[PATCH 3.16 218/410] crypto: caam - fix endless loop when DECO acquire fails

[Ben Hutchings]

3.16.57-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Horia GeantÄ <horia.geanta@xxxxxxx>

commit 225ece3e7dad4cfc44cca38ce7a3a80f255ea8f1 upstream.

In case DECO0 cannot be acquired - i.e. run_descriptor_deco0() fails
with -ENODEV, caam_probe() enters an endless loop:

run_descriptor_deco0
	ret -ENODEV
	-> instantiate_rng
		-ENODEV, overwritten by -EAGAIN
		ret -EAGAIN
		-> caam_probe
			-EAGAIN results in endless loop

It turns out the error path in instantiate_rng() is incorrect,
the checks are done in the wrong order.

Fixes: 1005bccd7a4a6 ("crypto: caam - enable instantiation of all RNG4 state handles")
Reported-by: Bryan O'Donoghue <pure.logic@xxxxxxxxxxxxxxxxx>
Suggested-by: Auer Lukas <lukas.auer@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Horia GeantÄ <horia.geanta@xxxxxxx>
Signed-off-by: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
 drivers/crypto/caam/ctrl.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

--- a/drivers/crypto/caam/ctrl.c
+++ b/drivers/crypto/caam/ctrl.c
@@ -196,12 +196,16 @@ static int instantiate_rng(struct device
 		 * without any error (HW optimizations for later
 		 * CAAM eras), then try again.
 		 */
+		if (ret)
+			break;
+
 		rdsta_val =
 			rd_reg32(&topregs->ctrl.r4tst[0].rdsta) & RDSTA_IFMASK;
-		if (status || !(rdsta_val & (1 << sh_idx)))
+		if (status || !(rdsta_val & (1 << sh_idx))) {
 			ret = -EAGAIN;
-		if (ret)
 			break;
+		}
+
 
 		dev_info(ctrldev, "Instantiated RNG4 SH%d\n", sh_idx);
 		/* Clear the contents before recreating the descriptor */