Fix redefinition errors for toolstack libs
cover Fix redefinition errors for toolstack libs
For replication I used gcc 10.3 on an Alpine system. In order to replicate the redefinition error for PAGE_SIZE one should install the 'fortify-headers' package which will change the chain of included headers by indirectly including /usr/include/limits.h where PAGE_SIZE and PATH_MAX are defined.
Costin Lupu (5):
tools/debugger: Fix PAGE_SIZE redefinition error tools/libfsimage: Fix PATH_MAX redefinition error tools/libs/foreignmemory: Fix PAGE_SIZE redefinition error tools/libs/gnttab: Fix PAGE_SIZE redefinition error tools/ocaml: Fix redefinition errors
tools/debugger/kdd/kdd-xen.c | 4 ++++ tools/debugger/kdd/kdd.c | 4 ++++ tools/libfsimage/ext2fs/fsys_ext2fs.c | 2 ++ tools/libfsimage/reiserfs/fsys_reiserfs.c | 2 ++ tools/libs/foreignmemory/private.h | 6 ++++-- tools/libs/gnttab/linux.c | 6 ++++++ tools/ocaml/libs/xc/xenctrl_stubs.c | 8 ++++++++ tools/ocaml/libs/xentoollog/xentoollog_stubs.c | 4 ++++ tools/ocaml/libs/xl/xenlight_stubs.c | 4 ++++ 9 files changed, 38 insertions(+), 2 deletions(-)
Signature · 1 line
2.20.1
PATCH 1/5 tools/debugger: Fix PAGE_SIZE redefinition error
If PAGE_SIZE is already defined in the system (e.g. in /usr/include/limits.h header) then gcc will trigger a redefinition error because of -Werror. This commit also protects PAGE_SHIFT definitions for keeping consistency.
--- tools/debugger/kdd/kdd-xen.c | 4 ++++ tools/debugger/kdd/kdd.c | 4 ++++ 2 files changed, 8 insertions(+)
tools/debugger/kdd/kdd-xen.c +4 -0
@@ -48,8 +48,12 @@#define MAPSIZE 4093/* Prime */ +#ifndef PAGE_SHIFT #define PAGE_SHIFT 12 +#endif +#ifndef PAGE_SIZE #define PAGE_SIZE (1U << PAGE_SHIFT) +#endif struct kdd_guest {struct xentoollog_logger xc_log;/* Must be first for xc log callbacks */
tools/debugger/kdd/kdd.c +4 -0
@@ -288,8 +288,12 @@ static void kdd_log_pkt(kdd_state *s, const char *name, kdd_pkt *p)* Memory access: virtual addresses and syntactic sugar. */ +#ifndef PAGE_SHIFT #define PAGE_SHIFT (12) +#endif +#ifndef PAGE_SIZE #define PAGE_SIZE (1ULL << PAGE_SHIFT) +#endif static uint32_t kdd_read_physical (kdd_state *s,uint64_t addr,uint32_t len,void *buf)
Signature · 1 line
2.20.1
Re: [PATCH 1/5] tools/debugger: Fix PAGE_SIZE redefinition error
Hi,
Quoted 4 lines · Costin Lupu
If
PAGE_SIZEis already defined in the system (e.g. in/usr/include/limits.hheader) then gcc will trigger a redefinition error because of -Werror. This commit also protectsPAGE_SHIFTdefinitions for keeping consistency.
Thanks for looking into this! I think properly speaking we should fix this by defining KDD_PAGE_SHIFT and KDD_PAGE_SIZE in kdd.h and using those everywhere we currently use PAGE_SIZE/PAGE_SHIFT. in kdd.c and kdd-xen.c. If for some reason we ever ended up with a system-defined PAGE_SIZE that wasn't 4096u then we would not want to use it here because it would break our guest operations.
Cheers,
Tim
Re: [PATCH 1/5] tools/debugger: Fix PAGE_SIZE redefinition error
Hi Tim,
Quoted 8 lines · Tim Deegan
Hi,
Quoted 4 lines · Costin Lupu
If
PAGE_SIZEis already defined in the system (e.g. in/usr/include/limits.hheader) then gcc will trigger a redefinition error because of -Werror. This commit also protectsPAGE_SHIFTdefinitions for keeping consistency.
Quoted 6 lines
Thanks for looking into this! I think properly speaking we should fix this by defining
KDD_PAGE_SHIFTandKDD_PAGE_SIZEinkdd.hand using those everywhere we currently use PAGE_SIZE/PAGE_SHIFT. inkdd.candkdd-xen.c. If for some reason we ever ended up with a system-definedPAGE_SIZEthat wasn't 4096u then we would not want to use it here because it would break our guest operations.
As discussed for another patch of the series, an agreed solution that would apply for other libs as well would be to use XC_PAGE_* macros instead of PAGE_* macros. I've just sent a v2 doing that. Please let me know if you think it would be better to introduce the KDD_PAGE_* definitions instead.
Cheers,
Costin
Re: [PATCH 1/5] tools/debugger: Fix PAGE_SIZE redefinition error
Quoted 18 lines · Costin Lupu
Hi Tim,
Quoted 8 lines · Tim Deegan
Hi,
Quoted 4 lines · Costin Lupu
If
PAGE_SIZEis already defined in the system (e.g. in/usr/include/limits.hheader) then gcc will trigger a redefinition error because of -Werror. This commit also protectsPAGE_SHIFTdefinitions for keeping consistency.Quoted 6 lines
Thanks for looking into this! I think properly speaking we should fix this by defining
KDD_PAGE_SHIFTandKDD_PAGE_SIZEinkdd.hand using those everywhere we currently use PAGE_SIZE/PAGE_SHIFT. inkdd.candkdd-xen.c. If for some reason we ever ended up with a system-definedPAGE_SIZEthat wasn't 4096u then we would not want to use it here because it would break our guest operations.
Quoted 5 lines
As discussed for another patch of the series, an agreed solution that would apply for other libs as well would be to use XC_PAGE_* macros instead of PAGE_* macros. I've just sent a v2 doing that. Please let me know if you think it would be better to introduce the KDD_PAGE_* definitions instead.
Sorry to be annoying, but yes, please do introduce the KDD_ versions. All the xen-specific code in KDD lives in kdd-xen.c; kdd.c shouldn't include any xen headers.
Cheers,
Tim.
Re: [PATCH 1/5] tools/debugger: Fix PAGE_SIZE redefinition error
Quoted 25 lines · Tim Deegan
Quoted 18 lines · Costin Lupu
Hi Tim,
Quoted 8 lines · Tim Deegan
Hi,
Quoted 4 lines · Costin Lupu
If
PAGE_SIZEis already defined in the system (e.g. in/usr/include/limits.hheader) then gcc will trigger a redefinition error because of -Werror. This commit also protectsPAGE_SHIFTdefinitions for keeping consistency.Quoted 6 lines
Thanks for looking into this! I think properly speaking we should fix this by defining
KDD_PAGE_SHIFTandKDD_PAGE_SIZEinkdd.hand using those everywhere we currently use PAGE_SIZE/PAGE_SHIFT. inkdd.candkdd-xen.c. If for some reason we ever ended up with a system-definedPAGE_SIZEthat wasn't 4096u then we would not want to use it here because it would break our guest operations.Quoted 5 lines
As discussed for another patch of the series, an agreed solution that would apply for other libs as well would be to use XC_PAGE_* macros instead of PAGE_* macros. I've just sent a v2 doing that. Please let me know if you think it would be better to introduce the KDD_PAGE_* definitions instead.
Quoted 3 lines
Sorry to be annoying, but yes, please do introduce the
KDD_versions. All the xen-specific code in KDD lives inkdd-xen.c;kdd.cshouldn't include any xen headers.
No worries, will do. I imagined that might be the case for kdd.c, but I wasn't sure.
Cheers,
Costin
PATCH 2/5 tools/libfsimage: Fix PATH_MAX redefinition error
In reply to Costin LupuIf PATH_MAX is already defined in the system (e.g. in /usr/include/limits.h header) then gcc will trigger a redefinition error because of -Werror.
--- tools/libfsimage/ext2fs/fsys_ext2fs.c | 2 ++ tools/libfsimage/reiserfs/fsys_reiserfs.c | 2 ++ 2 files changed, 4 insertions(+)
tools/libfsimage/ext2fs/fsys_ext2fs.c +2 -0
@@ -278,7 +278,9 @@ struct ext4_extent_header {#define EXT2_SUPER_MAGIC 0xEF53/* include/linux/ext2_fs.h */ #define EXT2_ROOT_INO 2/* include/linux/ext2_fs.h */ +#ifndef PATH_MAX #define PATH_MAX 1024/* include/linux/limits.h */ +#endif #define MAX_LINK_COUNT 5/* number of symbolic links to follow */ /* made up, these are pointers into FSYS_BUF */
tools/libfsimage/reiserfs/fsys_reiserfs.c +2 -0
@@ -284,7 +284,9 @@ struct reiserfs_de_head#define S_ISDIR (mode) (((mode) & 0170000) == 0040000)#define S_ISLNK (mode) (((mode) & 0170000) == 0120000) +#ifndef PATH_MAX #define PATH_MAX 1024/* include/linux/limits.h */ +#endif #define MAX_LINK_COUNT 5/* number of symbolic links to follow */ /* The size of the node cache */
Signature · 1 line
2.20.1
Re: [PATCH 2/5] tools/libfsimage: Fix PATH_MAX redefinition error
Quoted 14 lines · Costin Lupu
If
PATH_MAXis already defined in the system (e.g. in/usr/include/limits.hheader) then gcc will trigger a redefinition error because of -Werror.Signed-off-by Costin Lupu <[email protected]>--- tools/libfsimage/ext2fs/fsys_ext2fs.c | 2 ++ tools/libfsimage/reiserfs/fsys_reiserfs.c | 2 ++ 2 files changed, 4 insertions(+)tools/libfsimage/ext2fs/fsys_ext2fs.c +0
-0
tools/libfsimage/ext2fs/fsys_ext2fs.c struct ext4_extent_header { · line 283
Quoted 6 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c +2
-0@@ -278,7 +278,9 @@ struct ext4_extent_header {#define EXT2_SUPER_MAGIC 0xEF53/* include/linux/ext2_fs.h */ #define EXT2_ROOT_INO 2/* include/linux/ext2_fs.h */ +#ifndef PATH_MAX #define PATH_MAX 1024/* include/linux/limits.h */ +#endif
Can we drop it completely and just rely on limits.h?
Quoted 18 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c +0
-0#define MAX_LINK_COUNT 5/* number of symbolic links to follow */ /* made up, these are pointers into FSYS_BUF */ tools/libfsimage/reiserfs/fsys_reiserfs.c +2
-0@@ -284,7 +284,9 @@ struct reiserfs_de_head#define S_ISDIR (mode) (((mode) & 0170000) == 0040000)#define S_ISLNK (mode) (((mode) & 0170000) == 0120000) +#ifndef PATH_MAX #define PATH_MAX 1024/* include/linux/limits.h */ +#endif #define MAX_LINK_COUNT 5/* number of symbolic links to follow */ /* The size of the node cache */
Signature · 1 line
Julien Grall
Re: [PATCH 2/5] tools/libfsimage: Fix PATH_MAX redefinition error
Quoted 13 lines · Julien Grall
Quoted 10 lines · Costin Lupu
If
PATH_MAXis already defined in the system (e.g. in/usr/include/limits.h
header) then gcc will trigger a redefinition error because of -Werror.Signed-off-by Costin Lupu <[email protected]>---
tools/libfsimage/ext2fs/fsys_ext2fs.c | 2 ++tools/libfsimage/reiserfs/fsys_reiserfs.c| 2 ++2 files changed, 4 insertions(+)
Quoted 14 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c struct ext4_extent_header { · line 282
Quoted 11 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c +2
-0@@ -278,7 +278,9 @@ struct ext4_extent_header { #define EXT2_SUPER_MAGIC 0xEF53 /* include/linux/ext2_fs.h */ #define EXT2_ROOT_INO 2 /* include/linux/ext2_fs.h */ +#ifndef PATH_MAX #define PATH_MAX 1024 /* include/linux/limits.h */ +#endif Can we drop it completely and just rely on
limits.h?
One problem here is that the system limits.h header doesn't necessarily include linux/limits.h, which would mean we would have to include linux/limits.h. But this is problematic for other systems such as BSD.
I had a look on a FreeBSD source tree to see how this is done there. It seems that there are lots of submodules, apps and libs that redefine PATH_MAX in case it wasn't defined before so the changes introduced by the current patch seem to be very popular. Another clean approach I saw was for jemalloc [1] which includes unistd.h. They redefine PATH_MAX only for MS C compiler, but AFAIK we don't need that.
So IMHO the current changes seem to be the most portable, but I'm open to any suggestions.
Quoted 18 lines
Quoted 17 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c +0
-0#define MAX_LINK_COUNT 5 /* number of symbolic linksto follow */
/* made up, these are pointers intoFSYS_BUF*/tools/libfsimage/reiserfs/fsys_reiserfs.c +0
-0@@ -284,7 +284,9 @@ struct reiserfs_de_head #define S_ISDIR(mode) (((mode) & 0170000) == 0040000) #define S_ISLNK(mode) (((mode) & 0170000) == 0120000)+#ifndef
PATH_MAX
#define PATH_MAX 1024 /*include/linux/limits.h*/ +#endif
#define MAX_LINK_COUNT 5 /* number of symbolic links to follow */ /* The size of the node cache */
Julien Grall
Reviewed-byRe: [PATCH 2/5] tools/libfsimage: Fix PATH_MAX redefinition error
Hi Costin,
Quoted 29 lines · Costin Lupu
Quoted 13 lines · Julien Grall
Quoted 10 lines · Costin Lupu
If
PATH_MAXis already defined in the system (e.g. in/usr/include/limits.h
header) then gcc will trigger a redefinition error because of -Werror.Signed-off-by Costin Lupu <[email protected]>--- tools/libfsimage/ext2fs/fsys_ext2fs.c | 2 ++ tools/libfsimage/reiserfs/fsys_reiserfs.c | 2 ++ 2 files changed, 4 insertions(+)Quoted 14 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c struct ext4_extent_header { · line 282
Quoted 11 lines
tools/libfsimage/ext2fs/fsys_ext2fs.c +2
-0@@ -278,7 +278,9 @@ struct ext4_extent_header { #define EXT2_SUPER_MAGIC 0xEF53 /* include/linux/ext2_fs.h */ #define EXT2_ROOT_INO 2 /* include/linux/ext2_fs.h */ +#ifndef PATH_MAX #define PATH_MAX 1024 /* include/linux/limits.h */ +#endif Can we drop it completely and just rely on
limits.h?
Quoted 3 lines
One problem here is that the system
limits.hheader doesn't necessarily includelinux/limits.h, which would mean we would have to includelinux/limits.h. But this is problematic for other systems such as BSD.
That's annoying :).
Quoted 7 lines
I had a look on a FreeBSD source tree to see how this is done there. It seems that there are lots of submodules, apps and libs that redefine
PATH_MAXin case it wasn't defined before so the changes introduced by the current patch seem to be very popular. Another clean approach I saw was for jemalloc [1] which includesunistd.h. They redefinePATH_MAXonly for MS C compiler, but AFAIK we don't need that.
I am not aware of anyone using MS C compiler to build the tools.
Quoted 3 lines
So IMHO the current changes seem to be the most portable, but I'm open to any suggestions.
Right, this is the good thing of your approach. I can't see a better solution if the system limits.h doesn't always define PATH_MAX. So:
Cheers,
Signature · 1 line
Julien Grall
PATCH 3/5 tools/libs/foreignmemory: Fix PAGE_SIZE redefinition error
In reply to Costin LupuIf PAGE_SIZE is already defined in the system (e.g. in /usr/include/limits.h header) then gcc will trigger a redefinition error because of -Werror. This commit also protects PAGE_SHIFT and PAGE_MASK definitions for keeping consistency.
--- tools/libs/foreignmemory/private.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
tools/libs/foreignmemory/private.h +4 -2
@@ -10,11 +10,13 @@#include <xen/xen.h> #include <xen/sys/privcmd.h> -#ifndef PAGE_SHIFT /* Mini-os, Yukk */ +#ifndef PAGE_SHIFT #define PAGE_SHIFT 12#endif -#ifndef __MINIOS__/* Yukk */ +#ifndef PAGE_SIZE #define PAGE_SIZE (1UL << PAGE_SHIFT) +#endif +#ifndef PAGE_MASK #define PAGE_MASK (~(PAGE_SIZE-1))#endif
Signature · 1 line
2.20.1
Re: [PATCH 3/5] tools/libs/foreignmemory: Fix PAGE_SIZE redefinition error
Hi Costin,
Quoted 12 lines · Costin Lupu
tools/libs/foreignmemory/private.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)tools/libs/foreignmemory/private.h +0
-1@@ -10,11 +10,13 @@#include <xen/xen.h> #include <xen/sys/privcmd.h> -#ifndef PAGE_SHIFT /* Mini-os, Yukk */
tools/libs/foreignmemory/private.h line 21
Quoted 10 lines
tools/libs/foreignmemory/private.h +4
-1+#ifndef PAGE_SHIFT #define PAGE_SHIFT 12#endif -#ifndef __MINIOS__/* Yukk */ +#ifndef PAGE_SIZE #define PAGE_SIZE (1UL << PAGE_SHIFT) +#endif +#ifndef PAGE_MASK #define PAGE_MASK (~(PAGE_SIZE-1))#endif
Looking at the usage, I believe PAGE_* are referring to the page granularity of Xen rather than the page granularity of the control domain itself.
So it would be incorrect to use the domain's page granularity here and would break dom0 using 64KB page granularity on Arm.
Instead, we should replace PAGE_* with XC_PAGE_*. If some of them are still referring to the control domain granularity, then we should use getpageshift() (Or equivalent) because the userspace shouldn't rely on a specific page granularity.
Cheers,
Signature · 1 line
Julien Grall
Re: [PATCH 3/5] tools/libs/foreignmemory: Fix PAGE_SIZE redefinition error
Hi Julien,
Quoted 30 lines · Julien Grall
Hi Costin,
Quoted 20 lines · Costin Lupu
tools/libs/foreignmemory/private.h| 6 ++++--1 file changed, 4 insertions(+), 2 deletions(-)tools/libs/foreignmemory/private.h +0
-0@@ -10,11 +10,13 @@ #include <xen/xen.h> #include <xen/sys/privcmd.h>-#ifndef
PAGE_SHIFT/* Mini-os, Yukk */
+#ifndefPAGE_SHIFT
#define PAGE_SHIFT 12
#endiftools/libs/foreignmemory/private.h +1
-1-#ifndef __MINIOS__/* Yukk */ +#ifndef PAGE_SIZE #define PAGE_SIZE (1UL <<
PAGE_SHIFT)tools/libs/foreignmemory/private.h +2
-0+#endif +#ifndef PAGE_MASK Quoted 2 lines
#define PAGE_MASK (~(PAGE_SIZE-1))
#endifLooking at the usage, I believe PAGE_* are referring to the page granularity of Xen rather than the page granularity of the control domain itself.
Quoted 7 lines
So it would be incorrect to use the domain's page granularity here and would break dom0 using 64KB page granularity on Arm.
Instead, we should replace PAGE_* with XC_PAGE_*. If some of them are still referring to the control domain granularity, then we should use
getpageshift()(Or equivalent) because the userspace shouldn't rely on a specific page granularity.
Yes, this makes sense. One thing I need to clarify: what does XC stand for? I thought for some time it stands for Xen Control.
Thanks,
Costin
Re: [PATCH 3/5] tools/libs/foreignmemory: Fix PAGE_SIZE redefinition error
Quoted 1 line · Costin Lupu
Hi Julien,
Hi Costin,
Quoted 34 lines · Julien Grall
Quoted 33 lines
Hi Costin,
Quoted 12 lines · Costin Lupu
tools/libs/foreignmemory/private.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)tools/libs/foreignmemory/private.h +0
-0@@ -10,11 +10,13 @@ #include <xen/xen.h> #include <xen/sys/privcmd.h> -#ifndef PAGE_SHIFT /* Mini-os, Yukk */tools/libs/foreignmemory/private.h line 21
Quoted 10 lines
tools/libs/foreignmemory/private.h +4
-1+#ifndef PAGE_SHIFT #define PAGE_SHIFT 12 #endif -#ifndef __MINIOS__/* Yukk */ +#ifndef PAGE_SIZE #define PAGE_SIZE (1UL << PAGE_SHIFT) +#endif +#ifndef PAGE_MASK #define PAGE_MASK (~(PAGE_SIZE-1)) #endifLooking at the usage, I believe PAGE_* are referring to the page granularity of Xen rather than the page granularity of the control domain itself.
So it would be incorrect to use the domain's page granularity here and would break dom0 using 64KB page granularity on Arm.
Quoted 7 lines
Quoted 4 lines
Instead, we should replace PAGE_* with XC_PAGE_*. If some of them are still referring to the control domain granularity, then we should use
getpageshift()(Or equivalent) because the userspace shouldn't rely on a specific page granularity.Yes, this makes sense. One thing I need to clarify: what does XC stand for? I thought for some time it stands for Xen Control.
I think it is Xen Control, which is a bit confusing for that specific define.
Cheers,
Signature · 1 line
Julien Grall
PATCH 4/5 tools/libs/gnttab: Fix PAGE_SIZE redefinition error
In reply to Costin LupuIf PAGE_SIZE is already defined in the system (e.g. in /usr/include/limits.h header) then gcc will trigger a redefinition error because of -Werror. This commit also protects PAGE_SHIFT and PAGE_MASK definitions for keeping consistency.
--- tools/libs/gnttab/linux.c | 6 ++++++ 1 file changed, 6 insertions(+)
tools/libs/gnttab/linux.c +6 -0
@@ -36,9 +36,15 @@#include "private.h" +#ifndef PAGE_SHIFT #define PAGE_SHIFT 12 +#endif +#ifndef PAGE_SIZE #define PAGE_SIZE (1UL << PAGE_SHIFT) +#endif +#ifndef PAGE_MASK #define PAGE_MASK (~(PAGE_SIZE-1)) +#endif #define DEVXEN "/dev/xen/"
Signature · 1 line
2.20.1
PATCH 5/5 tools/ocaml: Fix redefinition errors
In reply to Costin LupuIf PAGE_SIZE is already defined in the system (e.g. in /usr/include/limits.h header) then gcc will trigger a redefinition error because of -Werror. This commit also protects PAGE_SHIFT and PAGE_MASK definitions for keeping consistency.
Same issue applies for redefinitions of Val_none and Some_val macros which can be already define in the OCaml system headers (e.g. /usr/lib/ocaml/caml/mlvalues.h).
--- tools/ocaml/libs/xc/xenctrl_stubs.c | 8 ++++++++ tools/ocaml/libs/xentoollog/xentoollog_stubs.c | 4 ++++ tools/ocaml/libs/xl/xenlight_stubs.c | 4 ++++ 3 files changed, 16 insertions(+)
tools/ocaml/libs/xc/xenctrl_stubs.c +8 -0
@@ -36,14 +36,22 @@#include "mmap_stubs.h" +#ifndef PAGE_SHIFT #define PAGE_SHIFT 12 +#endif +#ifndef PAGE_SIZE #define PAGE_SIZE (1UL << PAGE_SHIFT) +#endif +#ifndef PAGE_MASK #define PAGE_MASK (~(PAGE_SIZE-1)) +#endif #define _H (__h) ((xc_interface *)(__h))#define _D (__d) ((uint32_t)Int_val(__d)) +#ifndef Val_none#define Val_none (Val_int(0)) +#endif #define string_of_option_array (array, index) \ ((Field(array, index) == Val_none) ? NULL : String_val(Field(Field(array, index), 0)))
tools/ocaml/libs/xentoollog/xentoollog_stubs.c +4 -0
@@ -53,8 +53,12 @@ static char * dup_String_val(value s)#include "_xtl_levels.inc" /* Option type support as per http://www.linux-nantes.org/~fmonnier/ocaml/ocaml-wrapping-c.php */ +#ifndef Val_none#define Val_none Val_int(0) +#endif +#ifndef Some_val#define Some_val (v) Field(v,0) +#endif static value Val_some (value v) {
tools/ocaml/libs/xl/xenlight_stubs.c +4 -0
@@ -227,8 +227,12 @@ static value Val_string_list(libxl_string_list *c_val) }/* Option type support as per http://www.linux-nantes.org/~fmonnier/ocaml/ocaml-wrapping-c.php */ +#ifndef Val_none#define Val_none Val_int(0) +#endif +#ifndef Some_val#define Some_val (v) Field(v,0) +#endif static value Val_some (value v) {
Signature · 1 line
2.20.1
On 27 Apr 2021, at 13:05, Costin Lupu <[email protected]<mailto:[email protected]>> wrote:
For replication I used gcc 10.3 on an Alpine system. In order to replicate the redefinition error for PAGE_SIZE one should install the 'fortify-headers' package which will change the chain of included headers by indirectly including /usr/include/limits.h where PAGE_SIZE and PATH_MAX are defined.
Costin Lupu (5):
tools/debugger: Fix PAGE_SIZE redefinition error tools/libfsimage: Fix PATH_MAX redefinition error tools/libs/foreignmemory: Fix PAGE_SIZE redefinition error tools/libs/gnttab: Fix PAGE_SIZE redefinition error tools/ocaml: Fix redefinition errors
tools/debugger/kdd/kdd-xen.c | 4 ++++ tools/debugger/kdd/kdd.c | 4 ++++ tools/libfsimage/ext2fs/fsys_ext2fs.c | 2 ++ tools/libfsimage/reiserfs/fsys_reiserfs.c | 2 ++ tools/libs/foreignmemory/private.h | 6 ++++-- tools/libs/gnttab/linux.c | 6 ++++++ tools/ocaml/libs/xc/xenctrl_stubs.c | 8 ++++++++ tools/ocaml/libs/xentoollog/xentoollog_stubs.c | 4 ++++ tools/ocaml/libs/xl/xenlight_stubs.c | 4 ++++ 9 files changed, 38 insertions(+), 2 deletions(-)
—
2.20.1
For the OCaml bindings, this avoids redefinitions as you say. Looks good to me.
Quoted 34 lines · Christian Lindig
Quoted 28 lines
On 27 Apr 2021, at 13:05, Costin Lupu <[email protected] <mailto:[email protected]>> wrote:
For replication I used gcc 10.3 on an Alpine system. In order to replicate the
redefinition error forPAGE_SIZEone should install the 'fortify-headers' package which will change the chain of included headers by indirectly including/usr/include/limits.hwherePAGE_SIZEandPATH_MAXare defined.Costin Lupu (5):
tools/debugger: FixPAGE_SIZEredefinition errortools/libfsimage: FixPATH_MAXredefinition errortools/libs/foreignmemory: FixPAGE_SIZEredefinition errortools/libs/gnttab: FixPAGE_SIZEredefinition errortools/ocaml: Fix redefinition errors
tools/debugger/kdd/kdd-xen.c| 4 ++++tools/debugger/kdd/kdd.c| 4 ++++tools/libfsimage/ext2fs/fsys_ext2fs.c| 2 ++tools/libfsimage/reiserfs/fsys_reiserfs.c| 2 ++tools/libs/foreignmemory/private.h| 6 ++++--tools/libs/gnttab/linux.c| 6 ++++++tools/ocaml/libs/xc/xenctrl_stubs.c| 8 ++++++++tools/ocaml/libs/xentoollog/xentoollog_stubs.c| 4 ++++tools/ocaml/libs/xl/xenlight_stubs.c| 4 ++++Quoted 5 lines
9 files changed, 38 insertions(+), 2 deletions(-)—
2.20.1
Quoted 6 lines
For the OCaml bindings, this avoids redefinitions as you say. Looks good to me.
Acked-by Christian Lindig <[email protected]<mailto:[email protected]>>
Thanks, Christian, I'll add your ack on the Ocaml patch for the next version of the series.
Cheers,
Costin