pstore/ram: Add ftrace messages handling

Hello Anton Vorontsov,

This is a semi-automatic email about new static checker warnings.

The patch a694d1b5916a: "pstore/ram: Add ftrace messages handling" from Jul 9, 2012, leads to the following Smatch complaint:

fs/pstore/ram.c:423 ramoops_probe()

	 error: we previously assumed 'cxt->cprz' could be null (see line 408)

fs/pstore/ram.c
407

   408		if (!cxt->przs && !cxt->cprz && !cxt->fprz) {
                                  ^^^^^^^^^^

Checked here.

409 pr_err("memory size too small, minimum is %lu\n",
410 cxt->console_size + cxt->record_size +
411 cxt->ftrace_size);
412 goto fail_cnt;
413 }
414
415 cxt->pstore.data = cxt;
416 /*
417 * Console can handle any buffer size, so prefer dumps buffer
418 * size since usually it is smaller.
419 */
420 if (cxt->przs)
421 cxt->pstore.bufsize = cxt->przs[0]->buffer_size;
422 else

   423			cxt->pstore.bufsize = cxt->cprz->buffer_size;
                                              ^^^^^^^^^

Dereferenced here. What about if only "cxt->fprz" is non-NULL?

Also these are crap variable names, "przs" and "cprz" look so similar. It makes my head hurt to keep them appart.

   424		cxt->pstore.buf = kmalloc(cxt->pstore.bufsize, GFP_KERNEL);
   425		spin_lock_init(&cxt->pstore.buf_lock);

regards,
dan carpenter

Hi Dan,

A nice tool. The homepage of Smatch doesn't explicitly say that, so I have to ask: is it a complete superset of sparse (i.e. does it produce all the warnings that the pure sparse can produce)? If so, I'll probably switch to it from the vanilla sparse.

Thanks,

Anton Vorontsov
Email: [email protected]

No. It just uses Sparse as a parser.

regards,
dan carpenter

We can dereference 'cxt->cprz' if console and dump logging are disabled (which is unlikely, but still possible to do). This patch fixes the issue by changing the code so that we don't dereference przs at all, we can just calculate bufsize from console_size and record_size values.

Plus, while at it, the patch improves the buffer size calculation.

After Kay's printk rework, we know the optimal buffer size for console logging -- it is LOG_LINE_MAX (defined privately in printk.c). Previously, if only console logging was enabled, we would allocate unnecessary large buffer in pstore, while we only need LOG_LINE_MAX. (Pstore console logging is still capable of handling buffers > LOG_LINE_MAX, it will just do multiple calls to psinfo->write).

Note that I don't export the constant, since we will do even a better thing soon: we will switch console logging to a new write_buf API, which will eliminate the need for the additional buffer; and so we won't need the constant.

---
 fs/pstore/ram.c |   13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

index b86b2b7..c34fccf 100644
@@ -414,13 +414,14 @@ static int __devinit ramoops_probe(struct platform_device *pdev)
 
 	cxt->pstore.data = cxt;
 	/*
-	 * Console can handle any buffer size, so prefer dumps buffer
-	 * size since usually it is smaller.
+	 * Console can handle any buffer size, so prefer LOG_LINE_MAX. If we
+	 * have to handle dumps, we must have at least record_size buffer. And
+	 * for ftrace, bufsize is irrelevant (if bufsize is 0, buf will be
+	 * ZERO_SIZE_PTR).
 	 */
-	if (cxt->przs)
-		cxt->pstore.bufsize = cxt->przs[0]->buffer_size;
-	else
-		cxt->pstore.bufsize = cxt->cprz->buffer_size;
+	if (cxt->console_size)
+		cxt->pstore.bufsize = 1024; /* LOG_LINE_MAX */
+	cxt->pstore.bufsize = max(cxt->record_size, cxt->pstore.bufsize);
 	cxt->pstore.buf = kmalloc(cxt->pstore.bufsize, GFP_KERNEL);
 	spin_lock_init(&cxt->pstore.buf_lock);
 	if (!cxt->pstore.buf) {

1.7.10.4

-Kees

Kees Cook
Chrome OS Security