[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH] Fix stubdom build failure for RELEASE-4.3.1



On 31/10/13 14:44, Alex Sharp wrote:
Hey,

I noticed that RELEASE-4.3.1 (commit 4e8e0bdef96859e6f428af74be6f4a3bc7fd11f3) was failing to make stubdom, complaining of various forms of format string mismatch. The following patch at least gets things to compile, but I'm not sure if there's a deeper underlying issue. I'm using gcc version 4.8.1 (Ubuntu/Linaro 4.8.1-10ubuntu8) on x86_64. My .config contains the line "PYTHON_PREFIX_ARG=--install-layout=deb" and I'm ./configuring with --prefix=

Hope this helps.

diff --git a/extras/mini-os/fbfront.c b/extras/mini-os/fbfront.c
index 54a5e67..b6aaa6b 100644
--- a/extras/mini-os/fbfront.c
+++ b/extras/mini-os/fbfront.c
@@ -105,7 +105,7 @@ again:
         free(err);
     }
 
-    err = xenbus_printf(xbt, nodename, "page-ref","%u", virt_to_mfn(s));
+    err = xenbus_printf(xbt, nodename, "page-ref","%lu", virt_to_mfn(s));
     if (err) {
         message = "writing page-ref";
         goto abort_transaction;
@@ -463,7 +463,7 @@ again:
         free(err);
     }
 
-    err = xenbus_printf(xbt, nodename, "page-ref","%u", virt_to_mfn(s));
+    err = xenbus_printf(xbt, nodename, "page-ref","%lu", virt_to_mfn(s));

The two above here are certainly valid.  virt_to_mfn() goes through many macros and ends up reading out of "unsigned long *phys_to_machine_mapping".

     if (err) {
         message = "writing page-ref";
         goto abort_transaction;
diff --git a/extras/mini-os/pcifront.c b/extras/mini-os/pcifront.c
index bbe21e0..9d0cb17 100644
--- a/extras/mini-os/pcifront.c
+++ b/extras/mini-os/pcifront.c
@@ -413,7 +413,7 @@ int pcifront_physical_to_virtual (struct pcifront_dev *dev,
                 continue;
             }
 
-            if (sscanf(s, "%x:%x:%x.%x", dom, bus, slot, fun) != 4) {
+            if (sscanf(s, "%x:%x:%x.%lx", dom, bus, slot, fun) != 4) {

This is a little awkward.  It has been changed in unstable and is now correct.  Perhaps a subset of the change needs backporting

                 printk("\"%s\" does not look like a PCI device address\n", s);
                 free(s);
                 continue;
diff --git a/extras/mini-os/xenbus/xenbus.c b/extras/mini-os/xenbus/xenbus.c
index ee1691b..fa2e872 100644
--- a/extras/mini-os/xenbus/xenbus.c
+++ b/extras/mini-os/xenbus/xenbus.c
@@ -672,7 +672,7 @@ char *xenbus_transaction_start(xenbus_transaction_t *xbt)
     err = errmsg(rep);
     if (err)
        return err;
-    sscanf((char *)(rep + 1), "%u", xbt);
+    sscanf((char *)(rep + 1), "%lu", xbt);

This is good

     free(rep);
     return NULL;
 }
@@ -769,7 +769,7 @@ domid_t xenbus_get_self_id(void)
     domid_t ret;
 
     BUG_ON(xenbus_read(XBT_NIL, "domid", &dom_id));
-    sscanf(dom_id, "%d", &ret);
+    sscanf(dom_id, "%u", (unsigned int*)&ret);

This needs to be

sscanf(dom_id, "%"SCNd16, &ret);

That pointer casting there will result in stack corruption.

Could you please follow the instructions http://wiki.xen.org/wiki/Submitting_Xen_Patches (specifically the Signed-off-by line) and submit a patch against xen-unstable, which can then be backported appropriately.

~Andrew

 
     return ret;
 }



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.