This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix cygwin build error with i386-linux-tdep.c


Hui Zhu wrote:

Hi guys,

I make a new patch that change regcache_raw_read to regcache_raw_read_unsigned.
Please help me review it.

Thanks,
Hui

2009-08-31 Hui Zhu <teawater@gmail.com>

	* i386-linux-tdep.c (i386_linux_intx80_sysenter_record): Change
	regcache_raw_read to regcache_raw_read_unsigned.

---
 i386-linux-tdep.c |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

--- a/i386-linux-tdep.c
+++ b/i386-linux-tdep.c
@@ -367,18 +367,18 @@ static int
 i386_linux_intx80_sysenter_record (struct regcache *regcache)
 {
   int ret;
-  uint32_t tmpu32;
+  ULONGEST num;

I like Mark's suggestion of calling it "syscall". Calling it "num" is not much better than calling it "tmp". ;-)

Now, talking about signed vs. unsigned vs. casts...

You declared it a ULONGEST because that's what
regcache_raw_read_unsigned expects, right?  -- fair enough.
See below.

-  regcache_raw_read (regcache, I386_EAX_REGNUM, (gdb_byte *) &tmpu32);
+  regcache_raw_read_unsigned (regcache, I386_EAX_REGNUM, &num);

-  if (tmpu32 > 499)
+  if (num > 499)
     {
       printf_unfiltered (_("Process record and replay target doesn't "
-                           "support syscall number %u\n"), tmpu32);
+                           "support syscall number %d\n"), (int) num);
       return -1;
     }

-  ret = record_linux_system_call (tmpu32, regcache,
+  ret = record_linux_system_call ((int) num, regcache,
 				  &i386_linux_record_tdep);

But here you cast it to int because that's what record_linux_system_call expects, right?

And since we know at this point that the value is between
0 and 499, the cast can't really do any harm, so it seems
fair.

I could suggest casting it to (unsigned int), but it wouldn't
really make any difference, would it?  Mark -- Jiang -- would
that make you guys more comfortable?

Or would it be more correct to change the prototype of
linux-record-system-call so that it expects an unsigned int?


   if (ret)
     return ret;



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]