[Qemu-devel] [PATCH] libvixl: Fix format strings for several int64_t values

[Qemu-devel] [PATCH] libvixl: Fix format strings for several int64_t values

[Stefan Weil]

"%d" or "%x" won't work on hosts where int values are smaller than 64 bit.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
---
 disas/libvixl/a64/disasm-a64.cc |   20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/disas/libvixl/a64/disasm-a64.cc b/disas/libvixl/a64/disasm-a64.cc
index 5c6b898..5f172da 100644
--- a/disas/libvixl/a64/disasm-a64.cc
+++ b/disas/libvixl/a64/disasm-a64.cc
@@ -1342,7 +1342,7 @@ int Disassembler::SubstituteImmediateField(Instruction* instr,
         ASSERT(format[5] == 'L');
         AppendToOutput("#0x%" PRIx64, instr->ImmMoveWide());
         if (instr->ShiftMoveWide() > 0) {
-          AppendToOutput(", lsl #%d", 16 * instr->ShiftMoveWide());
+          AppendToOutput(", lsl #%" PRId64, 16 * instr->ShiftMoveWide());
         }
       }
       return 8;
@@ -1391,7 +1391,7 @@ int Disassembler::SubstituteImmediateField(Instruction* instr,
     }
     case 'F': {  // IFPSingle, IFPDouble or IFPFBits.
       if (format[3] == 'F') {  // IFPFbits.
-        AppendToOutput("#%d", 64 - instr->FPScale());
+        AppendToOutput("#%" PRId64, 64 - instr->FPScale());
         return 8;
       } else {
         AppendToOutput("#0x%" PRIx64 " (%.4f)", instr->ImmFP(),
@@ -1412,23 +1412,23 @@ int Disassembler::SubstituteImmediateField(Instruction* instr,
       return 5;
     }
     case 'P': {  // IP - Conditional compare.
-      AppendToOutput("#%d", instr->ImmCondCmp());
+      AppendToOutput("#%" PRId64, instr->ImmCondCmp());
       return 2;
     }
     case 'B': {  // Bitfields.
       return SubstituteBitfieldImmediateField(instr, format);
     }
     case 'E': {  // IExtract.
-      AppendToOutput("#%d", instr->ImmS());
+      AppendToOutput("#%" PRId64, instr->ImmS());
       return 8;
     }
     case 'S': {  // IS - Test and branch bit.
-      AppendToOutput("#%d", (instr->ImmTestBranchBit5() << 5) |
-                            instr->ImmTestBranchBit40());
+      AppendToOutput("#%" PRId64, (instr->ImmTestBranchBit5() << 5) |
+                                  instr->ImmTestBranchBit40());
       return 2;
     }
     case 'D': {  // IDebug - HLT and BRK instructions.
-      AppendToOutput("#0x%x", instr->ImmException());
+      AppendToOutput("#0x%" PRIx64, instr->ImmException());
       return 6;
     }
     default: {
@@ -1598,12 +1598,12 @@ int Disassembler::SubstituteExtendField(Instruction* instr,
       (((instr->ExtendMode() == UXTW) && (instr->SixtyFourBits() == 0)) ||
        (instr->ExtendMode() == UXTX))) {
     if (instr->ImmExtendShift() > 0) {
-      AppendToOutput(", lsl #%d", instr->ImmExtendShift());
+      AppendToOutput(", lsl #%" PRId64, instr->ImmExtendShift());
     }
   } else {
     AppendToOutput(", %s", extend_mode[instr->ExtendMode()]);
     if (instr->ImmExtendShift() > 0) {
-      AppendToOutput(" #%d", instr->ImmExtendShift());
+      AppendToOutput(" #%" PRId64, instr->ImmExtendShift());
     }
   }
   return 3;
@@ -1632,7 +1632,7 @@ int Disassembler::SubstituteLSRegOffsetField(Instruction* instr,
   if (!((ext == UXTX) && (shift == 0))) {
     AppendToOutput(", %s", extend_mode[ext]);
     if (shift != 0) {
-      AppendToOutput(" #%d", instr->SizeLS());
+      AppendToOutput(" #%" PRId64, instr->SizeLS());
     }
   }
   return 9;
-- 
1.7.9.5

Re: [Qemu-devel] [PATCH] libvixl: Fix format strings for several int64_t values

[Andreas Färber]

Am 07.03.2014 20:15, schrieb Stefan Weil:
> "%d" or "%x" won't work on hosts where int values are smaller than 64 bit.
> 
> Signed-off-by: Stefan Weil <sw@weilnetz.de>
> ---
>  disas/libvixl/a64/disasm-a64.cc |   20 ++++++++++----------
>  1 file changed, 10 insertions(+), 10 deletions(-)

Patch looks correct, but I see no indication that this is a backport
from upstream libvixl. Have you submitted it there too?

Regards,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

Re: [Qemu-devel] [PATCH] libvixl: Fix format strings for several int64_t values

[Stefan Weil]

Am 07.03.2014 20:42, schrieb Andreas Färber:
> Am 07.03.2014 20:15, schrieb Stefan Weil:
>> "%d" or "%x" won't work on hosts where int values are smaller than 64 bit.
>>
>> Signed-off-by: Stefan Weil <sw@weilnetz.de>
>> ---
>>  disas/libvixl/a64/disasm-a64.cc |   20 ++++++++++----------
>>  1 file changed, 10 insertions(+), 10 deletions(-)
> 
> Patch looks correct, but I see no indication that this is a backport
> from upstream libvixl. Have you submitted it there too?
> 
> Regards,
> Andreas


If I knew how to submit it to libvixl, I'd have done that already. I did
not find any hint (e-mail address, keyword contrib) in the sources from
github. Can you help me? It would also be fine if someone just takes my
patch and applies it there.

Thanks,
Stefan

Re: [Qemu-devel] [PATCH] libvixl: Fix format strings for several int64_t values

[Andreas Färber]

Am 07.03.2014 21:01, schrieb Stefan Weil:
> Am 07.03.2014 20:42, schrieb Andreas Färber:
>> Am 07.03.2014 20:15, schrieb Stefan Weil:
>>> "%d" or "%x" won't work on hosts where int values are smaller than 64 bit.
>>>
>>> Signed-off-by: Stefan Weil <sw@weilnetz.de>
>>> ---
>>>  disas/libvixl/a64/disasm-a64.cc |   20 ++++++++++----------
>>>  1 file changed, 10 insertions(+), 10 deletions(-)
>>
>> Patch looks correct, but I see no indication that this is a backport
>> from upstream libvixl. Have you submitted it there too?
> 
> If I knew how to submit it to libvixl, I'd have done that already. I did
> not find any hint (e-mail address, keyword contrib) in the sources from
> github. Can you help me?

Google led me to the following:
https://github.com/armvixl/vixl

vixl@arm.com

> It would also be fine if someone just takes my
> patch and applies it there.

Hopefully Peter can help; I just remembered someone asking not to apply
spelling fixes to not complicate functional backports.

Cheers,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

Re: [Qemu-devel] [PATCH] libvixl: Fix format strings for several int64_t values

[Peter Maydell]

On 7 March 2014 22:01, Andreas Färber <afaerber@suse.de> wrote:
> Am 07.03.2014 21:01, schrieb Stefan Weil:
>> Am 07.03.2014 20:42, schrieb Andreas Färber:
>>> Am 07.03.2014 20:15, schrieb Stefan Weil:
>>>> "%d" or "%x" won't work on hosts where int values are smaller than 64 bit.
>>>>
>>>> Signed-off-by: Stefan Weil <sw@weilnetz.de>
>>>> ---
>>>>  disas/libvixl/a64/disasm-a64.cc |   20 ++++++++++----------
>>>>  1 file changed, 10 insertions(+), 10 deletions(-)
>>>
>>> Patch looks correct, but I see no indication that this is a backport
>>> from upstream libvixl. Have you submitted it there too?
>>
>> If I knew how to submit it to libvixl, I'd have done that already. I did
>> not find any hint (e-mail address, keyword contrib) in the sources from
>> github. Can you help me?
>
> Google led me to the following:
> https://github.com/armvixl/vixl
>
> vixl@arm.com

This is the right email address and git repo, yes.
(our disas/libvixl/README gives the github URL, at least.)

I would encourage you to email them patches, it will
act as a prod for them to get themselves set up to
handle incoming contributions properly :-)

>> It would also be fine if someone just takes my
>> patch and applies it there.
>
> Hopefully Peter can help; I just remembered someone asking not to apply
> spelling fixes to not complicate functional backports.

I will apply this patch to target-arm.next. I'm happy
that we carry genuine bugfixes to libvixl (hopefully
in anticipation of them going away on a future update),
it's only non-functional changes I'd rather we don't
carry in our tree.

thanks
-- PMM