On 23.04.2015 04:42, Xiao Guang Chen wrote:
From: Bo Tu <t...@linux.vnet.ibm.com>
Hm, why is Bo Tu the patch author, but doesn't have an S-o-b in the
commit message?
when creating an image qemu-img enable us specifying the size of the
image using -o size=xx options. But when we specify an invalid size
such as a negtive size then different platform gives different result.
parse_option_size() function in util/qemu-option.c will be called to
parse the size, a cast was called in the function to cast the input
(saved as a double in the function) size to an unsigned int64 value,
when the input is a negtive value or exceeds the maximum of uint64, then
the result is undefined.
Language spec 6.3.1.4 Real floating and integers:
the result of this assignment/cast is undefined if the float is not
in the open interval (-1, U<type>_MAX+1).
Thank you for pointing to the specific section. I guess there are always
new things to discover in C...
Signed-off-by: Xiao Guang Chen <che...@linux.vnet.ibm.com>
---
tests/qemu-iotests/049.out | 10 ++++------
util/qemu-option.c | 5 +++++
2 files changed, 9 insertions(+), 6 deletions(-)
diff --git a/tests/qemu-iotests/049.out b/tests/qemu-iotests/049.out
index 9f93666..75d90b2 100644
--- a/tests/qemu-iotests/049.out
+++ b/tests/qemu-iotests/049.out
@@ -95,17 +95,15 @@ qemu-img create -f qcow2 TEST_DIR/t.qcow2 -- -1024
qemu-img: Image size must be less than 8 EiB!
qemu-img create -f qcow2 -o size=-1024 TEST_DIR/t.qcow2
-qemu-img: qcow2 doesn't support shrinking images yet
-qemu-img: TEST_DIR/t.qcow2: Could not resize image: Operation not supported
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=-1024 encryption=off
cluster_size=65536 lazy_refcounts=off refcount_bits=16
+qemu-img: Parameter 'size' expects a positive number and must not exceeds the
maximum UINT64
+qemu-img: TEST_DIR/t.qcow2: Invalid options for file format 'qcow2'
qemu-img create -f qcow2 TEST_DIR/t.qcow2 -- -1k
qemu-img: Image size must be less than 8 EiB!
qemu-img create -f qcow2 -o size=-1k TEST_DIR/t.qcow2
-qemu-img: qcow2 doesn't support shrinking images yet
-qemu-img: TEST_DIR/t.qcow2: Could not resize image: Operation not supported
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=-1024 encryption=off
cluster_size=65536 lazy_refcounts=off refcount_bits=16
+qemu-img: Parameter 'size' expects a positive number and must not exceeds the
maximum UINT64
+qemu-img: TEST_DIR/t.qcow2: Invalid options for file format 'qcow2'
qemu-img create -f qcow2 TEST_DIR/t.qcow2 -- 1kilobyte
qemu-img: Invalid image size specified! You may use k, M, G, T, P or E
suffixes for
diff --git a/util/qemu-option.c b/util/qemu-option.c
index fda4e5f..1c50fa4 100644
--- a/util/qemu-option.c
+++ b/util/qemu-option.c
@@ -179,6 +179,11 @@ void parse_option_size(const char *name, const char *value,
if (value != NULL) {
sizef = strtod(value, &postfix);
+ if (sizef < 0 || sizef > UINT64_MAX) {
+ error_set(errp, QERR_INVALID_PARAMETER_VALUE, name, "a positive "
+ "number and must not exceeds the maximum UINT64");
I think Markus would like to see these error macros not getting used
anymore, so I think it should be dropped and the full string should be
given here. I'll let him do the arguing, though. :-)
If you keep the macro, I'd propose "a non-negative number below 2^64"
(or actually give the decimal value of UINT64_MAX, using 'a non-negative
number not exceeding "%" PRIu64, UINT64_MAX'). Remember that 0 is not
positive, but still a valid choice.
If you drop the macro, I'd propose error_setg(errp, "'%s' must be a
non-negative number below 2^64", name) or, like it is now,
error_setg(errp, "Parameter '%s' expects a non-negative number below
2^64", name).
Max
+ return;
+ }
switch (*postfix) {
case 'T':
sizef *= 1024;
