Re: [libvirt] [PATCH v3] buf: Fix possible infinite loop in EscapeString, VSnprintf

On 09/09/2010 08:09 AM, Cole Robinson wrote: > The current code will go into an infinite loop if the printf generated > string is>= 1000, AND exactly 1 character smaller than the amount of free > space in the buffer. When this happens, we are dropped into the loop body, > but nothing will actually change, because count == (buf->size - buf->use - 1), > and virBufferGrow returns unchanged if count< (buf->size - buf->use) > > Fix this by removing the '- 1' bit from 'size'. The *nprintf functions handle > the NULL byte for us anyways, so we shouldn't need to manually accommodate > for it. > > Here's a bug where we are actually hitting this issue: > https://bugzilla.redhat.com/show_bug.cgi?id=602772 > > v2: Eric's improvements: while -> if (), remove extra va_list variable, > make sure we report buffer error if snprintf fails > > v3: Add tests/virbuftest which reproduces the infinite loop before this > patch, works correctly after I confirm that the only change in v3 is the addition of a test. > +static int testBufInfiniteLoop(const void *data ATTRIBUTE_UNUSED) > +{ > + virBuffer bufinit = VIR_BUFFER_INITIALIZER; > + virBufferPtr buf =&bufinit; > + char *addstr = NULL, *bufret = NULL; > + int ret = -1; > + const struct testInfo *info = data; > + > + virBufferAddChar(buf, 'a'); > + if (buf->a != 1002 || buf->b != 1) { > + TEST_ERROR("Buf was not expected size, size=%d use=%d\n", > + buf->a, buf->b); This seems a bit fragile; it is inspecting the innards of virBuffer. Maybe add a comment here that any change to the innards of virBuffer may warrant a change to this test at the same time. But having the test, even if it is fragile, is a good thing, so... ACK.

We could add further tests to this file to cover more lines of code, such as: printing a string that needs escapes added, and actually verifying that the escaped string matches expectations; passing a bad format string (such as "%$2s" to trigger EINVAL, "%*d%*d",1<<31,1,1<<31,1 to trigger EOVERFLOW) and verifying that error gets set. But those can be in a separate patch; let's get the bug fix in now before 0.8.4.