[PATCH] ioemu: use asprintf instead of PATH_MAX

[PATCH] ioemu: use asprintf instead of PATH_MAX

[Samuel Thibault]

Use asprintf instead of PATH_MAX, which POSIX says to be facultative.

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>

diff -r c364f80eb4b5 tools/ioemu/target-i386-dm/helper2.c
--- a/tools/ioemu/target-i386-dm/helper2.c	Wed Jan 23 13:27:21 2008 +0000
+++ b/tools/ioemu/target-i386-dm/helper2.c	Wed Jan 23 17:18:56 2008 +0000
@@ -635,7 +635,7 @@
     extern int suspend_requested;
     CPUState *env = cpu_single_env;
     int evtchn_fd = xce_handle == -1 ? -1 : xc_evtchn_fd(xce_handle);
-    char qemu_file[PATH_MAX];
+    char *qemu_file;
     fd_set fds;
     int ret = 0;
 
@@ -665,9 +665,9 @@
         main_loop_wait(1); /* For the select() on events */
 
         /* Save the device state */
-        snprintf(qemu_file, sizeof(qemu_file), 
-                 "/var/lib/xen/qemu-save.%d", domid);
+        asprintf(&qemu_file, "/var/lib/xen/qemu-save.%d", domid);
         do_savevm(qemu_file);
+        free(qemu_file);
 
         xenstore_record_dm_state("paused");

Re: ioemu: use asprintf instead of PATH_MAX

[Mike D. Day]

erOn 23/01/08 17:20 +0000, Samuel Thibault wrote:
> Use asprintf instead of PATH_MAX, which POSIX says to be facultative.
> 
> Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>

> -        snprintf(qemu_file, sizeof(qemu_file), 
> -                 "/var/lib/xen/qemu-save.%d", domid);
> +        asprintf(&qemu_file, "/var/lib/xen/qemu-save.%d", domid);

Just a question - afaik asprintf is a unix-ism that isn't supported on
windows. Will this cause problems if also pushed to upstream qemu?

Mike
-- 
Mike D. Day
IBM LTC
Cell: 919 412-3900
Sametime: ncmike@us.ibm.com AIM: ncmikeday  Yahoo: ultra.runner
PGP key: http://www.ncultra.org/ncmike/pubkey.asc

Re: ioemu: use asprintf instead of PATH_MAX

[Samuel Thibault]

Mike D. Day, le Wed 23 Jan 2008 15:53:47 -0500, a écrit :
> erOn 23/01/08 17:20 +0000, Samuel Thibault wrote:
> > Use asprintf instead of PATH_MAX, which POSIX says to be facultative.
> > 
> > Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
> 
> > -        snprintf(qemu_file, sizeof(qemu_file), 
> > -                 "/var/lib/xen/qemu-save.%d", domid);
> > +        asprintf(&qemu_file, "/var/lib/xen/qemu-save.%d", domid);
> 
> Just a question - afaik asprintf is a unix-ism that isn't supported on
> windows. Will this cause problems if also pushed to upstream qemu?

qemu is quite often used in windows too so that would pose problem
indeed.  But I guess the /var/lib patch would too anyway :)  And this
part of the code is actually xen-specific.

Samuel

Re: ioemu: use asprintf instead of PATH_MAX

[Mike D. Day]

On 24/01/08 01:38 +0000, Samuel Thibault wrote:
> Mike D. Day, le Wed 23 Jan 2008 15:53:47 -0500, a écrit :
> > erOn 23/01/08 17:20 +0000, Samuel Thibault wrote:
> > > Use asprintf instead of PATH_MAX, which POSIX says to be facultative.
> > > 
> > > Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
> > 
> > > -        snprintf(qemu_file, sizeof(qemu_file), 
> > > -                 "/var/lib/xen/qemu-save.%d", domid);
> > > +        asprintf(&qemu_file, "/var/lib/xen/qemu-save.%d", domid);
> > 
> > Just a question - afaik asprintf is a unix-ism that isn't supported on
> > windows. Will this cause problems if also pushed to upstream qemu?
> 
> qemu is quite often used in windows too so that would pose problem
> indeed.  But I guess the /var/lib patch would too anyway :)  And this
> part of the code is actually xen-specific.

Yes, true. At a higher scope, what are current plans for a merge with
upstream Qemu? Is the process to merge upstream or to continue
maintaining the Xen differences out-of-tree?

thanks, 

Mike

-- 
Mike D. Day
IBM LTC
Cell: 919 412-3900
Sametime: ncmike@us.ibm.com AIM: ncmikeday  Yahoo: ultra.runner
PGP key: http://www.ncultra.org/ncmike/pubkey.asc

Re: Re: ioemu: use asprintf instead of PATH_MAX

[Keir Fraser]

On 24/1/08 14:45, "Mike D. Day" <ncmike@us.ibm.com> wrote:

>>> Just a question - afaik asprintf is a unix-ism that isn't supported on
>>> windows. Will this cause problems if also pushed to upstream qemu?
>> 
>> qemu is quite often used in windows too so that would pose problem
>> indeed.  But I guess the /var/lib patch would too anyway :)  And this
>> part of the code is actually xen-specific.
> 
> Yes, true. At a higher scope, what are current plans for a merge with
> upstream Qemu? Is the process to merge upstream or to continue
> maintaining the Xen differences out-of-tree?

We'd like the diff to be reduced of course, but it depends if anyone is
enthusiastic enough to do the work.

 -- Keir