Re: [PATCH] tools/libxc: Annotate xc_report_error with __attribute__((format))

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 05/06/14 14:11, Ian Campbell wrote:
> On Tue, 2014-06-03 at 18:45 +0100, Andrew Cooper wrote:
>> diff --git a/tools/libxc/xc_domain_restore.c b/tools/libxc/xc_domain_restore.c
>> index 89af1ad..d950639 100644
>> --- a/tools/libxc/xc_domain_restore.c
>> +++ b/tools/libxc/xc_domain_restore.c
>> @@ -102,7 +102,7 @@ static ssize_t rdexact(xc_interface *xch, struct restore_ctx *ctx,
>>              errno = 0;
>>          }
>>          if ( len <= 0 ) {
>> -            ERROR("%s failed (read rc: %d, errno: %d)", __func__, len, errno);
>> +            ERROR("%s failed (read rc: %ld, errno: %d)", __func__, (long)len, errno);
> The correct way to print an ssize_t is %zd.

Odd - I thought I tried that and it failed, but clearly I failed to try
it correctly.

>
>>              return -1;
>>          }
>>          offset += len;
>> @@ -443,7 +443,7 @@ static int compat_buffer_qemu(xc_interface *xch, struct restore_ctx *ctx,
>>      }
>>  
>>      if ( memcmp(qbuf, "QEVM", 4) ) {
>> -        ERROR("Invalid QEMU magic: 0x%08x", *(unsigned long*)qbuf);
>> +        ERROR("Invalid QEMU magic: 0x%08x", *(uint32_t*)qbuf);
> 0x%08lx with unsigned long or 0x%PRIx32 with uint32_t, not some hybrid,
> please. (I thin %..lx is the right answer here)

qbuf is strictly 4 bytes, as identified in the line above.  PRIx32 it
is.  (I remember deciding on PRIx32 and clearly failed to implement it
properly)

v2 on its way.

~Andrew