two uninitialized
variables down to a function.

Pádraig Brady wrote:
> gnulib docs
> state the following platforms have neither getutent or getutxent,
> and so might have issues with this?

OpenBSD 6.7, Minix 3.1.8, mingw, MSVC 14

On these platforms the #include "readutmp.h" would already provoke a
mass of syntax errors.

Btw, we don't test on Minix any longer (since Minix is dead),
and OpenBSD 6.x is end-of-life already [1].

Bruno

https://en.wikipedia.org/wiki/OpenBSD#Releases

bug#64954: GNU 'uptime' on OpenBSD always prints "0 users"

[Pádraig Brady]

On 30/07/2023 20:43, Paul Eggert wrote:

On 2023-07-30 11:41, Pádraig Brady wrote:

I'm fine with the change, but we'll also need to adjust
the sc_prohibit_always_true_header_tests syntax check in gnulib


I looked into that but it's such a hassle that I came up with the
attached simpler patch to Coreutils. How about installing it instead? No
Gnulib change should be needed.


Maybe, though I notice that gnulib docs
state the following platforms have neither getutent or getutxent,
and so might have issues with this?

OpenBSD 6.7, Minix 3.1.8, mingw, MSVC 14

bug#64937: "who" reports funny dates

[Paul Eggert]

On 2023-07-31 00:08, Thorsten Kukuk wrote:

1. you need to extend ut_tv from 32bit time_t to 64bit time_t, or your
patch will not survive 2038.


Yes, that's the plan. Gnulib's readutmp module already does this, in 
apps that also use the year2038 or year2038-recommended modules (which 
most do).



2. the string entries are size limited in struct utmp, but are not with
libsystemd. So could be that you need to truncate device names, user
names and host names.


This suggests that Gnulib needs to redo the readutmp API so that it's 
not tied to struct utmpx's fixed string lengths. That's a bit of a 
hassle but should be doable (change char array to char *). An advantage 
is that Gnulib is a source-code library and so we can make binary- (and 
even source-) incompatible changes to the API without too much trouble.


Thanks for the advice.