* [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s
@ 2018-07-13 12:34 Shivaprasad G Bhat
2018-07-13 12:37 ` Laurent Vivier
2018-07-14 1:07 ` David Gibson
0 siblings, 2 replies; 5+ messages in thread
From: Shivaprasad G Bhat @ 2018-07-13 12:34 UTC (permalink / raw)
To: riku.voipio, laurent; +Cc: qemu-devel, david
Qemu includes the glibc headers for the host defines and target headers are
part of the qemu source themselves. The glibc has the F_GETLK64, F_SETLK64
and F_SETLKW64 defined to 12, 13 and 14 for all archs in
sysdeps/unix/sysv/linux/bits/fcntl-linux.h. The linux kernel generic
definition for F_*LK is 5, 6 & 7 and F_*LK64* is 12,13, and 14 as seen in
include/uapi/asm-generic/fcntl.h. On 64bit machine, by default the kernel
assumes all F_*LK to 64bit calls and doesnt support use of F_*LK64* as
can be seen in include/linux/fcntl.h in linux source.
On x86_64 host, the values for F_*LK64* are set to 5, 6 and 7
explicitly in /usr/include/x86_64-linux-gnu/bits/fcntl.h by the glibc.
Whereas, a PPC64 host doesn't have such a definition in
/usr/include/powerpc64le-linux-gnu/bits/fcntl.h by the glibc. So,
the sources on PPC64 host sees the default value of F_*LK64*
as 12, 13 & 14(fcntl-linux.h).
Since the 64bit kernel doesnt support 12, 13 & 14; the glibc fcntl syscall
implementation(__libc_fcntl*(), __fcntl64_nocancel) does the F_*LK64* value
convertion back to F_*LK* values on PPC64 as seen in
sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h with FCNTL_ADJUST_CMD()
macro. Whereas on x86_64 host the values for F_*LK64* are set to 5, 6 and 7
and no adjustments are needed.
Since qemu doesnt use the glibc fcntl, but makes the safe_syscall* on its
own, the PPC64 qemu is calling the syscall with 12, 13, and 14(without
adjustment) and they all fail. The fcntl calls to F_GETLK/F_SETLK|W all
fail by all pplications run on PPC64 host user emulation.
The fix here could be to see why on PPC64 the glibc is still keeping
F_*LK64* different from F_*LK and why adjusting them to 5, 6 and 7 before
the syscall for PPC only. See if we can make the
/usr/include/powerpc64le-linux-gnu/bits/fcntl.h to have the values
5, 6 & 7 just like x86_64 and remove the adjustment code in glibc. That
way, qemu sources see the kernel supported values in glibc headers.
OR
On PPC64 host, qemu sources see both F_*LK & F_*LK64* as same and set to
12, 13 and 14 because __USE_FILE_OFFSET64 is defined in qemu
sources(also refer sysdeps/unix/sysv/linux/bits/fcntl-linux.h).
Do the value adjustment just like it is done by glibc source by using
F_GETLK value of 5. That way, we make the syscalls with the actual
supported values in Qemu. The patch is taking this approach.
Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
---
v3 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02923.html
Changes from v3:
- Fixed the tabs for case statements
- Addressed the comments on v3 wrt to the variable initialisation
and break from default case.
v2 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02920.html
Changes from v2:
- Fixed the braces, and indentation for comments.
v1 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02567.html
Changes from v1:
- Changed the overwrite of F*LK64* with 5, 6 and 7 in using #define
instead using the adjustment code similar to glibc as suggested.
- Dropped __linux__ check for the adjustment code as suggested.
- Moved the adjustment code inside target_to_host_fcntl_cmd to address
all possible|future cases.
linux-user/syscall.c | 126 ++++++++++++++++++++++++++++++++------------------
1 file changed, 80 insertions(+), 46 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 643b8833de..b5274f657a 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -6475,63 +6475,97 @@ static int do_fork(CPUArchState *env, unsigned int flags, abi_ulong newsp,
/* warning : doesn't handle linux specific flags... */
static int target_to_host_fcntl_cmd(int cmd)
{
+ int ret;
+
switch(cmd) {
- case TARGET_F_DUPFD:
- case TARGET_F_GETFD:
- case TARGET_F_SETFD:
- case TARGET_F_GETFL:
- case TARGET_F_SETFL:
- return cmd;
- case TARGET_F_GETLK:
- return F_GETLK64;
- case TARGET_F_SETLK:
- return F_SETLK64;
- case TARGET_F_SETLKW:
- return F_SETLKW64;
- case TARGET_F_GETOWN:
- return F_GETOWN;
- case TARGET_F_SETOWN:
- return F_SETOWN;
- case TARGET_F_GETSIG:
- return F_GETSIG;
- case TARGET_F_SETSIG:
- return F_SETSIG;
+ case TARGET_F_DUPFD:
+ case TARGET_F_GETFD:
+ case TARGET_F_SETFD:
+ case TARGET_F_GETFL:
+ case TARGET_F_SETFL:
+ ret = cmd;
+ break;
+ case TARGET_F_GETLK:
+ ret = F_GETLK64;
+ break;
+ case TARGET_F_SETLK:
+ ret = F_SETLK64;
+ break;
+ case TARGET_F_SETLKW:
+ ret = F_SETLKW64;
+ break;
+ case TARGET_F_GETOWN:
+ ret = F_GETOWN;
+ break;
+ case TARGET_F_SETOWN:
+ ret = F_SETOWN;
+ break;
+ case TARGET_F_GETSIG:
+ ret = F_GETSIG;
+ break;
+ case TARGET_F_SETSIG:
+ ret = F_SETSIG;
+ break;
#if TARGET_ABI_BITS == 32
- case TARGET_F_GETLK64:
- return F_GETLK64;
- case TARGET_F_SETLK64:
- return F_SETLK64;
- case TARGET_F_SETLKW64:
- return F_SETLKW64;
-#endif
- case TARGET_F_SETLEASE:
- return F_SETLEASE;
- case TARGET_F_GETLEASE:
- return F_GETLEASE;
+ case TARGET_F_GETLK64:
+ ret = F_GETLK64;
+ break;
+ case TARGET_F_SETLK64:
+ ret = F_SETLK64;
+ break;
+ case TARGET_F_SETLKW64:
+ ret = F_SETLKW64;
+ break;
+#endif
+ case TARGET_F_SETLEASE:
+ ret = F_SETLEASE;
+ break;
+ case TARGET_F_GETLEASE:
+ ret = F_GETLEASE;
+ break;
#ifdef F_DUPFD_CLOEXEC
- case TARGET_F_DUPFD_CLOEXEC:
- return F_DUPFD_CLOEXEC;
+ case TARGET_F_DUPFD_CLOEXEC:
+ ret = F_DUPFD_CLOEXEC;
+ break;
#endif
- case TARGET_F_NOTIFY:
- return F_NOTIFY;
+ case TARGET_F_NOTIFY:
+ ret = F_NOTIFY;
+ break;
#ifdef F_GETOWN_EX
- case TARGET_F_GETOWN_EX:
- return F_GETOWN_EX;
+ case TARGET_F_GETOWN_EX:
+ ret = F_GETOWN_EX;
+ break;
#endif
#ifdef F_SETOWN_EX
- case TARGET_F_SETOWN_EX:
- return F_SETOWN_EX;
+ case TARGET_F_SETOWN_EX:
+ ret = F_SETOWN_EX;
+ break;
#endif
#ifdef F_SETPIPE_SZ
- case TARGET_F_SETPIPE_SZ:
- return F_SETPIPE_SZ;
- case TARGET_F_GETPIPE_SZ:
- return F_GETPIPE_SZ;
+ case TARGET_F_SETPIPE_SZ:
+ ret = F_SETPIPE_SZ;
+ break;
+ case TARGET_F_GETPIPE_SZ:
+ ret = F_GETPIPE_SZ;
+ break;
#endif
- default:
- return -TARGET_EINVAL;
+ default:
+ ret = -TARGET_EINVAL;
+ break;
}
- return -TARGET_EINVAL;
+
+#if defined(__powerpc64__)
+ /* On PPC64, glibc headers has the F_*LK* defined to 12, 13 and 14 and
+ * is not supported by kernel. The glibc fcntl call actually adjusts
+ * them to 5, 6 and 7 before making the syscall(). Since we make the
+ * syscall directly, adjust to what is supported by the kernel.
+ */
+ if (ret >= F_GETLK64 && ret <= F_SETLKW64) {
+ ret -= F_GETLK64 - 5;
+ }
+#endif
+
+ return ret;
}
#define TRANSTBL_CONVERT(a) { -1, TARGET_##a, -1, a }
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s
2018-07-13 12:34 [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s Shivaprasad G Bhat
@ 2018-07-13 12:37 ` Laurent Vivier
2018-07-14 1:07 ` David Gibson
1 sibling, 0 replies; 5+ messages in thread
From: Laurent Vivier @ 2018-07-13 12:37 UTC (permalink / raw)
To: Shivaprasad G Bhat, riku.voipio; +Cc: qemu-devel, david
Le 13/07/2018 à 14:34, Shivaprasad G Bhat a écrit :
> Qemu includes the glibc headers for the host defines and target headers are
> part of the qemu source themselves. The glibc has the F_GETLK64, F_SETLK64
> and F_SETLKW64 defined to 12, 13 and 14 for all archs in
> sysdeps/unix/sysv/linux/bits/fcntl-linux.h. The linux kernel generic
> definition for F_*LK is 5, 6 & 7 and F_*LK64* is 12,13, and 14 as seen in
> include/uapi/asm-generic/fcntl.h. On 64bit machine, by default the kernel
> assumes all F_*LK to 64bit calls and doesnt support use of F_*LK64* as
> can be seen in include/linux/fcntl.h in linux source.
>
> On x86_64 host, the values for F_*LK64* are set to 5, 6 and 7
> explicitly in /usr/include/x86_64-linux-gnu/bits/fcntl.h by the glibc.
> Whereas, a PPC64 host doesn't have such a definition in
> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h by the glibc. So,
> the sources on PPC64 host sees the default value of F_*LK64*
> as 12, 13 & 14(fcntl-linux.h).
>
> Since the 64bit kernel doesnt support 12, 13 & 14; the glibc fcntl syscall
> implementation(__libc_fcntl*(), __fcntl64_nocancel) does the F_*LK64* value
> convertion back to F_*LK* values on PPC64 as seen in
> sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h with FCNTL_ADJUST_CMD()
> macro. Whereas on x86_64 host the values for F_*LK64* are set to 5, 6 and 7
> and no adjustments are needed.
>
> Since qemu doesnt use the glibc fcntl, but makes the safe_syscall* on its
> own, the PPC64 qemu is calling the syscall with 12, 13, and 14(without
> adjustment) and they all fail. The fcntl calls to F_GETLK/F_SETLK|W all
> fail by all pplications run on PPC64 host user emulation.
>
> The fix here could be to see why on PPC64 the glibc is still keeping
> F_*LK64* different from F_*LK and why adjusting them to 5, 6 and 7 before
> the syscall for PPC only. See if we can make the
> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h to have the values
> 5, 6 & 7 just like x86_64 and remove the adjustment code in glibc. That
> way, qemu sources see the kernel supported values in glibc headers.
>
> OR
>
> On PPC64 host, qemu sources see both F_*LK & F_*LK64* as same and set to
> 12, 13 and 14 because __USE_FILE_OFFSET64 is defined in qemu
> sources(also refer sysdeps/unix/sysv/linux/bits/fcntl-linux.h).
> Do the value adjustment just like it is done by glibc source by using
> F_GETLK value of 5. That way, we make the syscalls with the actual
> supported values in Qemu. The patch is taking this approach.
>
> Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
> ---
> v3 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02923.html
> Changes from v3:
> - Fixed the tabs for case statements
> - Addressed the comments on v3 wrt to the variable initialisation
> and break from default case.
> v2 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02920.html
> Changes from v2:
> - Fixed the braces, and indentation for comments.
> v1 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02567.html
> Changes from v1:
> - Changed the overwrite of F*LK64* with 5, 6 and 7 in using #define
> instead using the adjustment code similar to glibc as suggested.
> - Dropped __linux__ check for the adjustment code as suggested.
> - Moved the adjustment code inside target_to_host_fcntl_cmd to address
> all possible|future cases.
>
> linux-user/syscall.c | 126 ++++++++++++++++++++++++++++++++------------------
> 1 file changed, 80 insertions(+), 46 deletions(-)
>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s
2018-07-13 12:34 [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s Shivaprasad G Bhat
2018-07-13 12:37 ` Laurent Vivier
@ 2018-07-14 1:07 ` David Gibson
2018-07-14 5:37 ` Laurent Vivier
1 sibling, 1 reply; 5+ messages in thread
From: David Gibson @ 2018-07-14 1:07 UTC (permalink / raw)
To: Shivaprasad G Bhat; +Cc: riku.voipio, laurent, qemu-devel
[-- Attachment #1: Type: text/plain, Size: 8211 bytes --]
On Fri, Jul 13, 2018 at 07:34:46AM -0500, Shivaprasad G Bhat wrote:
> Qemu includes the glibc headers for the host defines and target headers are
> part of the qemu source themselves. The glibc has the F_GETLK64, F_SETLK64
> and F_SETLKW64 defined to 12, 13 and 14 for all archs in
> sysdeps/unix/sysv/linux/bits/fcntl-linux.h. The linux kernel generic
> definition for F_*LK is 5, 6 & 7 and F_*LK64* is 12,13, and 14 as seen in
> include/uapi/asm-generic/fcntl.h. On 64bit machine, by default the kernel
> assumes all F_*LK to 64bit calls and doesnt support use of F_*LK64* as
> can be seen in include/linux/fcntl.h in linux source.
>
> On x86_64 host, the values for F_*LK64* are set to 5, 6 and 7
> explicitly in /usr/include/x86_64-linux-gnu/bits/fcntl.h by the glibc.
> Whereas, a PPC64 host doesn't have such a definition in
> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h by the glibc. So,
> the sources on PPC64 host sees the default value of F_*LK64*
> as 12, 13 & 14(fcntl-linux.h).
>
> Since the 64bit kernel doesnt support 12, 13 & 14; the glibc fcntl syscall
> implementation(__libc_fcntl*(), __fcntl64_nocancel) does the F_*LK64* value
> convertion back to F_*LK* values on PPC64 as seen in
> sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h with FCNTL_ADJUST_CMD()
> macro. Whereas on x86_64 host the values for F_*LK64* are set to 5, 6 and 7
> and no adjustments are needed.
>
> Since qemu doesnt use the glibc fcntl, but makes the safe_syscall* on its
> own, the PPC64 qemu is calling the syscall with 12, 13, and 14(without
> adjustment) and they all fail. The fcntl calls to F_GETLK/F_SETLK|W all
> fail by all pplications run on PPC64 host user emulation.
>
> The fix here could be to see why on PPC64 the glibc is still keeping
> F_*LK64* different from F_*LK and why adjusting them to 5, 6 and 7 before
> the syscall for PPC only. See if we can make the
> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h to have the values
> 5, 6 & 7 just like x86_64 and remove the adjustment code in glibc. That
> way, qemu sources see the kernel supported values in glibc headers.
>
> OR
>
> On PPC64 host, qemu sources see both F_*LK & F_*LK64* as same and set to
> 12, 13 and 14 because __USE_FILE_OFFSET64 is defined in qemu
> sources(also refer sysdeps/unix/sysv/linux/bits/fcntl-linux.h).
> Do the value adjustment just like it is done by glibc source by using
> F_GETLK value of 5. That way, we make the syscalls with the actual
> supported values in Qemu. The patch is taking this approach.
>
> Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
I'm not sure if this should go in through my tree or not.
> ---
> v3 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02923.html
> Changes from v3:
> - Fixed the tabs for case statements
> - Addressed the comments on v3 wrt to the variable initialisation
> and break from default case.
> v2 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02920.html
> Changes from v2:
> - Fixed the braces, and indentation for comments.
> v1 - https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg02567.html
> Changes from v1:
> - Changed the overwrite of F*LK64* with 5, 6 and 7 in using #define
> instead using the adjustment code similar to glibc as suggested.
> - Dropped __linux__ check for the adjustment code as suggested.
> - Moved the adjustment code inside target_to_host_fcntl_cmd to address
> all possible|future cases.
>
> linux-user/syscall.c | 126 ++++++++++++++++++++++++++++++++------------------
> 1 file changed, 80 insertions(+), 46 deletions(-)
>
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index 643b8833de..b5274f657a 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -6475,63 +6475,97 @@ static int do_fork(CPUArchState *env, unsigned int flags, abi_ulong newsp,
> /* warning : doesn't handle linux specific flags... */
> static int target_to_host_fcntl_cmd(int cmd)
> {
> + int ret;
> +
> switch(cmd) {
> - case TARGET_F_DUPFD:
> - case TARGET_F_GETFD:
> - case TARGET_F_SETFD:
> - case TARGET_F_GETFL:
> - case TARGET_F_SETFL:
> - return cmd;
> - case TARGET_F_GETLK:
> - return F_GETLK64;
> - case TARGET_F_SETLK:
> - return F_SETLK64;
> - case TARGET_F_SETLKW:
> - return F_SETLKW64;
> - case TARGET_F_GETOWN:
> - return F_GETOWN;
> - case TARGET_F_SETOWN:
> - return F_SETOWN;
> - case TARGET_F_GETSIG:
> - return F_GETSIG;
> - case TARGET_F_SETSIG:
> - return F_SETSIG;
> + case TARGET_F_DUPFD:
> + case TARGET_F_GETFD:
> + case TARGET_F_SETFD:
> + case TARGET_F_GETFL:
> + case TARGET_F_SETFL:
> + ret = cmd;
> + break;
> + case TARGET_F_GETLK:
> + ret = F_GETLK64;
> + break;
> + case TARGET_F_SETLK:
> + ret = F_SETLK64;
> + break;
> + case TARGET_F_SETLKW:
> + ret = F_SETLKW64;
> + break;
> + case TARGET_F_GETOWN:
> + ret = F_GETOWN;
> + break;
> + case TARGET_F_SETOWN:
> + ret = F_SETOWN;
> + break;
> + case TARGET_F_GETSIG:
> + ret = F_GETSIG;
> + break;
> + case TARGET_F_SETSIG:
> + ret = F_SETSIG;
> + break;
> #if TARGET_ABI_BITS == 32
> - case TARGET_F_GETLK64:
> - return F_GETLK64;
> - case TARGET_F_SETLK64:
> - return F_SETLK64;
> - case TARGET_F_SETLKW64:
> - return F_SETLKW64;
> -#endif
> - case TARGET_F_SETLEASE:
> - return F_SETLEASE;
> - case TARGET_F_GETLEASE:
> - return F_GETLEASE;
> + case TARGET_F_GETLK64:
> + ret = F_GETLK64;
> + break;
> + case TARGET_F_SETLK64:
> + ret = F_SETLK64;
> + break;
> + case TARGET_F_SETLKW64:
> + ret = F_SETLKW64;
> + break;
> +#endif
> + case TARGET_F_SETLEASE:
> + ret = F_SETLEASE;
> + break;
> + case TARGET_F_GETLEASE:
> + ret = F_GETLEASE;
> + break;
> #ifdef F_DUPFD_CLOEXEC
> - case TARGET_F_DUPFD_CLOEXEC:
> - return F_DUPFD_CLOEXEC;
> + case TARGET_F_DUPFD_CLOEXEC:
> + ret = F_DUPFD_CLOEXEC;
> + break;
> #endif
> - case TARGET_F_NOTIFY:
> - return F_NOTIFY;
> + case TARGET_F_NOTIFY:
> + ret = F_NOTIFY;
> + break;
> #ifdef F_GETOWN_EX
> - case TARGET_F_GETOWN_EX:
> - return F_GETOWN_EX;
> + case TARGET_F_GETOWN_EX:
> + ret = F_GETOWN_EX;
> + break;
> #endif
> #ifdef F_SETOWN_EX
> - case TARGET_F_SETOWN_EX:
> - return F_SETOWN_EX;
> + case TARGET_F_SETOWN_EX:
> + ret = F_SETOWN_EX;
> + break;
> #endif
> #ifdef F_SETPIPE_SZ
> - case TARGET_F_SETPIPE_SZ:
> - return F_SETPIPE_SZ;
> - case TARGET_F_GETPIPE_SZ:
> - return F_GETPIPE_SZ;
> + case TARGET_F_SETPIPE_SZ:
> + ret = F_SETPIPE_SZ;
> + break;
> + case TARGET_F_GETPIPE_SZ:
> + ret = F_GETPIPE_SZ;
> + break;
> #endif
> - default:
> - return -TARGET_EINVAL;
> + default:
> + ret = -TARGET_EINVAL;
> + break;
> }
> - return -TARGET_EINVAL;
> +
> +#if defined(__powerpc64__)
> + /* On PPC64, glibc headers has the F_*LK* defined to 12, 13 and 14 and
> + * is not supported by kernel. The glibc fcntl call actually adjusts
> + * them to 5, 6 and 7 before making the syscall(). Since we make the
> + * syscall directly, adjust to what is supported by the kernel.
> + */
> + if (ret >= F_GETLK64 && ret <= F_SETLKW64) {
> + ret -= F_GETLK64 - 5;
> + }
> +#endif
> +
> + return ret;
> }
>
> #define TRANSTBL_CONVERT(a) { -1, TARGET_##a, -1, a }
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s
2018-07-14 1:07 ` David Gibson
@ 2018-07-14 5:37 ` Laurent Vivier
2018-07-14 9:26 ` David Gibson
0 siblings, 1 reply; 5+ messages in thread
From: Laurent Vivier @ 2018-07-14 5:37 UTC (permalink / raw)
To: David Gibson, Shivaprasad G Bhat; +Cc: riku.voipio, qemu-devel
[-- Attachment #1: Type: text/plain, Size: 2961 bytes --]
Le 14/07/2018 à 03:07, David Gibson a écrit :
> On Fri, Jul 13, 2018 at 07:34:46AM -0500, Shivaprasad G Bhat wrote:
>> Qemu includes the glibc headers for the host defines and target headers are
>> part of the qemu source themselves. The glibc has the F_GETLK64, F_SETLK64
>> and F_SETLKW64 defined to 12, 13 and 14 for all archs in
>> sysdeps/unix/sysv/linux/bits/fcntl-linux.h. The linux kernel generic
>> definition for F_*LK is 5, 6 & 7 and F_*LK64* is 12,13, and 14 as seen in
>> include/uapi/asm-generic/fcntl.h. On 64bit machine, by default the kernel
>> assumes all F_*LK to 64bit calls and doesnt support use of F_*LK64* as
>> can be seen in include/linux/fcntl.h in linux source.
>>
>> On x86_64 host, the values for F_*LK64* are set to 5, 6 and 7
>> explicitly in /usr/include/x86_64-linux-gnu/bits/fcntl.h by the glibc.
>> Whereas, a PPC64 host doesn't have such a definition in
>> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h by the glibc. So,
>> the sources on PPC64 host sees the default value of F_*LK64*
>> as 12, 13 & 14(fcntl-linux.h).
>>
>> Since the 64bit kernel doesnt support 12, 13 & 14; the glibc fcntl syscall
>> implementation(__libc_fcntl*(), __fcntl64_nocancel) does the F_*LK64* value
>> convertion back to F_*LK* values on PPC64 as seen in
>> sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h with FCNTL_ADJUST_CMD()
>> macro. Whereas on x86_64 host the values for F_*LK64* are set to 5, 6 and 7
>> and no adjustments are needed.
>>
>> Since qemu doesnt use the glibc fcntl, but makes the safe_syscall* on its
>> own, the PPC64 qemu is calling the syscall with 12, 13, and 14(without
>> adjustment) and they all fail. The fcntl calls to F_GETLK/F_SETLK|W all
>> fail by all pplications run on PPC64 host user emulation.
>>
>> The fix here could be to see why on PPC64 the glibc is still keeping
>> F_*LK64* different from F_*LK and why adjusting them to 5, 6 and 7 before
>> the syscall for PPC only. See if we can make the
>> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h to have the values
>> 5, 6 & 7 just like x86_64 and remove the adjustment code in glibc. That
>> way, qemu sources see the kernel supported values in glibc headers.
>>
>> OR
>>
>> On PPC64 host, qemu sources see both F_*LK & F_*LK64* as same and set to
>> 12, 13 and 14 because __USE_FILE_OFFSET64 is defined in qemu
>> sources(also refer sysdeps/unix/sysv/linux/bits/fcntl-linux.h).
>> Do the value adjustment just like it is done by glibc source by using
>> F_GETLK value of 5. That way, we make the syscalls with the actual
>> supported values in Qemu. The patch is taking this approach.
>>
>> Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
>
> Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
>
> I'm not sure if this should go in through my tree or not.
I will take it through my linux-user tree, I have some fixes to send for
the next -rc.
Thanks,
Laurent
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s
2018-07-14 5:37 ` Laurent Vivier
@ 2018-07-14 9:26 ` David Gibson
0 siblings, 0 replies; 5+ messages in thread
From: David Gibson @ 2018-07-14 9:26 UTC (permalink / raw)
To: Laurent Vivier; +Cc: Shivaprasad G Bhat, riku.voipio, qemu-devel
[-- Attachment #1: Type: text/plain, Size: 3321 bytes --]
On Sat, Jul 14, 2018 at 07:37:43AM +0200, Laurent Vivier wrote:
> Le 14/07/2018 à 03:07, David Gibson a écrit :
> > On Fri, Jul 13, 2018 at 07:34:46AM -0500, Shivaprasad G Bhat wrote:
> >> Qemu includes the glibc headers for the host defines and target headers are
> >> part of the qemu source themselves. The glibc has the F_GETLK64, F_SETLK64
> >> and F_SETLKW64 defined to 12, 13 and 14 for all archs in
> >> sysdeps/unix/sysv/linux/bits/fcntl-linux.h. The linux kernel generic
> >> definition for F_*LK is 5, 6 & 7 and F_*LK64* is 12,13, and 14 as seen in
> >> include/uapi/asm-generic/fcntl.h. On 64bit machine, by default the kernel
> >> assumes all F_*LK to 64bit calls and doesnt support use of F_*LK64* as
> >> can be seen in include/linux/fcntl.h in linux source.
> >>
> >> On x86_64 host, the values for F_*LK64* are set to 5, 6 and 7
> >> explicitly in /usr/include/x86_64-linux-gnu/bits/fcntl.h by the glibc.
> >> Whereas, a PPC64 host doesn't have such a definition in
> >> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h by the glibc. So,
> >> the sources on PPC64 host sees the default value of F_*LK64*
> >> as 12, 13 & 14(fcntl-linux.h).
> >>
> >> Since the 64bit kernel doesnt support 12, 13 & 14; the glibc fcntl syscall
> >> implementation(__libc_fcntl*(), __fcntl64_nocancel) does the F_*LK64* value
> >> convertion back to F_*LK* values on PPC64 as seen in
> >> sysdeps/unix/sysv/linux/powerpc/powerpc64/sysdep.h with FCNTL_ADJUST_CMD()
> >> macro. Whereas on x86_64 host the values for F_*LK64* are set to 5, 6 and 7
> >> and no adjustments are needed.
> >>
> >> Since qemu doesnt use the glibc fcntl, but makes the safe_syscall* on its
> >> own, the PPC64 qemu is calling the syscall with 12, 13, and 14(without
> >> adjustment) and they all fail. The fcntl calls to F_GETLK/F_SETLK|W all
> >> fail by all pplications run on PPC64 host user emulation.
> >>
> >> The fix here could be to see why on PPC64 the glibc is still keeping
> >> F_*LK64* different from F_*LK and why adjusting them to 5, 6 and 7 before
> >> the syscall for PPC only. See if we can make the
> >> /usr/include/powerpc64le-linux-gnu/bits/fcntl.h to have the values
> >> 5, 6 & 7 just like x86_64 and remove the adjustment code in glibc. That
> >> way, qemu sources see the kernel supported values in glibc headers.
> >>
> >> OR
> >>
> >> On PPC64 host, qemu sources see both F_*LK & F_*LK64* as same and set to
> >> 12, 13 and 14 because __USE_FILE_OFFSET64 is defined in qemu
> >> sources(also refer sysdeps/unix/sysv/linux/bits/fcntl-linux.h).
> >> Do the value adjustment just like it is done by glibc source by using
> >> F_GETLK value of 5. That way, we make the syscalls with the actual
> >> supported values in Qemu. The patch is taking this approach.
> >>
> >> Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
> >
> > Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
> >
> > I'm not sure if this should go in through my tree or not.
>
> I will take it through my linux-user tree, I have some fixes to send for
> the next -rc.
Great, thanks.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2018-07-14 9:32 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2018-07-13 12:34 [Qemu-devel] [PATCH v4] linux-user: ppc64: use the correct values for F_*LK64s Shivaprasad G Bhat
2018-07-13 12:37 ` Laurent Vivier
2018-07-14 1:07 ` David Gibson
2018-07-14 5:37 ` Laurent Vivier
2018-07-14 9:26 ` David Gibson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).