* [PATCH RFC] make error codes a formal part of the ABI
@ 2015-01-13 16:21 Jan Beulich
2015-01-13 16:35 ` Ian Campbell
2015-01-13 16:40 ` Andrew Cooper
0 siblings, 2 replies; 14+ messages in thread
From: Jan Beulich @ 2015-01-13 16:21 UTC (permalink / raw)
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Ian Jackson,
Tim Deegan, Ian Campbell, xen-devel
[-- Attachment #1: Type: text/plain, Size: 14047 bytes --]
Now that we have two cases where patches against hvmloader got
submitted needing to include the hypervisor's errno.h (for the host's
system header not necessarily reflecting the correct numbers), take
this as a strong sign that we need to make the error return values part
of the hypervisor ABI (which de-fact they've always been).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
There's on small block commented with TBD left in the public header.
This is the main reason for the submission being RFC. While we don't
currently use these error codes, I'm not sure if we should leave all
or some of them out for the time being.
--- a/xen/include/asm-arm/page.h
+++ b/xen/include/asm-arm/page.h
@@ -2,7 +2,6 @@
#define __ARM_PAGE_H__
#include <xen/config.h>
-#include <xen/errno.h>
#include <public/xen.h>
#include <asm/processor.h>
@@ -83,6 +82,7 @@
#ifndef __ASSEMBLY__
+#include <xen/errno.h>
#include <xen/types.h>
#include <xen/lib.h>
--- a/xen/include/asm-x86/multicall.h
+++ b/xen/include/asm-x86/multicall.h
@@ -24,7 +24,7 @@
" callq *%%rax; " \
"1: movq %%rax,%c4(%0)\n" \
".section .fixup,\"ax\"\n" \
- "2: movq $-"STR(ENOSYS)",%%rax\n" \
+ "2: movq %5,%%rax\n" \
" jmp 1b\n" \
".previous\n" \
: \
@@ -32,7 +32,8 @@
"i" (offsetof(__typeof__(*_call), op)), \
"i" (offsetof(__typeof__(*_call), args)), \
"i" (sizeof(*(_call)->args)), \
- "i" (offsetof(__typeof__(*_call), result)) \
+ "i" (offsetof(__typeof__(*_call), result)), \
+ "i" (-ENOSYS) \
/* all the caller-saves registers */ \
: "rax", "rcx", "rdx", "rsi", "rdi", \
"r8", "r9", "r10", "r11" ); \
@@ -54,7 +55,7 @@
" callq *%%rax; " \
"1: movl %%eax,%c4(%0)\n" \
".section .fixup,\"ax\"\n" \
- "2: movl $-"STR(ENOSYS)",%%eax\n" \
+ "2: movl %5,%%eax\n" \
" jmp 1b\n" \
".previous\n" \
: \
@@ -62,7 +63,8 @@
"i" (offsetof(__typeof__(*_call), op)), \
"i" (offsetof(__typeof__(*_call), args)), \
"i" (sizeof(*(_call)->args)), \
- "i" (offsetof(__typeof__(*_call), result)) \
+ "i" (offsetof(__typeof__(*_call), result)), \
+ "i" (-ENOSYS) \
/* all the caller-saves registers */ \
: "rax", "rcx", "rdx", "rsi", "rdi", \
"r8", "r9", "r10", "r11" ) \
--- /dev/null
+++ b/xen/include/public/errno.h
@@ -0,0 +1,94 @@
+#ifndef __XEN_PUBLIC_ERRNO_H__
+
+#ifndef __ASSEMBLY__
+
+#define XEN_ERRNO(name, value) XEN_##name = value,
+enum xen_errno {
+
+#else /* !__ASSEMBLY__ */
+
+#define XEN_ERRNO(name, value) .equ XEN_##name, value
+
+#endif /* __ASSEMBLY__ */
+
+/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
+/* ` enum errnoval { */
+
+#endif /* __XEN_PUBLIC_ERRNO_H__ */
+
+#ifdef XEN_ERRNO
+
+XEN_ERRNO(EPERM, 1) /* Operation not permitted */
+XEN_ERRNO(ENOENT, 2) /* No such file or directory */
+XEN_ERRNO(ESRCH, 3) /* No such process */
+#ifdef __XEN__
+XEN_ERRNO(EINTR, 4) /* Interrupted system call */
+#endif
+XEN_ERRNO(EIO, 5) /* I/O error */
+XEN_ERRNO(ENXIO, 6) /* No such device or address */
+XEN_ERRNO(E2BIG, 7) /* Arg list too long */
+XEN_ERRNO(ENOEXEC, 8) /* Exec format error */
+XEN_ERRNO(EBADF, 9) /* Bad file number */
+XEN_ERRNO(ECHILD, 10) /* No child processes */
+XEN_ERRNO(EAGAIN, 11) /* Try again */
+XEN_ERRNO(ENOMEM, 12) /* Out of memory */
+XEN_ERRNO(EACCES, 13) /* Permission denied */
+XEN_ERRNO(EFAULT, 14) /* Bad address */
+XEN_ERRNO(EBUSY, 16) /* Device or resource busy */
+XEN_ERRNO(EEXIST, 17) /* File exists */
+XEN_ERRNO(EXDEV, 18) /* Cross-device link */
+XEN_ERRNO(ENODEV, 19) /* No such device */
+XEN_ERRNO(EINVAL, 22) /* Invalid argument */
+XEN_ERRNO(ENFILE, 23) /* File table overflow */
+XEN_ERRNO(EMFILE, 24) /* Too many open files */
+XEN_ERRNO(ENOSPC, 28) /* No space left on device */
+XEN_ERRNO(EMLINK, 31) /* Too many links */
+XEN_ERRNO(EDOM, 33) /* Math argument out of domain of func */
+XEN_ERRNO(ERANGE, 34) /* Math result not representable */
+XEN_ERRNO(EDEADLK, 35) /* Resource deadlock would occur */
+XEN_ERRNO(ENAMETOOLONG, 36) /* File name too long */
+XEN_ERRNO(ENOLCK, 37) /* No record locks available */
+XEN_ERRNO(ENOSYS, 38) /* Function not implemented */
+#if 0/* TBD */
+#define ENOTEMPTY 39 /* Directory not empty */
+#define ELOOP 40 /* Too many symbolic links encountered */
+#define ENOMSG 42 /* No message of desired type */
+#define EIDRM 43 /* Identifier removed */
+#endif/* TBD */
+XEN_ERRNO(EBADRQC, 56) /* Invalid request code */
+XEN_ERRNO(EBADSLT, 57) /* Invalid slot */
+XEN_ERRNO(ENODATA, 61) /* No data available */
+XEN_ERRNO(ETIME, 62) /* Timer expired */
+XEN_ERRNO(EBADMSG, 74) /* Not a data message */
+XEN_ERRNO(EOVERFLOW, 75) /* Value too large for defined data type */
+XEN_ERRNO(EILSEQ, 84) /* Illegal byte sequence */
+#ifdef __XEN__
+XEN_ERRNO(ERESTART, 85) /* Interrupted system call should be restarted */
+#endif
+XEN_ERRNO(EUSERS, 87) /* Too many users */
+XEN_ERRNO(EOPNOTSUPP, 95) /* Operation not supported on transport endpoint */
+XEN_ERRNO(EADDRINUSE, 98) /* Address already in use */
+XEN_ERRNO(EADDRNOTAVAIL, 99) /* Cannot assign requested address */
+XEN_ERRNO(ENOBUFS, 105) /* No buffer space available */
+XEN_ERRNO(EISCONN, 106) /* Transport endpoint is already connected */
+XEN_ERRNO(ENOTCONN, 107) /* Transport endpoint is not connected */
+XEN_ERRNO(ESHUTDOWN, 108) /* Cannot send after transport endpoint shutdown */
+XEN_ERRNO(ETOOMANYREFS, 109) /* Too many references: cannot splice */
+XEN_ERRNO(ETIMEDOUT, 110) /* Connection timed out */
+
+#undef XEN_ERRNO
+#endif /* XEN_ERRNO */
+
+#ifndef __XEN_PUBLIC_ERRNO_H__
+#define __XEN_PUBLIC_ERRNO_H__
+
+/* ` } */
+
+#ifndef __ASSEMBLY__
+};
+#endif
+
+#define XEN_EWOULDBLOCK XEN_EAGAIN /* Operation would block */
+#define XEN_EDEADLOCK XEN_EDEADLK /* Resource deadlock would occur */
+
+#endif /* __XEN_PUBLIC_ERRNO_H__ */
--- a/xen/include/xen/errno.h
+++ b/xen/include/xen/errno.h
@@ -1,137 +1,20 @@
-#ifndef _I386_ERRNO_H
-#define _I386_ERRNO_H
+#ifndef __XEN_ERRNO_H__
+#define __XEN_ERRNO_H__
-/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
-/* ` enum errnoval { */
+#include <public/errno.h>
-#define EPERM 1 /* Operation not permitted */
-#define ENOENT 2 /* No such file or directory */
-#define ESRCH 3 /* No such process */
-#define EINTR 4 /* Interrupted system call */
-#define EIO 5 /* I/O error */
-#define ENXIO 6 /* No such device or address */
-#define E2BIG 7 /* Arg list too long */
-#define ENOEXEC 8 /* Exec format error */
-#define EBADF 9 /* Bad file number */
-#define ECHILD 10 /* No child processes */
-#define EAGAIN 11 /* Try again */
-#define ENOMEM 12 /* Out of memory */
-#define EACCES 13 /* Permission denied */
-#define EFAULT 14 /* Bad address */
-#define ENOTBLK 15 /* Block device required */
-#define EBUSY 16 /* Device or resource busy */
-#define EEXIST 17 /* File exists */
-#define EXDEV 18 /* Cross-device link */
-#define ENODEV 19 /* No such device */
-#define ENOTDIR 20 /* Not a directory */
-#define EISDIR 21 /* Is a directory */
-#define EINVAL 22 /* Invalid argument */
-#define ENFILE 23 /* File table overflow */
-#define EMFILE 24 /* Too many open files */
-#define ENOTTY 25 /* Not a typewriter */
-#define ETXTBSY 26 /* Text file busy */
-#define EFBIG 27 /* File too large */
-#define ENOSPC 28 /* No space left on device */
-#define ESPIPE 29 /* Illegal seek */
-#define EROFS 30 /* Read-only file system */
-#define EMLINK 31 /* Too many links */
-#define EPIPE 32 /* Broken pipe */
-#define EDOM 33 /* Math argument out of domain of func */
-#define ERANGE 34 /* Math result not representable */
-#define EDEADLK 35 /* Resource deadlock would occur */
-#define ENAMETOOLONG 36 /* File name too long */
-#define ENOLCK 37 /* No record locks available */
-#define ENOSYS 38 /* Function not implemented */
-#define ENOTEMPTY 39 /* Directory not empty */
-#define ELOOP 40 /* Too many symbolic links encountered */
-#define EWOULDBLOCK EAGAIN /* Operation would block */
-#define ENOMSG 42 /* No message of desired type */
-#define EIDRM 43 /* Identifier removed */
-#define ECHRNG 44 /* Channel number out of range */
-#define EL2NSYNC 45 /* Level 2 not synchronized */
-#define EL3HLT 46 /* Level 3 halted */
-#define EL3RST 47 /* Level 3 reset */
-#define ELNRNG 48 /* Link number out of range */
-#define EUNATCH 49 /* Protocol driver not attached */
-#define ENOCSI 50 /* No CSI structure available */
-#define EL2HLT 51 /* Level 2 halted */
-#define EBADE 52 /* Invalid exchange */
-#define EBADR 53 /* Invalid request descriptor */
-#define EXFULL 54 /* Exchange full */
-#define ENOANO 55 /* No anode */
-#define EBADRQC 56 /* Invalid request code */
-#define EBADSLT 57 /* Invalid slot */
-
-#define EDEADLOCK EDEADLK
-
-#define EBFONT 59 /* Bad font file format */
-#define ENOSTR 60 /* Device not a stream */
-#define ENODATA 61 /* No data available */
-#define ETIME 62 /* Timer expired */
-#define ENOSR 63 /* Out of streams resources */
-#define ENONET 64 /* Machine is not on the network */
-#define ENOPKG 65 /* Package not installed */
-#define EREMOTE 66 /* Object is remote */
-#define ENOLINK 67 /* Link has been severed */
-#define EADV 68 /* Advertise error */
-#define ESRMNT 69 /* Srmount error */
-#define ECOMM 70 /* Communication error on send */
-#define EPROTO 71 /* Protocol error */
-#define EMULTIHOP 72 /* Multihop attempted */
-#define EDOTDOT 73 /* RFS specific error */
-#define EBADMSG 74 /* Not a data message */
-#define EOVERFLOW 75 /* Value too large for defined data type */
-#define ENOTUNIQ 76 /* Name not unique on network */
-#define EBADFD 77 /* File descriptor in bad state */
-#define EREMCHG 78 /* Remote address changed */
-#define ELIBACC 79 /* Can not access a needed shared library */
-#define ELIBBAD 80 /* Accessing a corrupted shared library */
-#define ELIBSCN 81 /* .lib section in a.out corrupted */
-#define ELIBMAX 82 /* Attempting to link in too many shared libraries */
-#define ELIBEXEC 83 /* Cannot exec a shared library directly */
-#define EILSEQ 84 /* Illegal byte sequence */
-#define ERESTART 85 /* Interrupted system call should be restarted */
-#define ESTRPIPE 86 /* Streams pipe error */
-#define EUSERS 87 /* Too many users */
-#define ENOTSOCK 88 /* Socket operation on non-socket */
-#define EDESTADDRREQ 89 /* Destination address required */
-#define EMSGSIZE 90 /* Message too long */
-#define EPROTOTYPE 91 /* Protocol wrong type for socket */
-#define ENOPROTOOPT 92 /* Protocol not available */
-#define EPROTONOSUPPORT 93 /* Protocol not supported */
-#define ESOCKTNOSUPPORT 94 /* Socket type not supported */
-#define EOPNOTSUPP 95 /* Operation not supported on transport endpoint */
-#define EPFNOSUPPORT 96 /* Protocol family not supported */
-#define EAFNOSUPPORT 97 /* Address family not supported by protocol */
-#define EADDRINUSE 98 /* Address already in use */
-#define EADDRNOTAVAIL 99 /* Cannot assign requested address */
-#define ENETDOWN 100 /* Network is down */
-#define ENETUNREACH 101 /* Network is unreachable */
-#define ENETRESET 102 /* Network dropped connection because of reset */
-#define ECONNABORTED 103 /* Software caused connection abort */
-#define ECONNRESET 104 /* Connection reset by peer */
-#define ENOBUFS 105 /* No buffer space available */
-#define EISCONN 106 /* Transport endpoint is already connected */
-#define ENOTCONN 107 /* Transport endpoint is not connected */
-#define ESHUTDOWN 108 /* Cannot send after transport endpoint shutdown */
-#define ETOOMANYREFS 109 /* Too many references: cannot splice */
-#define ETIMEDOUT 110 /* Connection timed out */
-#define ECONNREFUSED 111 /* Connection refused */
-#define EHOSTDOWN 112 /* Host is down */
-#define EHOSTUNREACH 113 /* No route to host */
-#define EALREADY 114 /* Operation already in progress */
-#define EINPROGRESS 115 /* Operation now in progress */
-#define ESTALE 116 /* Stale NFS file handle */
-#define EUCLEAN 117 /* Structure needs cleaning */
-#define ENOTNAM 118 /* Not a XENIX named type file */
-#define ENAVAIL 119 /* No XENIX semaphores available */
-#define EISNAM 120 /* Is a named type file */
-#define EREMOTEIO 121 /* Remote I/O error */
-#define EDQUOT 122 /* Quota exceeded */
+#ifndef __ASSEMBLY__
-#define ENOMEDIUM 123 /* No medium found */
-#define EMEDIUMTYPE 124 /* Wrong medium type */
+#define XEN_ERRNO(name, value) name = XEN_##name,
+enum {
+#include <public/errno.h>
+};
-/* ` } */
+#else /* !__ASSEMBLY__ */
-#endif
+#define XEN_ERRNO(name, value) .equ name, XEN_##name
+#include <public/errno.h>
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* __XEN_ERRNO_H__ */
[-- Attachment #2: errno-ABI.patch --]
[-- Type: text/plain, Size: 14088 bytes --]
make error codes a formal part of the ABI
Now that we have two cases where patches against hvmloader got
submitted needing to include the hypervisor's errno.h (for the host's
system header not necessarily reflecting the correct numbers), take
this as a strong sign that we need to make the error return values part
of the hypervisor ABI (which de-fact they've always been).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
There's on small block commented with TBD left in the public header.
This is the main reason for the submission being RFC. While we don't
currently use these error codes, I'm not sure if we should leave all
or some of them out for the time being.
--- a/xen/include/asm-arm/page.h
+++ b/xen/include/asm-arm/page.h
@@ -2,7 +2,6 @@
#define __ARM_PAGE_H__
#include <xen/config.h>
-#include <xen/errno.h>
#include <public/xen.h>
#include <asm/processor.h>
@@ -83,6 +82,7 @@
#ifndef __ASSEMBLY__
+#include <xen/errno.h>
#include <xen/types.h>
#include <xen/lib.h>
--- a/xen/include/asm-x86/multicall.h
+++ b/xen/include/asm-x86/multicall.h
@@ -24,7 +24,7 @@
" callq *%%rax; " \
"1: movq %%rax,%c4(%0)\n" \
".section .fixup,\"ax\"\n" \
- "2: movq $-"STR(ENOSYS)",%%rax\n" \
+ "2: movq %5,%%rax\n" \
" jmp 1b\n" \
".previous\n" \
: \
@@ -32,7 +32,8 @@
"i" (offsetof(__typeof__(*_call), op)), \
"i" (offsetof(__typeof__(*_call), args)), \
"i" (sizeof(*(_call)->args)), \
- "i" (offsetof(__typeof__(*_call), result)) \
+ "i" (offsetof(__typeof__(*_call), result)), \
+ "i" (-ENOSYS) \
/* all the caller-saves registers */ \
: "rax", "rcx", "rdx", "rsi", "rdi", \
"r8", "r9", "r10", "r11" ); \
@@ -54,7 +55,7 @@
" callq *%%rax; " \
"1: movl %%eax,%c4(%0)\n" \
".section .fixup,\"ax\"\n" \
- "2: movl $-"STR(ENOSYS)",%%eax\n" \
+ "2: movl %5,%%eax\n" \
" jmp 1b\n" \
".previous\n" \
: \
@@ -62,7 +63,8 @@
"i" (offsetof(__typeof__(*_call), op)), \
"i" (offsetof(__typeof__(*_call), args)), \
"i" (sizeof(*(_call)->args)), \
- "i" (offsetof(__typeof__(*_call), result)) \
+ "i" (offsetof(__typeof__(*_call), result)), \
+ "i" (-ENOSYS) \
/* all the caller-saves registers */ \
: "rax", "rcx", "rdx", "rsi", "rdi", \
"r8", "r9", "r10", "r11" ) \
--- /dev/null
+++ b/xen/include/public/errno.h
@@ -0,0 +1,94 @@
+#ifndef __XEN_PUBLIC_ERRNO_H__
+
+#ifndef __ASSEMBLY__
+
+#define XEN_ERRNO(name, value) XEN_##name = value,
+enum xen_errno {
+
+#else /* !__ASSEMBLY__ */
+
+#define XEN_ERRNO(name, value) .equ XEN_##name, value
+
+#endif /* __ASSEMBLY__ */
+
+/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
+/* ` enum errnoval { */
+
+#endif /* __XEN_PUBLIC_ERRNO_H__ */
+
+#ifdef XEN_ERRNO
+
+XEN_ERRNO(EPERM, 1) /* Operation not permitted */
+XEN_ERRNO(ENOENT, 2) /* No such file or directory */
+XEN_ERRNO(ESRCH, 3) /* No such process */
+#ifdef __XEN__
+XEN_ERRNO(EINTR, 4) /* Interrupted system call */
+#endif
+XEN_ERRNO(EIO, 5) /* I/O error */
+XEN_ERRNO(ENXIO, 6) /* No such device or address */
+XEN_ERRNO(E2BIG, 7) /* Arg list too long */
+XEN_ERRNO(ENOEXEC, 8) /* Exec format error */
+XEN_ERRNO(EBADF, 9) /* Bad file number */
+XEN_ERRNO(ECHILD, 10) /* No child processes */
+XEN_ERRNO(EAGAIN, 11) /* Try again */
+XEN_ERRNO(ENOMEM, 12) /* Out of memory */
+XEN_ERRNO(EACCES, 13) /* Permission denied */
+XEN_ERRNO(EFAULT, 14) /* Bad address */
+XEN_ERRNO(EBUSY, 16) /* Device or resource busy */
+XEN_ERRNO(EEXIST, 17) /* File exists */
+XEN_ERRNO(EXDEV, 18) /* Cross-device link */
+XEN_ERRNO(ENODEV, 19) /* No such device */
+XEN_ERRNO(EINVAL, 22) /* Invalid argument */
+XEN_ERRNO(ENFILE, 23) /* File table overflow */
+XEN_ERRNO(EMFILE, 24) /* Too many open files */
+XEN_ERRNO(ENOSPC, 28) /* No space left on device */
+XEN_ERRNO(EMLINK, 31) /* Too many links */
+XEN_ERRNO(EDOM, 33) /* Math argument out of domain of func */
+XEN_ERRNO(ERANGE, 34) /* Math result not representable */
+XEN_ERRNO(EDEADLK, 35) /* Resource deadlock would occur */
+XEN_ERRNO(ENAMETOOLONG, 36) /* File name too long */
+XEN_ERRNO(ENOLCK, 37) /* No record locks available */
+XEN_ERRNO(ENOSYS, 38) /* Function not implemented */
+#if 0/* TBD */
+#define ENOTEMPTY 39 /* Directory not empty */
+#define ELOOP 40 /* Too many symbolic links encountered */
+#define ENOMSG 42 /* No message of desired type */
+#define EIDRM 43 /* Identifier removed */
+#endif/* TBD */
+XEN_ERRNO(EBADRQC, 56) /* Invalid request code */
+XEN_ERRNO(EBADSLT, 57) /* Invalid slot */
+XEN_ERRNO(ENODATA, 61) /* No data available */
+XEN_ERRNO(ETIME, 62) /* Timer expired */
+XEN_ERRNO(EBADMSG, 74) /* Not a data message */
+XEN_ERRNO(EOVERFLOW, 75) /* Value too large for defined data type */
+XEN_ERRNO(EILSEQ, 84) /* Illegal byte sequence */
+#ifdef __XEN__
+XEN_ERRNO(ERESTART, 85) /* Interrupted system call should be restarted */
+#endif
+XEN_ERRNO(EUSERS, 87) /* Too many users */
+XEN_ERRNO(EOPNOTSUPP, 95) /* Operation not supported on transport endpoint */
+XEN_ERRNO(EADDRINUSE, 98) /* Address already in use */
+XEN_ERRNO(EADDRNOTAVAIL, 99) /* Cannot assign requested address */
+XEN_ERRNO(ENOBUFS, 105) /* No buffer space available */
+XEN_ERRNO(EISCONN, 106) /* Transport endpoint is already connected */
+XEN_ERRNO(ENOTCONN, 107) /* Transport endpoint is not connected */
+XEN_ERRNO(ESHUTDOWN, 108) /* Cannot send after transport endpoint shutdown */
+XEN_ERRNO(ETOOMANYREFS, 109) /* Too many references: cannot splice */
+XEN_ERRNO(ETIMEDOUT, 110) /* Connection timed out */
+
+#undef XEN_ERRNO
+#endif /* XEN_ERRNO */
+
+#ifndef __XEN_PUBLIC_ERRNO_H__
+#define __XEN_PUBLIC_ERRNO_H__
+
+/* ` } */
+
+#ifndef __ASSEMBLY__
+};
+#endif
+
+#define XEN_EWOULDBLOCK XEN_EAGAIN /* Operation would block */
+#define XEN_EDEADLOCK XEN_EDEADLK /* Resource deadlock would occur */
+
+#endif /* __XEN_PUBLIC_ERRNO_H__ */
--- a/xen/include/xen/errno.h
+++ b/xen/include/xen/errno.h
@@ -1,137 +1,20 @@
-#ifndef _I386_ERRNO_H
-#define _I386_ERRNO_H
+#ifndef __XEN_ERRNO_H__
+#define __XEN_ERRNO_H__
-/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
-/* ` enum errnoval { */
+#include <public/errno.h>
-#define EPERM 1 /* Operation not permitted */
-#define ENOENT 2 /* No such file or directory */
-#define ESRCH 3 /* No such process */
-#define EINTR 4 /* Interrupted system call */
-#define EIO 5 /* I/O error */
-#define ENXIO 6 /* No such device or address */
-#define E2BIG 7 /* Arg list too long */
-#define ENOEXEC 8 /* Exec format error */
-#define EBADF 9 /* Bad file number */
-#define ECHILD 10 /* No child processes */
-#define EAGAIN 11 /* Try again */
-#define ENOMEM 12 /* Out of memory */
-#define EACCES 13 /* Permission denied */
-#define EFAULT 14 /* Bad address */
-#define ENOTBLK 15 /* Block device required */
-#define EBUSY 16 /* Device or resource busy */
-#define EEXIST 17 /* File exists */
-#define EXDEV 18 /* Cross-device link */
-#define ENODEV 19 /* No such device */
-#define ENOTDIR 20 /* Not a directory */
-#define EISDIR 21 /* Is a directory */
-#define EINVAL 22 /* Invalid argument */
-#define ENFILE 23 /* File table overflow */
-#define EMFILE 24 /* Too many open files */
-#define ENOTTY 25 /* Not a typewriter */
-#define ETXTBSY 26 /* Text file busy */
-#define EFBIG 27 /* File too large */
-#define ENOSPC 28 /* No space left on device */
-#define ESPIPE 29 /* Illegal seek */
-#define EROFS 30 /* Read-only file system */
-#define EMLINK 31 /* Too many links */
-#define EPIPE 32 /* Broken pipe */
-#define EDOM 33 /* Math argument out of domain of func */
-#define ERANGE 34 /* Math result not representable */
-#define EDEADLK 35 /* Resource deadlock would occur */
-#define ENAMETOOLONG 36 /* File name too long */
-#define ENOLCK 37 /* No record locks available */
-#define ENOSYS 38 /* Function not implemented */
-#define ENOTEMPTY 39 /* Directory not empty */
-#define ELOOP 40 /* Too many symbolic links encountered */
-#define EWOULDBLOCK EAGAIN /* Operation would block */
-#define ENOMSG 42 /* No message of desired type */
-#define EIDRM 43 /* Identifier removed */
-#define ECHRNG 44 /* Channel number out of range */
-#define EL2NSYNC 45 /* Level 2 not synchronized */
-#define EL3HLT 46 /* Level 3 halted */
-#define EL3RST 47 /* Level 3 reset */
-#define ELNRNG 48 /* Link number out of range */
-#define EUNATCH 49 /* Protocol driver not attached */
-#define ENOCSI 50 /* No CSI structure available */
-#define EL2HLT 51 /* Level 2 halted */
-#define EBADE 52 /* Invalid exchange */
-#define EBADR 53 /* Invalid request descriptor */
-#define EXFULL 54 /* Exchange full */
-#define ENOANO 55 /* No anode */
-#define EBADRQC 56 /* Invalid request code */
-#define EBADSLT 57 /* Invalid slot */
-
-#define EDEADLOCK EDEADLK
-
-#define EBFONT 59 /* Bad font file format */
-#define ENOSTR 60 /* Device not a stream */
-#define ENODATA 61 /* No data available */
-#define ETIME 62 /* Timer expired */
-#define ENOSR 63 /* Out of streams resources */
-#define ENONET 64 /* Machine is not on the network */
-#define ENOPKG 65 /* Package not installed */
-#define EREMOTE 66 /* Object is remote */
-#define ENOLINK 67 /* Link has been severed */
-#define EADV 68 /* Advertise error */
-#define ESRMNT 69 /* Srmount error */
-#define ECOMM 70 /* Communication error on send */
-#define EPROTO 71 /* Protocol error */
-#define EMULTIHOP 72 /* Multihop attempted */
-#define EDOTDOT 73 /* RFS specific error */
-#define EBADMSG 74 /* Not a data message */
-#define EOVERFLOW 75 /* Value too large for defined data type */
-#define ENOTUNIQ 76 /* Name not unique on network */
-#define EBADFD 77 /* File descriptor in bad state */
-#define EREMCHG 78 /* Remote address changed */
-#define ELIBACC 79 /* Can not access a needed shared library */
-#define ELIBBAD 80 /* Accessing a corrupted shared library */
-#define ELIBSCN 81 /* .lib section in a.out corrupted */
-#define ELIBMAX 82 /* Attempting to link in too many shared libraries */
-#define ELIBEXEC 83 /* Cannot exec a shared library directly */
-#define EILSEQ 84 /* Illegal byte sequence */
-#define ERESTART 85 /* Interrupted system call should be restarted */
-#define ESTRPIPE 86 /* Streams pipe error */
-#define EUSERS 87 /* Too many users */
-#define ENOTSOCK 88 /* Socket operation on non-socket */
-#define EDESTADDRREQ 89 /* Destination address required */
-#define EMSGSIZE 90 /* Message too long */
-#define EPROTOTYPE 91 /* Protocol wrong type for socket */
-#define ENOPROTOOPT 92 /* Protocol not available */
-#define EPROTONOSUPPORT 93 /* Protocol not supported */
-#define ESOCKTNOSUPPORT 94 /* Socket type not supported */
-#define EOPNOTSUPP 95 /* Operation not supported on transport endpoint */
-#define EPFNOSUPPORT 96 /* Protocol family not supported */
-#define EAFNOSUPPORT 97 /* Address family not supported by protocol */
-#define EADDRINUSE 98 /* Address already in use */
-#define EADDRNOTAVAIL 99 /* Cannot assign requested address */
-#define ENETDOWN 100 /* Network is down */
-#define ENETUNREACH 101 /* Network is unreachable */
-#define ENETRESET 102 /* Network dropped connection because of reset */
-#define ECONNABORTED 103 /* Software caused connection abort */
-#define ECONNRESET 104 /* Connection reset by peer */
-#define ENOBUFS 105 /* No buffer space available */
-#define EISCONN 106 /* Transport endpoint is already connected */
-#define ENOTCONN 107 /* Transport endpoint is not connected */
-#define ESHUTDOWN 108 /* Cannot send after transport endpoint shutdown */
-#define ETOOMANYREFS 109 /* Too many references: cannot splice */
-#define ETIMEDOUT 110 /* Connection timed out */
-#define ECONNREFUSED 111 /* Connection refused */
-#define EHOSTDOWN 112 /* Host is down */
-#define EHOSTUNREACH 113 /* No route to host */
-#define EALREADY 114 /* Operation already in progress */
-#define EINPROGRESS 115 /* Operation now in progress */
-#define ESTALE 116 /* Stale NFS file handle */
-#define EUCLEAN 117 /* Structure needs cleaning */
-#define ENOTNAM 118 /* Not a XENIX named type file */
-#define ENAVAIL 119 /* No XENIX semaphores available */
-#define EISNAM 120 /* Is a named type file */
-#define EREMOTEIO 121 /* Remote I/O error */
-#define EDQUOT 122 /* Quota exceeded */
+#ifndef __ASSEMBLY__
-#define ENOMEDIUM 123 /* No medium found */
-#define EMEDIUMTYPE 124 /* Wrong medium type */
+#define XEN_ERRNO(name, value) name = XEN_##name,
+enum {
+#include <public/errno.h>
+};
-/* ` } */
+#else /* !__ASSEMBLY__ */
-#endif
+#define XEN_ERRNO(name, value) .equ name, XEN_##name
+#include <public/errno.h>
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* __XEN_ERRNO_H__ */
[-- Attachment #3: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:21 [PATCH RFC] make error codes a formal part of the ABI Jan Beulich
@ 2015-01-13 16:35 ` Ian Campbell
2015-01-13 16:57 ` Ian Jackson
2015-01-14 8:46 ` Jan Beulich
2015-01-13 16:40 ` Andrew Cooper
1 sibling, 2 replies; 14+ messages in thread
From: Ian Campbell @ 2015-01-13 16:35 UTC (permalink / raw)
To: Jan Beulich
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
xen-devel, Ian Jackson
On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
> Now that we have two cases where patches against hvmloader got
> submitted needing to include the hypervisor's errno.h (for the host's
> system header not necessarily reflecting the correct numbers), take
> this as a strong sign that we need to make the error return values part
> of the hypervisor ABI (which de-fact they've always been).
Yes, I think this is a good move.
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> ---
> There's on small block commented with TBD left in the public header.
> This is the main reason for the submission being RFC. While we don't
> currently use these error codes, I'm not sure if we should leave all
> or some of them out for the time being.
I say lets omit any we don't use for now.
> --- /dev/null
> +++ b/xen/include/public/errno.h
> @@ -0,0 +1,94 @@
> +#ifndef __XEN_PUBLIC_ERRNO_H__
> +
> +#ifndef __ASSEMBLY__
> +
> +#define XEN_ERRNO(name, value) XEN_##name = value,
> +enum xen_errno {
The switch to an enum doesn't seem related to the main purpose of the
patch, unless I'm missing something?
> +#else /* !__ASSEMBLY__ */
> +
> +#define XEN_ERRNO(name, value) .equ XEN_##name, value
So here public/errno.h defines it's own XEN_ERRNO for ASM vs none. But
then later xen/errno.h also defines it before including the public
version. Also the enum xen_errno seems to be similarly duplicated. (I
suspect you changed your mind and forgot to save one or the other
file?). I think the includer chooses the namespace approach makes most
sense.
(I suppose someone needs to patch libxc et al to actually use this)
> +
> +#endif /* __ASSEMBLY__ */
> +
> +/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
> +/* ` enum errnoval { */
> +
> +#endif /* __XEN_PUBLIC_ERRNO_H__ */
> +
> +#ifdef XEN_ERRNO
> +
> +XEN_ERRNO(EPERM, 1) /* Operation not permitted */
> +XEN_ERRNO(ENOENT, 2) /* No such file or directory */
> +XEN_ERRNO(ESRCH, 3) /* No such process */
> +#ifdef __XEN__
> +XEN_ERRNO(EINTR, 4) /* Interrupted system call */
> +#endif
I take it this is because something prevents this value ever getting
exposes to userspace? (Continuations?). I think keeping that away from
guest API is a good idea, but if it's completely internal perhaps we
should move it up into a region which we reserve for ourselves?
Ian.
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:35 ` Ian Campbell
@ 2015-01-13 16:57 ` Ian Jackson
2015-01-13 17:10 ` Ian Campbell
2015-01-13 17:14 ` Jan Beulich
2015-01-14 8:46 ` Jan Beulich
1 sibling, 2 replies; 14+ messages in thread
From: Ian Jackson @ 2015-01-13 16:57 UTC (permalink / raw)
To: Ian Campbell
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
Jan Beulich, xen-devel
Ian Campbell writes ("Re: [PATCH RFC] make error codes a formal part of the ABI"):
> On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
> > There's on small block commented with TBD left in the public header.
> > This is the main reason for the submission being RFC. While we don't
> > currently use these error codes, I'm not sure if we should leave all
> > or some of them out for the time being.
>
> I say lets omit any we don't use for now.
Is it possible that anyone is using the existing header file where
these values were defined ? If so their code might say
case ELOOP:
which would not compile when they switched to the new header.
I don't know whether this is likely, or a problem.
Ian.
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:57 ` Ian Jackson
@ 2015-01-13 17:10 ` Ian Campbell
2015-01-13 17:14 ` Jan Beulich
1 sibling, 0 replies; 14+ messages in thread
From: Ian Campbell @ 2015-01-13 17:10 UTC (permalink / raw)
To: Ian Jackson
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
Jan Beulich, xen-devel
On Tue, 2015-01-13 at 16:57 +0000, Ian Jackson wrote:
> Ian Campbell writes ("Re: [PATCH RFC] make error codes a formal part of the ABI"):
> > On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
> > > There's on small block commented with TBD left in the public header.
> > > This is the main reason for the submission being RFC. While we don't
> > > currently use these error codes, I'm not sure if we should leave all
> > > or some of them out for the time being.
> >
> > I say lets omit any we don't use for now.
>
> Is it possible that anyone is using the existing header file where
> these values were defined ?
It's not installed or in the regular header paths, so it seems unlikely,
or at least they would have had to jump through some hoops and no doubt
have a big comment about their fragile hack...
> If so their code might say
> case ELOOP:
> which would not compile when they switched to the new header.
>
> I don't know whether this is likely, or a problem.
>
> Ian.
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:57 ` Ian Jackson
2015-01-13 17:10 ` Ian Campbell
@ 2015-01-13 17:14 ` Jan Beulich
1 sibling, 0 replies; 14+ messages in thread
From: Jan Beulich @ 2015-01-13 17:14 UTC (permalink / raw)
To: Ian Jackson
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
Ian Campbell, xen-devel
>>> On 13.01.15 at 17:57, <Ian.Jackson@eu.citrix.com> wrote:
> Ian Campbell writes ("Re: [PATCH RFC] make error codes a formal part of the
> ABI"):
>> On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
>> > There's on small block commented with TBD left in the public header.
>> > This is the main reason for the submission being RFC. While we don't
>> > currently use these error codes, I'm not sure if we should leave all
>> > or some of them out for the time being.
>>
>> I say lets omit any we don't use for now.
>
> Is it possible that anyone is using the existing header file where
> these values were defined ? If so their code might say
> case ELOOP:
> which would not compile when they switched to the new header.
The existing header is a hypervisor private one. Any code outside
the hypervisor using it imo deserves to get broken.
Jan
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:35 ` Ian Campbell
2015-01-13 16:57 ` Ian Jackson
@ 2015-01-14 8:46 ` Jan Beulich
2015-01-14 10:28 ` Ian Campbell
1 sibling, 1 reply; 14+ messages in thread
From: Jan Beulich @ 2015-01-14 8:46 UTC (permalink / raw)
To: Ian Campbell
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
xen-devel, Ian Jackson
>>> On 13.01.15 at 17:35, <Ian.Campbell@eu.citrix.com> wrote:
> On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
>> --- /dev/null
>> +++ b/xen/include/public/errno.h
>> @@ -0,0 +1,94 @@
>> +#ifndef __XEN_PUBLIC_ERRNO_H__
>> +
>> +#ifndef __ASSEMBLY__
>> +
>> +#define XEN_ERRNO(name, value) XEN_##name = value,
>> +enum xen_errno {
>
> The switch to an enum doesn't seem related to the main purpose of the
> patch, unless I'm missing something?
No, this is an integral part of the change: A macro can't be used to
generate preprocessor directives (i.e. #define-s).
>> +#else /* !__ASSEMBLY__ */
>> +
>> +#define XEN_ERRNO(name, value) .equ XEN_##name, value
>
> So here public/errno.h defines it's own XEN_ERRNO for ASM vs none. But
> then later xen/errno.h also defines it before including the public
> version. Also the enum xen_errno seems to be similarly duplicated. (I
> suspect you changed your mind and forgot to save one or the other
> file?). I think the includer chooses the namespace approach makes most
> sense.
No, this again is intentional and - imo - necessary: A plain
#include <public/errno.h> ought to suffice to get all XEN_E*
definitions. That's not so much for Xen's internal purposes, but more
for actual consumers of the public headers. For Xen's internal
purposes, a plain #include <xen/errno.h> ought to suffice and
produce (at least) the non-XEN_-prefixed values. Hence xen/errno.h
has to double-include public/errno.h, once without overriding
XEN_ERRNO() and then a second time with doing so.
> (I suppose someone needs to patch libxc et al to actually use this)
The primary consumer, as said in the description, is meant to be
hvmloader. But yes, other tools parts may also want to follow
that.
>> +
>> +#endif /* __ASSEMBLY__ */
>> +
>> +/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
>> +/* ` enum errnoval { */
>> +
>> +#endif /* __XEN_PUBLIC_ERRNO_H__ */
>> +
>> +#ifdef XEN_ERRNO
>> +
>> +XEN_ERRNO(EPERM, 1) /* Operation not permitted */
>> +XEN_ERRNO(ENOENT, 2) /* No such file or directory */
>> +XEN_ERRNO(ESRCH, 3) /* No such process */
>> +#ifdef __XEN__
>> +XEN_ERRNO(EINTR, 4) /* Interrupted system call */
>> +#endif
>
> I take it this is because something prevents this value ever getting
> exposes to userspace? (Continuations?).
Yes. The thus framed values are supposed to never reach the caller
of a hypercall.
> I think keeping that away from
> guest API is a good idea, but if it's completely internal perhaps we
> should move it up into a region which we reserve for ourselves?
That would be at the risk of (later) creating conflicting definitions. I
specifically wanted to preserve the original values and ordering.
Jan
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-14 8:46 ` Jan Beulich
@ 2015-01-14 10:28 ` Ian Campbell
2015-01-14 10:52 ` Julien Grall
2015-01-14 11:18 ` Jan Beulich
0 siblings, 2 replies; 14+ messages in thread
From: Ian Campbell @ 2015-01-14 10:28 UTC (permalink / raw)
To: Jan Beulich
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
xen-devel, Ian Jackson
On Wed, 2015-01-14 at 08:46 +0000, Jan Beulich wrote:
> >>> On 13.01.15 at 17:35, <Ian.Campbell@eu.citrix.com> wrote:
> > On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
> >> --- /dev/null
> >> +++ b/xen/include/public/errno.h
> >> @@ -0,0 +1,94 @@
> >> +#ifndef __XEN_PUBLIC_ERRNO_H__
> >> +
> >> +#ifndef __ASSEMBLY__
> >> +
> >> +#define XEN_ERRNO(name, value) XEN_##name = value,
> >> +enum xen_errno {
> >
> > The switch to an enum doesn't seem related to the main purpose of the
> > patch, unless I'm missing something?
>
> No, this is an integral part of the change: A macro can't be used to
> generate preprocessor directives (i.e. #define-s).
Oh right, yes.
Given the ABI pitfalls of enums (size etc) perhaps make it anonymous?
> >> +#else /* !__ASSEMBLY__ */
> >> +
> >> +#define XEN_ERRNO(name, value) .equ XEN_##name, value
> >
> > So here public/errno.h defines it's own XEN_ERRNO for ASM vs none. But
> > then later xen/errno.h also defines it before including the public
> > version. Also the enum xen_errno seems to be similarly duplicated. (I
> > suspect you changed your mind and forgot to save one or the other
> > file?). I think the includer chooses the namespace approach makes most
> > sense.
>
> No, this again is intentional and - imo - necessary: A plain
> #include <public/errno.h> ought to suffice to get all XEN_E*
> definitions. That's not so much for Xen's internal purposes, but more
> for actual consumers of the public headers. For Xen's internal
> purposes, a plain #include <xen/errno.h> ought to suffice and
> produce (at least) the non-XEN_-prefixed values. Hence xen/errno.h
> has to double-include public/errno.h, once without overriding
> XEN_ERRNO() and then a second time with doing so.
Oh I see now that you are relying on the multi-inclusion guard to also
guard the definition of the default version of XEN_ERRNO macro, which is
a bit tricksy (and wasn't obvious until I applied the patch and looked
at the result).
Can you make this more explicit while leaving the guard with its usual
semantics? e.g.
#ifndef XEN_ERRNO
#if .. Asm ...
#define ...
#else
#define ...
#endif
and in xen/errno.h:
#include <public/errno.h>
/* We explicitly want a second include with separate names. */
#undef __XEN_PUBLIC_ERRNO_H__
#if ... asm...
#define ...
#include <...>
#else
#define ...
enum {
#include <...>
}
#endif
If you don't like the #undef then perhaps move the value list into
public/errno-values.h as a bare list which requires XEN_ERRNO to be
defined by the includer and has no guard of its own, so it can be
included from both the public and private errno.h with the correct
context?
> > (I suppose someone needs to patch libxc et al to actually use this)
>
> The primary consumer, as said in the description, is meant to be
> hvmloader. But yes, other tools parts may also want to follow
> that.
/me wonders how libxc has been getting away with out this until now,
especially on non-Linux platforms.
Fixing libxc to correctly disambiguate xen errno's from genuine syscalls
ones is going to suck... Oh well, not your problem at least, lucky
you ;-)
>
> >> +
> >> +#endif /* __ASSEMBLY__ */
> >> +
> >> +/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
> >> +/* ` enum errnoval { */
> >> +
> >> +#endif /* __XEN_PUBLIC_ERRNO_H__ */
> >> +
> >> +#ifdef XEN_ERRNO
> >> +
> >> +XEN_ERRNO(EPERM, 1) /* Operation not permitted */
> >> +XEN_ERRNO(ENOENT, 2) /* No such file or directory */
> >> +XEN_ERRNO(ESRCH, 3) /* No such process */
> >> +#ifdef __XEN__
> >> +XEN_ERRNO(EINTR, 4) /* Interrupted system call */
> >> +#endif
> >
> > I take it this is because something prevents this value ever getting
> > exposes to userspace? (Continuations?).
>
> Yes. The thus framed values are supposed to never reach the caller
> of a hypercall.
>
> > I think keeping that away from
> > guest API is a good idea, but if it's completely internal perhaps we
> > should move it up into a region which we reserve for ourselves?
>
> That would be at the risk of (later) creating conflicting definitions. I
> specifically wanted to preserve the original values and ordering.
Good reason.
Perhaps
#ifdef __XEN__ /* Internal only, should never be exposed to the guest */
since there isn't so many of them to make that onerous.
Ian.
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-14 10:28 ` Ian Campbell
@ 2015-01-14 10:52 ` Julien Grall
2015-01-14 11:24 ` Ian Campbell
2015-01-14 11:18 ` Jan Beulich
1 sibling, 1 reply; 14+ messages in thread
From: Julien Grall @ 2015-01-14 10:52 UTC (permalink / raw)
To: Ian Campbell, Jan Beulich
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Ian Jackson,
Tim Deegan, xen-devel, Roger Pau Monné
Hi Ian,
On 14/01/2015 10:28, Ian Campbell wrote:
>>> (I suppose someone needs to patch libxc et al to actually use this)
>>
>> The primary consumer, as said in the description, is meant to be
>> hvmloader. But yes, other tools parts may also want to follow
>> that.
>
> /me wonders how libxc has been getting away with out this until now,
> especially on non-Linux platforms.
On FreeBSD, we translate the XEN errno to the guest one in the privcmd
drivers.
See
http://xenbits.xen.org/gitweb/?p=people/royger/freebsd.git;a=patch;h=f09eeed01bc884b37e978f6ec6e3e7a86778ef4b
Regards,
--
Julien Grall
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-14 10:52 ` Julien Grall
@ 2015-01-14 11:24 ` Ian Campbell
2015-01-14 11:37 ` Andrew Cooper
0 siblings, 1 reply; 14+ messages in thread
From: Ian Campbell @ 2015-01-14 11:24 UTC (permalink / raw)
To: Julien Grall
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
Jan Beulich, xen-devel, Ian Jackson, Roger Pau Monné
On Wed, 2015-01-14 at 10:52 +0000, Julien Grall wrote:
> Hi Ian,
>
> On 14/01/2015 10:28, Ian Campbell wrote:
> >>> (I suppose someone needs to patch libxc et al to actually use this)
> >>
> >> The primary consumer, as said in the description, is meant to be
> >> hvmloader. But yes, other tools parts may also want to follow
> >> that.
> >
> > /me wonders how libxc has been getting away with out this until now,
> > especially on non-Linux platforms.
>
> On FreeBSD, we translate the XEN errno to the guest one in the privcmd
> drivers.
Ah, I looked in tools/libxc/*bsd* but didn't think to look in the kernel
itself.
> See
> http://xenbits.xen.org/gitweb/?p=people/royger/freebsd.git;a=patch;h=f09eeed01bc884b37e978f6ec6e3e7a86778ef4b
OK, so essentially the requirement is that libxc should see the guest's
own ERRNO numbers and changing libxc to use xen's numbers would just be
wrong.
Which is good, because making that change would be a nightmare!
Ian.
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-14 11:24 ` Ian Campbell
@ 2015-01-14 11:37 ` Andrew Cooper
0 siblings, 0 replies; 14+ messages in thread
From: Andrew Cooper @ 2015-01-14 11:37 UTC (permalink / raw)
To: Ian Campbell, Julien Grall
Cc: Keir Fraser, Stefano Stabellini, Tim Deegan, Ian Jackson,
Jan Beulich, xen-devel, Roger Pau Monné
On 14/01/15 11:24, Ian Campbell wrote:
> On Wed, 2015-01-14 at 10:52 +0000, Julien Grall wrote:
>> Hi Ian,
>>
>> On 14/01/2015 10:28, Ian Campbell wrote:
>>>>> (I suppose someone needs to patch libxc et al to actually use this)
>>>> The primary consumer, as said in the description, is meant to be
>>>> hvmloader. But yes, other tools parts may also want to follow
>>>> that.
>>> /me wonders how libxc has been getting away with out this until now,
>>> especially on non-Linux platforms.
>> On FreeBSD, we translate the XEN errno to the guest one in the privcmd
>> drivers.
> Ah, I looked in tools/libxc/*bsd* but didn't think to look in the kernel
> itself.
>
>> See
>> http://xenbits.xen.org/gitweb/?p=people/royger/freebsd.git;a=patch;h=f09eeed01bc884b37e978f6ec6e3e7a86778ef4b
> OK, so essentially the requirement is that libxc should see the guest's
> own ERRNO numbers and changing libxc to use xen's numbers would just be
> wrong.
>
> Which is good, because making that change would be a nightmare!
It is not so simple.
Currently, the hypercall ioctl overloads errno from both the kernel and
Xen. This makes it impossible for libxc to programmaticly distinguish
an kernel error from a Xen error (e.g. where did an EFAULT come from).
Ideally, the correct solution would be for the ioctl to return -1/errno
for a kernel failure alone, and then have the top level stub inspect
args[0] for the Xen error. This would require libxc to know about both
the local errnos and Xen errnos.
~Andrew
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-14 10:28 ` Ian Campbell
2015-01-14 10:52 ` Julien Grall
@ 2015-01-14 11:18 ` Jan Beulich
2015-01-14 11:27 ` Ian Campbell
1 sibling, 1 reply; 14+ messages in thread
From: Jan Beulich @ 2015-01-14 11:18 UTC (permalink / raw)
To: Ian Campbell
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
xen-devel, Ian Jackson
>>> On 14.01.15 at 11:28, <Ian.Campbell@eu.citrix.com> wrote:
> On Wed, 2015-01-14 at 08:46 +0000, Jan Beulich wrote:
>> >>> On 13.01.15 at 17:35, <Ian.Campbell@eu.citrix.com> wrote:
>> > On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
>> >> --- /dev/null
>> >> +++ b/xen/include/public/errno.h
>> >> @@ -0,0 +1,94 @@
>> >> +#ifndef __XEN_PUBLIC_ERRNO_H__
>> >> +
>> >> +#ifndef __ASSEMBLY__
>> >> +
>> >> +#define XEN_ERRNO(name, value) XEN_##name = value,
>> >> +enum xen_errno {
>> >
>> > The switch to an enum doesn't seem related to the main purpose of the
>> > patch, unless I'm missing something?
>>
>> No, this is an integral part of the change: A macro can't be used to
>> generate preprocessor directives (i.e. #define-s).
>
> Oh right, yes.
>
> Given the ABI pitfalls of enums (size etc) perhaps make it anonymous?
Could do, but no-one is required to use the enum as a type. I just
wanted to _allow_ people using it if they want.
>> >> +#else /* !__ASSEMBLY__ */
>> >> +
>> >> +#define XEN_ERRNO(name, value) .equ XEN_##name, value
>> >
>> > So here public/errno.h defines it's own XEN_ERRNO for ASM vs none. But
>> > then later xen/errno.h also defines it before including the public
>> > version. Also the enum xen_errno seems to be similarly duplicated. (I
>> > suspect you changed your mind and forgot to save one or the other
>> > file?). I think the includer chooses the namespace approach makes most
>> > sense.
>>
>> No, this again is intentional and - imo - necessary: A plain
>> #include <public/errno.h> ought to suffice to get all XEN_E*
>> definitions. That's not so much for Xen's internal purposes, but more
>> for actual consumers of the public headers. For Xen's internal
>> purposes, a plain #include <xen/errno.h> ought to suffice and
>> produce (at least) the non-XEN_-prefixed values. Hence xen/errno.h
>> has to double-include public/errno.h, once without overriding
>> XEN_ERRNO() and then a second time with doing so.
>
> Oh I see now that you are relying on the multi-inclusion guard to also
> guard the definition of the default version of XEN_ERRNO macro, which is
> a bit tricksy (and wasn't obvious until I applied the patch and looked
> at the result).
>
> Can you make this more explicit while leaving the guard with its usual
> semantics? e.g.
>
> #ifndef XEN_ERRNO
> #if .. Asm ...
> #define ...
> #else
> #define ...
> #endif
I don't think that would work when including public/errno.h a second
time. Or maybe I'm not getting how you envision things to be...
> and in xen/errno.h:
>
> #include <public/errno.h>
>
> /* We explicitly want a second include with separate names. */
> #undef __XEN_PUBLIC_ERRNO_H__
> #if ... asm...
> #define ...
> #include <...>
> #else
> #define ...
> enum {
> #include <...>
> }
> #endif
>
> If you don't like the #undef then perhaps move the value list into
> public/errno-values.h as a bare list which requires XEN_ERRNO to be
> defined by the includer and has no guard of its own, so it can be
> included from both the public and private errno.h with the correct
> context?
To be honest I prefer a solution with just a single new public header,
even if it ends up using some trickery (so long as it doesn't violate
any language requirements).
>> > I think keeping that away from
>> > guest API is a good idea, but if it's completely internal perhaps we
>> > should move it up into a region which we reserve for ourselves?
>>
>> That would be at the risk of (later) creating conflicting definitions. I
>> specifically wanted to preserve the original values and ordering.
>
> Good reason.
>
> Perhaps
> #ifdef __XEN__ /* Internal only, should never be exposed to the guest */
> since there isn't so many of them to make that onerous.
Sure, easily added.
Jan
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-14 11:18 ` Jan Beulich
@ 2015-01-14 11:27 ` Ian Campbell
0 siblings, 0 replies; 14+ messages in thread
From: Ian Campbell @ 2015-01-14 11:27 UTC (permalink / raw)
To: Jan Beulich
Cc: Keir Fraser, Stefano Stabellini, Andrew Cooper, Tim Deegan,
xen-devel, Ian Jackson
On Wed, 2015-01-14 at 11:18 +0000, Jan Beulich wrote:
> >>> On 14.01.15 at 11:28, <Ian.Campbell@eu.citrix.com> wrote:
> > On Wed, 2015-01-14 at 08:46 +0000, Jan Beulich wrote:
> >> >>> On 13.01.15 at 17:35, <Ian.Campbell@eu.citrix.com> wrote:
> >> > On Tue, 2015-01-13 at 16:21 +0000, Jan Beulich wrote:
> >> >> --- /dev/null
> >> >> +++ b/xen/include/public/errno.h
> >> >> @@ -0,0 +1,94 @@
> >> >> +#ifndef __XEN_PUBLIC_ERRNO_H__
> >> >> +
> >> >> +#ifndef __ASSEMBLY__
> >> >> +
> >> >> +#define XEN_ERRNO(name, value) XEN_##name = value,
> >> >> +enum xen_errno {
> >> >
> >> > The switch to an enum doesn't seem related to the main purpose of the
> >> > patch, unless I'm missing something?
> >>
> >> No, this is an integral part of the change: A macro can't be used to
> >> generate preprocessor directives (i.e. #define-s).
> >
> > Oh right, yes.
> >
> > Given the ABI pitfalls of enums (size etc) perhaps make it anonymous?
>
> Could do, but no-one is required to use the enum as a type. I just
> wanted to _allow_ people using it if they want.
I was thinking to avoid presenting them with the temptation.
> >> >> +#else /* !__ASSEMBLY__ */
> >> >> +
> >> >> +#define XEN_ERRNO(name, value) .equ XEN_##name, value
> >> >
> >> > So here public/errno.h defines it's own XEN_ERRNO for ASM vs none. But
> >> > then later xen/errno.h also defines it before including the public
> >> > version. Also the enum xen_errno seems to be similarly duplicated. (I
> >> > suspect you changed your mind and forgot to save one or the other
> >> > file?). I think the includer chooses the namespace approach makes most
> >> > sense.
> >>
> >> No, this again is intentional and - imo - necessary: A plain
> >> #include <public/errno.h> ought to suffice to get all XEN_E*
> >> definitions. That's not so much for Xen's internal purposes, but more
> >> for actual consumers of the public headers. For Xen's internal
> >> purposes, a plain #include <xen/errno.h> ought to suffice and
> >> produce (at least) the non-XEN_-prefixed values. Hence xen/errno.h
> >> has to double-include public/errno.h, once without overriding
> >> XEN_ERRNO() and then a second time with doing so.
> >
> > Oh I see now that you are relying on the multi-inclusion guard to also
> > guard the definition of the default version of XEN_ERRNO macro, which is
> > a bit tricksy (and wasn't obvious until I applied the patch and looked
> > at the result).
> >
> > Can you make this more explicit while leaving the guard with its usual
> > semantics? e.g.
> >
> > #ifndef XEN_ERRNO
> > #if .. Asm ...
> > #define ...
> > #else
> > #define ...
enum xen_errno {
> > #endif
> I don't think that would work when including public/errno.h a second
> time. Or maybe I'm not getting how you envision things to be...
I forgot the enum stuff above, I've inserted it in the quote.
The second time around XEN_ERRNO would be defined and so none of that
stuff would happen, just the list of XEN_ERRNO(value..) would be
evaluated (in the context of xen/errno.h's "enum {").
Thinking about it now I guess some more trickery would be needed for the
closing } of the enum xen_errno, which could well end up being a cure
worse than the disease.
> To be honest I prefer a solution with just a single new public header,
> even if it ends up using some trickery (so long as it doesn't violate
> any language requirements).
I lean the other way, but not enough to continue arguing, so: ok.
Ian.
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:21 [PATCH RFC] make error codes a formal part of the ABI Jan Beulich
2015-01-13 16:35 ` Ian Campbell
@ 2015-01-13 16:40 ` Andrew Cooper
2015-01-14 8:54 ` Jan Beulich
1 sibling, 1 reply; 14+ messages in thread
From: Andrew Cooper @ 2015-01-13 16:40 UTC (permalink / raw)
To: Jan Beulich
Cc: Keir Fraser, Stefano Stabellini, Ian Jackson, Tim Deegan,
Ian Campbell, xen-devel
On 13/01/15 16:21, Jan Beulich wrote:
> Now that we have two cases where patches against hvmloader got
> submitted needing to include the hypervisor's errno.h (for the host's
> system header not necessarily reflecting the correct numbers), take
> this as a strong sign that we need to make the error return values part
> of the hypervisor ABI (which de-fact they've always been).
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> ---
> There's on small block commented with TBD left in the public header.
> This is the main reason for the submission being RFC. While we don't
> currently use these error codes, I'm not sure if we should leave all
> or some of them out for the time being.
I would suggest that we drop any error codes not in use. We can always
add them back in in the future if they are needed.
I would also suggest that we note the heritage, being taken from Linux
2.$mumble, which will act as a guide for introducing new codes in the
future.
Finally, I would suggest that we assign the Xen meaning (e.g. EACCES =
bad tool interface) rather than Linux meanings.
>
> --- a/xen/include/asm-arm/page.h
> +++ b/xen/include/asm-arm/page.h
> @@ -2,7 +2,6 @@
> #define __ARM_PAGE_H__
>
> #include <xen/config.h>
> -#include <xen/errno.h>
> #include <public/xen.h>
> #include <asm/processor.h>
>
> @@ -83,6 +82,7 @@
>
> #ifndef __ASSEMBLY__
>
> +#include <xen/errno.h>
> #include <xen/types.h>
> #include <xen/lib.h>
>
> --- a/xen/include/asm-x86/multicall.h
> +++ b/xen/include/asm-x86/multicall.h
> @@ -24,7 +24,7 @@
> " callq *%%rax; " \
> "1: movq %%rax,%c4(%0)\n" \
> ".section .fixup,\"ax\"\n" \
> - "2: movq $-"STR(ENOSYS)",%%rax\n" \
> + "2: movq %5,%%rax\n" \
> " jmp 1b\n" \
> ".previous\n" \
> : \
> @@ -32,7 +32,8 @@
> "i" (offsetof(__typeof__(*_call), op)), \
> "i" (offsetof(__typeof__(*_call), args)), \
> "i" (sizeof(*(_call)->args)), \
> - "i" (offsetof(__typeof__(*_call), result)) \
> + "i" (offsetof(__typeof__(*_call), result)), \
> + "i" (-ENOSYS) \
> /* all the caller-saves registers */ \
> : "rax", "rcx", "rdx", "rsi", "rdi", \
> "r8", "r9", "r10", "r11" ); \
> @@ -54,7 +55,7 @@
> " callq *%%rax; " \
> "1: movl %%eax,%c4(%0)\n" \
> ".section .fixup,\"ax\"\n" \
> - "2: movl $-"STR(ENOSYS)",%%eax\n" \
> + "2: movl %5,%%eax\n" \
> " jmp 1b\n" \
> ".previous\n" \
> : \
> @@ -62,7 +63,8 @@
> "i" (offsetof(__typeof__(*_call), op)), \
> "i" (offsetof(__typeof__(*_call), args)), \
> "i" (sizeof(*(_call)->args)), \
> - "i" (offsetof(__typeof__(*_call), result)) \
> + "i" (offsetof(__typeof__(*_call), result)), \
> + "i" (-ENOSYS) \
> /* all the caller-saves registers */ \
> : "rax", "rcx", "rdx", "rsi", "rdi", \
> "r8", "r9", "r10", "r11" ) \
> --- /dev/null
> +++ b/xen/include/public/errno.h
> @@ -0,0 +1,94 @@
> +#ifndef __XEN_PUBLIC_ERRNO_H__
> +
> +#ifndef __ASSEMBLY__
> +
> +#define XEN_ERRNO(name, value) XEN_##name = value,
> +enum xen_errno {
> +
> +#else /* !__ASSEMBLY__ */
> +
> +#define XEN_ERRNO(name, value) .equ XEN_##name, value
> +
> +#endif /* __ASSEMBLY__ */
> +
> +/* ` enum neg_errnoval { [ -Efoo for each Efoo in the list below ] } */
> +/* ` enum errnoval { */
> +
> +#endif /* __XEN_PUBLIC_ERRNO_H__ */
> +
> +#ifdef XEN_ERRNO
> +
> +XEN_ERRNO(EPERM, 1) /* Operation not permitted */
> +XEN_ERRNO(ENOENT, 2) /* No such file or directory */
> +XEN_ERRNO(ESRCH, 3) /* No such process */
> +#ifdef __XEN__
> +XEN_ERRNO(EINTR, 4) /* Interrupted system call */
> +#endif
Why is EINTR (and ERESTART lower) hidden from non-xen consumers but
still in the public api? Would it not be better to keep these in
xen/errno.h rather than public/errno.h ?
~Andrew
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [PATCH RFC] make error codes a formal part of the ABI
2015-01-13 16:40 ` Andrew Cooper
@ 2015-01-14 8:54 ` Jan Beulich
0 siblings, 0 replies; 14+ messages in thread
From: Jan Beulich @ 2015-01-14 8:54 UTC (permalink / raw)
To: Andrew Cooper
Cc: Keir Fraser, Stefano Stabellini, Ian Jackson, Tim Deegan,
Ian Campbell, xen-devel
>>> On 13.01.15 at 17:40, <andrew.cooper3@citrix.com> wrote:
> On 13/01/15 16:21, Jan Beulich wrote:
>> Now that we have two cases where patches against hvmloader got
>> submitted needing to include the hypervisor's errno.h (for the host's
>> system header not necessarily reflecting the correct numbers), take
>> this as a strong sign that we need to make the error return values part
>> of the hypervisor ABI (which de-fact they've always been).
>>
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>> ---
>> There's on small block commented with TBD left in the public header.
>> This is the main reason for the submission being RFC. While we don't
>> currently use these error codes, I'm not sure if we should leave all
>> or some of them out for the time being.
>
> I would suggest that we drop any error codes not in use. We can always
> add them back in in the future if they are needed.
>
> I would also suggest that we note the heritage, being taken from Linux
> 2.$mumble, which will act as a guide for introducing new codes in the
> future.
Good idea - I added a respective note.
> Finally, I would suggest that we assign the Xen meaning (e.g. EACCES =
> bad tool interface) rather than Linux meanings.
Looking through the uses, EACCES isn't solely used for that purpose,
and hence I'd like to keep the comment as is (assuming that it's the
comments associated with each entry that your remark was about).
>> +#ifdef XEN_ERRNO
>> +
>> +XEN_ERRNO(EPERM, 1) /* Operation not permitted */
>> +XEN_ERRNO(ENOENT, 2) /* No such file or directory */
>> +XEN_ERRNO(ESRCH, 3) /* No such process */
>> +#ifdef __XEN__
>> +XEN_ERRNO(EINTR, 4) /* Interrupted system call */
>> +#endif
>
> Why is EINTR (and ERESTART lower) hidden from non-xen consumers but
> still in the public api? Would it not be better to keep these in
> xen/errno.h rather than public/errno.h ?
As said in the reply to Ian - to avoid the risk of introducing conflicting
definitions later, I'd like to keep them all in a single place.
Jan
^ permalink raw reply [flat|nested] 14+ messages in thread
end of thread, other threads:[~2015-01-14 11:37 UTC | newest]
Thread overview: 14+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-01-13 16:21 [PATCH RFC] make error codes a formal part of the ABI Jan Beulich
2015-01-13 16:35 ` Ian Campbell
2015-01-13 16:57 ` Ian Jackson
2015-01-13 17:10 ` Ian Campbell
2015-01-13 17:14 ` Jan Beulich
2015-01-14 8:46 ` Jan Beulich
2015-01-14 10:28 ` Ian Campbell
2015-01-14 10:52 ` Julien Grall
2015-01-14 11:24 ` Ian Campbell
2015-01-14 11:37 ` Andrew Cooper
2015-01-14 11:18 ` Jan Beulich
2015-01-14 11:27 ` Ian Campbell
2015-01-13 16:40 ` Andrew Cooper
2015-01-14 8:54 ` Jan Beulich
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.