Skip to contentSkip to conversation

syscalls/ustat: Use __kernel_daddr_t instead of daddr_t

PATCH New Petr Vorel 4 participants 5 messages Jun 17, 2019 – Jul 29, 2019

lore
Share & tools
lore

Petr Vorel

author
#0 lore raw

PATCH 1/1 syscalls/ustat: Use __kernel_daddr_t instead of daddr_t

This fixes build for musl, which does not have <sys/ustat.h> nor typedef daddr_t. Musl does not define these as ustat is deprecated in favor of statfs() and daddr_t is ancient [1].

Other libc also does not have <sys/ustat.h> (ustat support was removed from glibc in 2.28), but they have daddr_t.

This brings <linux/types.h> dependency for ustat testing.

Typedef is needed as sparc and mips have __kernel_daddr_t long, otherwise it's int.

[1] https://www.openwall.com/lists/musl/2019/06/10/1

---
 configure.ac                              | 1 +
 include/lapi/ustat.h                      | 5 +++--
 testcases/kernel/syscalls/ustat/ustat01.c | 8 +++++++-
 testcases/kernel/syscalls/ustat/ustat02.c | 8 +++++++-
 4 files changed, 18 insertions(+), 4 deletions(-)
configure.ac +1 -0
index 53ad784d7..8ca7e8065 100644
@@ -46,6 +46,7 @@ AC_CHECK_HEADERS([ \
     linux/mempolicy.h \
     linux/module.h \
     linux/netlink.h \
+    linux/types.h \
     linux/userfaultfd.h \
     mm.h \
     netinet/sctp.h \
include/lapi/ustat.h +3 -2
index 12c073582..98633e749 100644
@@ -9,9 +9,10 @@
 
 #ifdef HAVE_SYS_USTAT_H
 # include <sys/ustat.h>
-#else
+#elif HAVE_LINUX_TYPES_H
+# include <linux/types.h>
 struct ustat {
-	daddr_t f_tfree;
+	__kernel_daddr_t f_tfree;
 	ino_t f_tinode;
 	char f_fname[6];
 	char f_fpack[6];
testcases/kernel/syscalls/ustat/ustat01.c +7 -1
index 2e7dcc9d7..0252858be 100644
@@ -5,6 +5,10 @@
  * Check that ustat() succeeds given correct parameters.
  */
 
+#include "config.h"
+#include "tst_test.h"
+
+#if defined(HAVE_SYS_USTAT_H) || defined(HAVE_LINUX_TYPES_H)
 #include <unistd.h>
 #include <errno.h>
 #include <sys/types.h>
@@ -12,7 +16,6 @@
 
 #include "lapi/syscalls.h"
 #include "lapi/ustat.h"
-#include "tst_test.h"
 
 static dev_t dev_num;
 
@@ -42,3 +45,6 @@ static struct tst_test test = {
 	.test_all = run,
 	.setup = setup,
 };
+#else
+TST_TEST_TCONF("testing ustat requires <sys/ustat.h> or <linux/types.h>");
+#endif
testcases/kernel/syscalls/ustat/ustat02.c +7 -1
index 9bbe4f3f5..4c171f7b8 100644
@@ -6,6 +6,10 @@
  * invalid dev_t parameter and for bad address paramater.
  */
 
+#include "config.h"
+#include "tst_test.h"
+
+#if defined(HAVE_SYS_USTAT_H) || defined(HAVE_LINUX_TYPES_H)
 #include <unistd.h>
 #include <errno.h>
 #include <sys/stat.h>
@@ -13,7 +17,6 @@
 
 #include "lapi/syscalls.h"
 #include "lapi/ustat.h"
-#include "tst_test.h"
 
 static dev_t invalid_dev = -1;
 static dev_t root_dev;
@@ -61,3 +64,6 @@ static struct tst_test test = {
 	.setup = setup,
 	.tcnt = ARRAY_SIZE(tc),
 };
+#else
+TST_TEST_TCONF("testing ustat requires <sys/ustat.h> or <linux/types.h>");
+#endif
Signature · 1 line
2.20.1
author
#1 lore raw

PATCH 1/1 syscalls/ustat: Use __kernel_daddr_t instead of daddr_t

Hi,

Quoted 7 lines

This fixes build for musl, which does not have <sys/ustat.h> nor typedef daddr_t. Musl does not define these as ustat is deprecated in favor of statfs() and daddr_t is ancient [1].

Other libc also does not have <sys/ustat.h> (ustat support was removed from glibc in 2.28), but they have daddr_t.

Quoted 6 lines

This brings <linux/types.h> dependency for ustat testing.

Typedef is needed as sparc and mips have __kernel_daddr_t long, otherwise it's int.

[1] https://www.openwall.com/lists/musl/2019/06/10/1

Ping, please.

Kind regards,
Petr

#2 lore raw

PATCH 1/1 syscalls/ustat: Use __kernel_daddr_t instead of daddr_t

Hi!

Quoted 7 lines
Quoted 6 lines

This fixes build for musl, which does not have <sys/ustat.h> nor typedef daddr_t. Musl does not define these as ustat is deprecated in favor of statfs() and daddr_t is ancient [1].

Other libc also does not have <sys/ustat.h> (ustat support was removed from glibc in 2.28), but they have daddr_t.

Quoted 6 lines
Quoted 6 lines

This brings <linux/types.h> dependency for ustat testing.

Typedef is needed as sparc and mips have __kernel_daddr_t long, otherwise it's int.

[1] https://www.openwall.com/lists/musl/2019/06/10/1

I'm always worried when we use types starting with double underscores but let's apply this one, we can always revert it in a case that it breaks build somewhere.

Signature · 2 lines
Cyril Hrubis
[email protected]
#3 lore raw

PATCH 1/1 syscalls/ustat: Use __kernel_daddr_t instead of daddr_t

Hi Cyril,

Quoted 9 lines
Quoted 8 lines
Quoted 8 lines

This fixes build for musl, which does not have <sys/ustat.h> nor typedef daddr_t. Musl does not define these as ustat is deprecated in favor of statfs() and daddr_t is ancient [1].

Other libc also does not have <sys/ustat.h> (ustat support was removed from glibc in 2.28), but they have daddr_t.

This brings <linux/types.h> dependency for ustat testing.

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

Typedef is needed as sparc and mips have __kernel_daddr_t long, otherwise it's int.

[1] https://www.openwall.com/lists/musl/2019/06/10/1

I'm always worried when we use types starting with double underscores but let's apply this one, we can always revert it in a case that it breaks build somewhere.

Thanks, merged.

BTW <linux/types.h> is also checked in m4/ltp-signalfd.m4 so there is duplicity in check, but I gave up cleanup it now.

Kind regards,
Petr

#4 lore raw

PATCH 1/1 syscalls/ustat: Use __kernel_daddr_t instead of daddr_t

In reply to Petr Vorel

Hi Petr,

Quoted 7 lines · Petr Vorel

This fixes build for musl, which does not have <sys/ustat.h> nor typedef daddr_t. Musl does not define these as ustat is deprecated in favor of statfs() and daddr_t is ancient [1].

Other libc also does not have <sys/ustat.h> (ustat support was removed from glibc in 2.28), but they have daddr_t.

Quoted 9 lines

This brings <linux/types.h> dependency for ustat testing.

Typedef is needed as sparc and mips have __kernel_daddr_t long, otherwise it's int.

[1] https://www.openwall.com/lists/musl/2019/06/10/1

I don't have a try to build this with musl, but from my code review, the patch makes sense.

Signature · 5 lines
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20190718/284ee6d4/attachment.htm>