Skip to contentSkip to conversation

qdd fadvise64_64 syscall

PATCH In review Riku Voipio 2 participants 4 messages Sep 17, 2008 – Sep 19, 2008

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 qdd fadvise64_64 syscall

Since it's only a hint, we can just add a happy passwtrough. Also from scratchbox.

Signature · 1 line
"rm -rf" only sounds scary if you don't have backups
linux-user/arm/syscall_nr.h +1 -0
Index: trunk/linux-user/arm/syscall_nr.h
===================================================================
@@ -273,6 +273,7 @@
 #define TARGET_NR_fstatfs64			267
 #define TARGET_NR_tgkill			268
 #define TARGET_NR_utimes			269
+#define TARGET_NR_fadvise64_64			270
 #define TARGET_NR_arm_fadvise64_64		270
 #define TARGET_NR_pciconfig_iobase		271
 #define TARGET_NR_pciconfig_read		272
linux-user/syscall.c +6 -0
Index: trunk/linux-user/syscall.c
===================================================================
@@ -5532,6 +5532,12 @@
 		}
         break;
 #endif
+#ifdef TARGET_NR_fadvise64_64
+     case TARGET_NR_fadvise64_64:
+        /* This is a hint, so ignoring and returning success is ok.  */
+        ret = get_errno(0);
+        break;
+#endif
 #ifdef TARGET_NR_madvise
     case TARGET_NR_madvise:
         /* A straight passthrough may not be safe because qemu sometimes
#1 lore raw
Quoted 12 lines · Riku Voipio

Since it's only a hint, we can just add a happy passwtrough. Also from scratchbox.

Signature · 4 lines
"rm -rf" only sounds scary if you don't have backups

Index: trunk/linux-user/arm/syscall_nr.h
===================================================================

linux-user/arm/syscall_nr.h line 279

Quoted 7 lines
linux-user/arm/syscall_nr.h +1 -0
@@ -273,6 +273,7 @@
 #define TARGET_NR_fstatfs64			267
 #define TARGET_NR_tgkill			268
 #define TARGET_NR_utimes			269
+#define TARGET_NR_fadvise64_64			270
 #define TARGET_NR_arm_fadvise64_64		270
 #define TARGET_NR_pciconfig_iobase		271
 #define TARGET_NR_pciconfig_read		272

It's wrong.
ARM doesn't provide fadvise64_64 syscall. It provides arm_fadvise64_64 with different argument order.

Quoted 7 lines

Index: trunk/linux-user/syscall.c
===================================================================

linux-user/syscall.c +0 -0
@@ -5532,6 +5532,12 @@
 		}
         break;

linux-user/syscall.c line 5543

Quoted 10 lines
linux-user/syscall.c +6 -0
 #endif
+#ifdef TARGET_NR_fadvise64_64
+     case TARGET_NR_fadvise64_64:
+        /* This is a hint, so ignoring and returning success is ok.  */
+        ret = get_errno(0);
+        break;
+#endif
 #ifdef TARGET_NR_madvise
     case TARGET_NR_madvise:
         /* A straight passthrough may not be safe because qemu sometimes

Who is this realization make happy?

Signature · 3 lines
Regards,  Kirill A. Shutemov
 + Belarus, Minsk
 + ALT Linux Team, http://www.altlinux.com/
#2 lore raw
Quoted 8 lines · Kirill A. Shutemov
Quoted 6 lines · Riku Voipio

Since it's only a hint, we can just add a happy passwtrough. Also from scratchbox.

Signature · 1 line
"rm -rf" only sounds scary if you don't have backups
Quoted 16 lines

linux-user/arm/syscall_nr.h line 279

Quoted 12 lines

Index: trunk/linux-user/arm/syscall_nr.h
===================================================================

linux-user/arm/syscall_nr.h +1 -0
@@ -273,6 +273,7 @@
 #define TARGET_NR_fstatfs64			267
 #define TARGET_NR_tgkill			268
 #define TARGET_NR_utimes			269
+#define TARGET_NR_fadvise64_64			270
 #define TARGET_NR_arm_fadvise64_64		270
 #define TARGET_NR_pciconfig_iobase		271
 #define TARGET_NR_pciconfig_read		272

It's wrong.
ARM doesn't provide fadvise64_64 syscall. It provides arm_fadvise64_64 with different argument order.

In this case, it doesn't matter since we don't use the arguments. Nevertheless it's decepting and I'll remove it.

Quoted 11 lines
Quoted 11 lines

Index: trunk/linux-user/syscall.c
===================================================================

linux-user/syscall.c +3 -0
@@ -5532,6 +5532,12 @@
 		}
         break;
 #endif
+#ifdef TARGET_NR_fadvise64_64
+     case TARGET_NR_fadvise64_64:
+        /* This is a hint, so ignoring and returning success is ok.  */
Quoted 8 lines

linux-user/syscall.c line 5543

Quoted 6 lines
linux-user/syscall.c +3 -0
+        ret = get_errno(0);
+        break;
+#endif
 #ifdef TARGET_NR_madvise
     case TARGET_NR_madvise:
         /* A straight passthrough may not be safe because qemu sometimes

Who is this realization make happy?

EPARSE.

Signature · 1 line
"rm -rf" only sounds scary if you don't have backups
#3 lore raw
Quoted 26 lines · Riku Voipio
Quoted 8 lines · Kirill A. Shutemov
Quoted 6 lines · Riku Voipio

Since it's only a hint, we can just add a happy passwtrough. Also from scratchbox.

Signature · 1 line
"rm -rf" only sounds scary if you don't have backups
Quoted 16 lines

linux-user/arm/syscall_nr.h line 279

Quoted 12 lines

Index: trunk/linux-user/arm/syscall_nr.h
===================================================================

linux-user/arm/syscall_nr.h +1 -0
@@ -273,6 +273,7 @@
 #define TARGET_NR_fstatfs64			267
 #define TARGET_NR_tgkill			268
 #define TARGET_NR_utimes			269
+#define TARGET_NR_fadvise64_64			270
 #define TARGET_NR_arm_fadvise64_64		270
 #define TARGET_NR_pciconfig_iobase		271
 #define TARGET_NR_pciconfig_read		272

It's wrong.
ARM doesn't provide fadvise64_64 syscall. It provides arm_fadvise64_64 with different argument order.

Quoted 1 line

In this case, it doesn't matter since we don't use the arguments.

It matters if someone will provide true implementation of the syscall.

Quoted 29 lines

Nevertheless it's decepting and I'll remove it.

Quoted 11 lines
Quoted 11 lines

Index: trunk/linux-user/syscall.c
===================================================================

linux-user/syscall.c +3 -0
@@ -5532,6 +5532,12 @@
 		}
         break;
 #endif
+#ifdef TARGET_NR_fadvise64_64
+     case TARGET_NR_fadvise64_64:
+        /* This is a hint, so ignoring and returning success is ok.  */
Quoted 8 lines

linux-user/syscall.c line 5543

Quoted 6 lines
linux-user/syscall.c +3 -0
+        ret = get_errno(0);
+        break;
+#endif
 #ifdef TARGET_NR_madvise
     case TARGET_NR_madvise:
         /* A straight passthrough may not be safe because qemu sometimes

Who is this realization make happy?

EPARSE.

Signature · 1 line
"rm -rf" only sounds scary if you don't have backups
Signature · 3 lines
Regards,  Kirill A. Shutemov
 + Belarus, Minsk
 + ALT Linux Team, http://www.altlinux.com/