[PATCH v2 1/2] drivers: caam: add RNG software self-test

[PATCH v2 1/2] drivers: caam: add RNG software self-test

[Roland Hieber]

This patch is based on a vendor patch in U-Boot, taken from
https://portland.source.codeaurora.org/patches/external/imxsupport/uboot-imx/imx_v2016.03_4.1.15_2.0.0_ga/HAB-238-Run-RNG-self-test-for-impacted-i.MX-chips.zip

|    HAB-238 Run RNG self test for impacted i.MX chips
|
|    Patch is only applicable to imx_v2016.03_4.1.15_2.0.0_ga branch of u-boot.
|    Please adapt the patch for your respective release version.
|
|    Background:
|    Few i.MX chips which have HAB 4.2.3 or beyond, have oberserved following
|    warning message generated by HAB due to incorrect implementation of drng
|    self test in boot ROM.
|
|        Event       |0xdb|0x0024|0x42| SRCE Field: 69 30 e1 1d
|                    |    |      |    |             STS = HAB_WARNING (0x69)
|                    |    |      |    |             RSN = HAB_ENG_FAIL (0x30)
|                    |    |      |    |             CTX = HAB_CTX_ENTRY (0xE1)
|                    |    |      |    |             ENG = HAB_ENG_CAAM (0x1D)
|                    |    |      |    | Evt Data (hex):
|                    |    |      |    |  00 08 00 02 40 00 36 06 55 55 00 03 00 00 00 00
|                    |    |      |    |  00 00 00 00 00 00 00 00 00 00 00 01
|
|    It is recommended to run this rng self test before any RNG related crypto
|    implementations are done.
|
[...]
|
|    Signed-off-by: Utkarsh Gupta <utkarsh.gupta@nxp.com>

Currently known impacted chips, as determined by NXP, include:
* i.MX6DQ+ silicon revision 1.1
* i.MX6DQ silicon revision 1.6
* i.MX6DLS silicon revision 1.4
* i.MX6SX silicon revision 1.4
* i.MX6UL silicon revision 1.2
* i.MX67SD silicon revision 1.3

Port the RNG software self-test from this patch to barebox. It can be
enabled by selecting CRYPTO_DEV_FSL_CAAM_RNG_SELF_TEST in Kconfig.

The original patch included a command line utility to run the self-test,
but we choose a different approach here, and run the software self-test
automatically when the respective HAB events indicating a RNG ROM
self-test failure are found when running habv4_get_status(). Note that
habv4_get_status() must be called by the board code before the CAAM
device driver is probed for this mechanism to work.

Until now there are at least two such known events. The first event was
observed on an i.MX6Solo, silicon revision 1.4; the second event is
mentioned in the original patch description given above. When an event
occured, habv4_get_status() tests if it is one of those known events,
and if so, indicates to the CAAM driver to run the software self-test.
In this case, printing the respective HAB warning is suppressed to
prevent confusion; the software self-test itself will error out in case
of recurring RNG self-test failure.

Signed-off-by: Roland Hieber <r.hieber@pengutronix.de>
---
v1 -> v2:
  * less pr_info and pr_err, more pr_debug where useful
  * habv4:
    - move external bool habv4_need_rng_software_self_test to <hab.h>
    - refactor comparison with known fail events into own function
      is_known_rng_fail_event()
  * rng_self_test:
    - rename rng_self_test() -> caam_rng_self_test()
    - remove unneeded sec_in32()
    - determine length of rng_{result,dsc}{1,2} arrays at compile time
      and use ARRAY_SIZE on them
    - make rng_result{1,2} arrays uint8 instead of uint32, and also fix
      their previously bogus dma_alloc sizes
    - use memcmp() instead of for loop for comparing results
    - dma_free(result) after use
    - improve logic in the caam_era/rng_vid/rng_rev check
    - simplify error handling in construct_rng_self_test_jobdesc(), can
      now return void
    - rng_self_test_done: remove unneccessary BUG_ON null pointer check
    - rng_self_test_done: remove unneccessary void* -> int* cast
---
 drivers/crypto/caam/Kconfig         |  24 +++
 drivers/crypto/caam/Makefile        |   1 +
 drivers/crypto/caam/ctrl.c          |  20 +++
 drivers/crypto/caam/regs.h          |   4 +
 drivers/crypto/caam/rng_self_test.c | 221 ++++++++++++++++++++++++++++
 drivers/crypto/caam/rng_self_test.h |  24 +++
 drivers/hab/habv4.c                 |  51 ++++++-
 include/hab.h                       |   1 +
 8 files changed, 343 insertions(+), 3 deletions(-)
 create mode 100644 drivers/crypto/caam/rng_self_test.c
 create mode 100644 drivers/crypto/caam/rng_self_test.h

diff --git a/drivers/crypto/caam/Kconfig b/drivers/crypto/caam/Kconfig
index 2ab509d110..56b90700b8 100644
--- a/drivers/crypto/caam/Kconfig
+++ b/drivers/crypto/caam/Kconfig
@@ -33,3 +33,27 @@ config CRYPTO_DEV_FSL_CAAM_RNG
 	default y
 	help
 	  Selecting this will register the SEC4 hardware rng.
+
+if CRYPTO_DEV_FSL_CAAM_RNG
+
+config CRYPTO_DEV_FSL_CAAM_RNG_SELF_TEST
+	bool "Run RNG software self-test on impacted chips"
+	depends on ARCH_IMX6
+	depends on HABV4
+	default y
+	help
+	  Some chips with HAB >= 4.2.3 have an incorrect implementation of the
+	  RNG self-test in ROM code. In this case, a software self-test should
+	  be run to ensure correctness of the RNG. By enabling this config
+	  option, the software self-test is run automatically when this case
+	  is detected.
+
+	  Currently known impacted chips:
+	  * i.MX6DQ+ silicon revision 1.1
+	  * i.MX6DQ silicon revision 1.6
+	  * i.MX6DLS silicon revision 1.4
+	  * i.MX6SX silicon revision 1.4
+	  * i.MX6UL silicon revision 1.2
+	  * i.MX67SD silicon revision 1.3
+
+endif
diff --git a/drivers/crypto/caam/Makefile b/drivers/crypto/caam/Makefile
index 74d32da00e..7bd6f3e23c 100644
--- a/drivers/crypto/caam/Makefile
+++ b/drivers/crypto/caam/Makefile
@@ -3,3 +3,4 @@
 #
 obj-$(CONFIG_CRYPTO_DEV_FSL_CAAM) += ctrl.o error.o jr.o
 obj-$(CONFIG_CRYPTO_DEV_FSL_CAAM_RNG) += caamrng.o
+obj-$(CONFIG_CRYPTO_DEV_FSL_CAAM_RNG_SELF_TEST) += rng_self_test.o
diff --git a/drivers/crypto/caam/ctrl.c b/drivers/crypto/caam/ctrl.c
index 9e62bd6fd6..39910131b1 100644
--- a/drivers/crypto/caam/ctrl.c
+++ b/drivers/crypto/caam/ctrl.c
@@ -8,6 +8,7 @@
 #include <common.h>
 #include <clock.h>
 #include <driver.h>
+#include <hab.h>
 #include <init.h>
 #include <linux/barebox-wrapper.h>
 #include <linux/spinlock.h>
@@ -19,6 +20,7 @@
 #include "desc_constr.h"
 #include "error.h"
 #include "ctrl.h"
+#include "rng_self_test.h"
 
 bool caam_little_end;
 EXPORT_SYMBOL(caam_little_end);
@@ -570,6 +572,24 @@ static int caam_probe(struct device_d *dev)
 
 	cha_vid_ls = rd_reg32(&ctrl->perfmon.cha_id_ls);
 
+	/* habv4_need_rng_software_self_test is determined by habv4_get_status() */
+	if (IS_ENABLED(CONFIG_CRYPTO_DEV_FSL_CAAM_RNG_SELF_TEST) &&
+	    habv4_need_rng_software_self_test) {
+		u8 caam_era;
+		u8 rngvid;
+		u8 rngrev;
+
+		caam_era = (rd_reg32(&ctrl->perfmon.ccb_id) & CCBVID_ERA_MASK) >> CCBVID_ERA_SHIFT;
+		rngvid = (cha_vid_ls & CHAVID_LS_RNGVID_MASK) >> CHAVID_LS_RNGVID_SHIFT;
+		rngrev = (rd_reg32(&ctrl->perfmon.cha_rev_ls) & CRNR_LS_RNGRN_MASK) >> CRNR_LS_RNGRN_SHIFT;
+
+		ret = caam_rng_self_test(ctrlpriv->jrpdev[0], caam_era, rngvid, rngrev);
+		if (ret != 0) {
+			caam_remove(dev);
+			return ret;
+		}
+	}
+
 	/*
 	 * If SEC has RNG version >= 4 and RNG state handle has not been
 	 * already instantiated, do RNG instantiation
diff --git a/drivers/crypto/caam/regs.h b/drivers/crypto/caam/regs.h
index 6c9d6d75a0..19e7d6d7e4 100644
--- a/drivers/crypto/caam/regs.h
+++ b/drivers/crypto/caam/regs.h
@@ -279,6 +279,8 @@ struct caam_perfmon {
 
 	/* CAAM Hardware Instantiation Parameters		fa0-fbf */
 	u32 cha_rev_ms;		/* CRNR - CHA Rev No. Most significant half*/
+#define CRNR_LS_RNGRN_SHIFT	16
+#define CRNR_LS_RNGRN_MASK	(0xfull << CRNR_LS_RNGRN_SHIFT)
 	u32 cha_rev_ls;		/* CRNR - CHA Rev No. Least significant half*/
 #define CTPR_MS_QI_SHIFT	25
 #define CTPR_MS_QI_MASK		(0x1ull << CTPR_MS_QI_SHIFT)
@@ -311,6 +313,8 @@ struct caam_perfmon {
 #define CCBVID_ERA_SHIFT	24
 	u32 ccb_id;		/* CCBVID - CCB Version ID	*/
 	u32 cha_id_ms;		/* CHAVID - CHA Version ID Most Significant*/
+#define CHAVID_LS_RNGVID_SHIFT	16
+#define CHAVID_LS_RNGVID_MASK	(0xfull << CHAVID_LS_RNGVID_SHIFT)
 	u32 cha_id_ls;		/* CHAVID - CHA Version ID Least Significant*/
 	u32 cha_num_ms;		/* CHANUM - CHA Number Most Significant	*/
 	u32 cha_num_ls;		/* CHANUM - CHA Number Least Significant*/
diff --git a/drivers/crypto/caam/rng_self_test.c b/drivers/crypto/caam/rng_self_test.c
new file mode 100644
index 0000000000..91cfbb9f9a
--- /dev/null
+++ b/drivers/crypto/caam/rng_self_test.c
@@ -0,0 +1,221 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2018 NXP
+ * Copyright (C) 2018 Pengutronix, Roland Hieber <r.hieber@pengutronix.de>
+ *
+ * CAAM RNG self-test -- based on the vendor patch for U-Boot:
+ * https://portland.source.codeaurora.org/patches/external/imxsupport/uboot-imx/imx_v2016.03_4.1.15_2.0.0_ga/HAB-238-Run-RNG-self-test-for-impacted-i.MX-chips.zip
+ *
+ * |    From: Utkarsh Gupta <utkarsh.gupta@nxp.com>
+ * |    Subject: [PATCH] HAB-238 Run RNG self test for impacted i.MX chips
+ * |
+ * |    Patch is only applicable to imx_v2016.03_4.1.15_2.0.0_ga branch of u-boot.
+ * |    Please adapt the patch for your respective release version.
+ * |
+ * |    Background:
+ * |    Few i.MX chips which have HAB 4.2.3 or beyond, have oberserved following
+ * |    warning message generated by HAB due to incorrect implementation of drng
+ * |    self test in boot ROM.
+ * |
+ * |        Event       |0xdb|0x0024|0x42| SRCE Field: 69 30 e1 1d
+ * |                    |    |      |    |             STS = HAB_WARNING (0x69)
+ * |                    |    |      |    |             RSN = HAB_ENG_FAIL (0x30)
+ * |                    |    |      |    |             CTX = HAB_CTX_ENTRY (0xE1)
+ * |                    |    |      |    |             ENG = HAB_ENG_CAAM (0x1D)
+ * |                    |    |      |    | Evt Data (hex):
+ * |                    |    |      |    |  00 08 00 02 40 00 36 06 55 55 00 03 00 00 00 00
+ * |                    |    |      |    |  00 00 00 00 00 00 00 00 00 00 00 01
+ * |
+ * |    It is recommended to run this rng self test before any RNG related crypto
+ * |    implementations are done.
+ * [...]
+ * |
+ * |    Signed-off-by: Utkarsh Gupta <utkarsh.gupta@nxp.com>
+ *
+ * Known impacted chips:
+ *
+ * - i.MX6DQ+ silicon revision 1.1
+ * - i.MX6DQ silicon revision 1.6
+ * - i.MX6DLS silicon revision 1.4
+ * - i.MX6SX silicon revision 1.4
+ * - i.MX6UL silicon revision 1.2
+ * - i.MX67SD silicon revision 1.3
+ */
+
+#define pr_fmt(fmt) "rng_self_test: " fmt
+
+#include <dma.h>
+#include <common.h>
+#include <linux/kernel.h>
+
+#include "error.h"
+#include "regs.h"
+#include "jr.h"
+
+static const u32 rng_dsc1[] = {
+	0xb0800036, 0x04800010, 0x3c85a15b, 0x50a9d0b1,
+	0x71a09fee, 0x2eecf20b, 0x02800020, 0xb267292e,
+	0x85bf712d, 0xe85ff43a, 0xa716b7fb, 0xc40bb528,
+	0x27b6f564, 0x8821cb5d, 0x9b5f6c26, 0x12a00020,
+	0x0a20de17, 0x6529357e, 0x316277ab, 0x2846254e,
+	0x34d23ba5, 0x6f5e9c32, 0x7abdc1bb, 0x0197a385,
+	0x82500405, 0xa2000001, 0x10880004, 0x00000005,
+	0x12820004, 0x00000020, 0x82500001, 0xa2000001,
+	0x10880004, 0x40000045, 0x02800020, 0x8f389cc7,
+	0xe7f7cbb0, 0x6bf2073d, 0xfc380b6d, 0xb22e9d1a,
+	0xee64fcb7, 0xa2b48d49, 0xdf9bc3a4, 0x82500009,
+	0xa2000001, 0x10880004, 0x00000005, 0x82500001,
+	0x60340020, 0xFFFFFFFF, 0xa2000001, 0x10880004,
+	0x00000005, 0x8250000d
+};
+
+static const u8 rng_result1[] = {
+	0x3a, 0xfe, 0x2c, 0x87, 0xcc, 0xb6, 0x44, 0x49,
+	0x19, 0x16, 0x9a, 0x74, 0xa1, 0x31, 0x8b, 0xef,
+	0xf4, 0x86, 0x0b, 0xb9, 0x5e, 0xee, 0xae, 0x91,
+	0x92, 0xf4, 0xa9, 0x8f, 0xb0, 0x37, 0x18, 0xa4
+};
+
+static const u32 rng_dsc2[] = {
+	0xb080003a, 0x04800020, 0x27b73130, 0x30b4b10f,
+	0x7c62b1ad, 0x77abe899, 0x67452301, 0xefcdab89,
+	0x98badcfe, 0x10325476, 0x02800020, 0x63f757cf,
+	0xb9165584, 0xc3c1b407, 0xcc4ce8ad, 0x1ffe8a58,
+	0xfb4fa893, 0xbb5f4af0, 0x3fb946a1, 0x12a00020,
+	0x56cbcaa5, 0xfff3adad, 0xe804dcbf, 0x9a900c71,
+	0xa42017e3, 0x826948e2, 0xd0cfeb3e, 0xaf1a136a,
+	0x82500405, 0xa2000001, 0x10880004, 0x00000005,
+	0x12820004, 0x00000020, 0x82500001, 0xa2000001,
+	0x10880004, 0x40000045, 0x02800020, 0x2e882f8a,
+	0xe929943e, 0x8132c0a8, 0x12037f90, 0x809fbd66,
+	0x8684ea04, 0x00cbafa7, 0x7b82d12a, 0x82500009,
+	0xa2000001, 0x10880004, 0x00000005, 0x82500001,
+	0x60340020, 0xFFFFFFFF, 0xa2000001, 0x10880004,
+	0x00000005, 0x8250000d
+};
+
+static const u8 rng_result2[] = {
+	0x76, 0x87, 0x66, 0x4e, 0xd8, 0x1d, 0x1f, 0x43,
+	0x76, 0x50, 0x85, 0x5d, 0x1e, 0x1d, 0x9d, 0x0f,
+	0x93, 0x75, 0x83, 0xff, 0x9a, 0x9b, 0x61, 0xa9,
+	0xa5, 0xeb, 0xa3, 0x28, 0x2a, 0x15, 0xc1, 0x57
+};
+
+/*
+ * construct_rng_self_test_jobdesc() - Implement destination address in RNG self test descriptors
+ * Returns zero on success, and negative on error.
+ */
+static void construct_rng_self_test_jobdesc(u32 *desc, const u32 *rng_st_dsc, u8 *res_addr, int desc_size)
+{
+	int result_addr_idx = desc_size - 5;
+	int i;
+
+	for (i = 0; i < desc_size; i++) {
+		desc[i] = rng_st_dsc[i];
+	}
+
+	/* Replace destination address in the descriptor */
+	desc[result_addr_idx] = (u32)res_addr;
+
+	pr_vdebug("RNG SELF TEST DESCRIPTOR:\n");
+	for (i = 0; i < desc_size; i++) {
+		pr_vdebug("0x%08X\n", desc[i]);
+	}
+}
+
+/* rng_self_test_done() - callback for caam_jr_enqueue */
+static void rng_self_test_done(struct device_d *dev, u32 *desc, u32 err, void *arg)
+{
+	int * job_err = arg;
+	*job_err = err;
+}
+
+/*
+ * caam_rng_self_test() - Perform RNG self test
+ * Returns zero on success, and negative on error.
+ */
+int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid, const u8 rngrev)
+{
+	int ret, i, desc_size = 0, result_size = 0, job_err = 0;
+	const u32 *rng_st_dsc;
+	const u8 *exp_result;
+	u32 *desc;
+	u8 *result;
+
+	pr_debug("got CAAM ERA %d, RNG Version ID %d, RNG revision %d\n", caam_era, rngvid, rngrev);
+
+	if (caam_era < 8 && rngvid == 4 && rngrev < 3) {
+		// older affected i.MX chipsets have CAAM < 8 and have RNG4 < 4.3
+		rng_st_dsc = rng_dsc1;
+		desc_size = ARRAY_SIZE(rng_dsc1);
+		exp_result = rng_result1;
+		result_size = ARRAY_SIZE(rng_result1);
+	} else if (caam_era >= 8 || (rngvid >= 4 && rngrev >= 3)) {
+		// newer affected chipsets have CAAM >= 8 or RNG4 >= 4.3
+		rng_st_dsc = rng_dsc2;
+		desc_size = ARRAY_SIZE(rng_dsc2);
+		exp_result = rng_result2;
+		result_size = ARRAY_SIZE(rng_result2);
+	} else {
+		pr_err("Error: Invalid CAAM ERA (%d) or RNG Version ID (%d) or RNG revision (%d)\n",
+				caam_era, rngvid, rngrev);
+		return -EINVAL;
+	}
+
+	result = dma_alloc(sizeof(*result) * result_size);
+	if (!result) {
+		return -ENOMEM;
+	}
+
+	desc = dma_alloc(sizeof(*desc) * desc_size);
+	if (!desc) {
+		return -ENOMEM;
+	}
+
+	construct_rng_self_test_jobdesc(desc, rng_st_dsc, result, desc_size);
+
+	dma_sync_single_for_device((unsigned long)desc,
+			desc_size * sizeof(*desc), DMA_TO_DEVICE);
+	dma_sync_single_for_device((unsigned long)result,
+			result_size * sizeof(*result), DMA_FROM_DEVICE);
+
+	/* wait for job completion */
+	ret = caam_jr_enqueue(dev, desc, rng_self_test_done, &job_err);
+	if (ret) {
+		pr_err("Error while running RNG self-test descriptor: %d %s\n",
+				ret, strerror(ret));
+		goto err;
+	}
+	if (job_err) {
+		pr_err("Job Error:\n");
+		caam_jr_strstatus(dev, job_err);
+		goto err;
+	}
+
+	dma_sync_single_for_cpu((unsigned long)result, result_size * sizeof(*result),
+			DMA_FROM_DEVICE);
+
+	pr_vdebug("Result\n");
+	for (i = 0; i < result_size; i++) {
+		pr_vdebug("%02X\n", result[i]);
+	}
+
+	pr_vdebug("Expected Result:\n");
+	for (i = 0; i < result_size; i++) {
+		pr_vdebug("%02X\n", exp_result[i]);
+	}
+
+	if (memcmp(result, exp_result, sizeof(*result) * result_size) != 0) {
+		pr_err("RNG self-test failed with unexpected result\n");
+		ret = -ERANGE;
+		goto err;
+	}
+
+	pr_notice("RNG software self-test passed\n");
+	ret = 0;
+
+err:
+	dma_free(desc);
+	dma_free(result);
+	return ret;
+}
diff --git a/drivers/crypto/caam/rng_self_test.h b/drivers/crypto/caam/rng_self_test.h
new file mode 100644
index 0000000000..1f5bf32628
--- /dev/null
+++ b/drivers/crypto/caam/rng_self_test.h
@@ -0,0 +1,24 @@
+/*
+ * CAAM RNG self test
+ *
+ * Copyright (C) 2018 Pengutronix, Roland Hieber <r.hieber@pengutronix.de>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef RNG_SELF_TEST_H
+#define RNG_SELF_TEST_H
+
+int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid, const u8 rngrev);
+
+#endif /* RNG_SELF_TEST_H */
diff --git a/drivers/hab/habv4.c b/drivers/hab/habv4.c
index 28fd42ecd7..e33a8784a0 100644
--- a/drivers/hab/habv4.c
+++ b/drivers/hab/habv4.c
@@ -387,6 +387,39 @@ static void habv4_display_event(uint8_t *data, uint32_t len)
 	habv4_display_event_record((struct hab_event_record *)data);
 }
 
+/* Some chips with HAB >= 4.2.3 have an incorrect implementation of the RNG
+ * self-test in ROM code. In this case, an HAB event is generated, and a
+ * software self-test should be run. This variable is set to @c true by
+ * habv4_get_status() when this occurs. */
+bool habv4_need_rng_software_self_test = false;
+EXPORT_SYMBOL(habv4_need_rng_software_self_test);
+
+#define RNG_FAIL_EVENT_SIZE 36
+static uint8_t habv4_known_rng_fail_events[][RNG_FAIL_EVENT_SIZE] = {
+	{ 0xdb, 0x00, 0x24, 0x42,  0x69, 0x30, 0xe1, 0x1d,
+	  0x00, 0x80, 0x00, 0x02,  0x40, 0x00, 0x36, 0x06,
+	  0x55, 0x55, 0x00, 0x03,  0x00, 0x00, 0x00, 0x00,
+	  0x00, 0x00, 0x00, 0x00,  0x00, 0x00, 0x00, 0x00,
+	  0x00, 0x00, 0x00, 0x01 },
+	{ 0xdb, 0x00, 0x24, 0x42,  0x69, 0x30, 0xe1, 0x1d,
+	  0x00, 0x04, 0x00, 0x02,  0x40, 0x00, 0x36, 0x06,
+	  0x55, 0x55, 0x00, 0x03,  0x00, 0x00, 0x00, 0x00,
+	  0x00, 0x00, 0x00, 0x00,  0x00, 0x00, 0x00, 0x00,
+	  0x00, 0x00, 0x00, 0x01 },
+};
+
+static bool is_known_rng_fail_event(const uint8_t *data, size_t len)
+{
+	int i;
+	for (i = 0; i < ARRAY_SIZE(habv4_known_rng_fail_events); i++) {
+		if (memcmp(data, habv4_known_rng_fail_events[i],
+			   min(len, (uint32_t)RNG_FAIL_EVENT_SIZE)) == 0) {
+			return true;
+		}
+	}
+	return false;
+}
+
 static int habv4_get_status(const struct habv4_rvt *rvt)
 {
 	uint8_t data[256];
@@ -413,10 +446,22 @@ static int habv4_get_status(const struct habv4_rvt *rvt)
 
 	len = sizeof(data);
 	while (rvt->report_event(HAB_STATUS_WARNING, index, data, &len) == HAB_STATUS_SUCCESS) {
-		pr_err("-------- HAB warning Event %d --------\n", index);
-		pr_err("event data:\n");
 
-		habv4_display_event(data, len);
+		/* suppress RNG self-test fail events if they can be handled in software */
+		bool is_rng_fail_event = false;
+		if (IS_ENABLED(CONFIG_CRYPTO_DEV_FSL_CAAM_RNG_SELF_TEST)) {
+			is_rng_fail_event = is_known_rng_fail_event(data, len);
+		}
+
+		if (is_rng_fail_event) {
+			pr_warning("RNG self-test failure detected, will run software self-test\n");
+			habv4_need_rng_software_self_test = true;
+		} else {
+			pr_err("-------- HAB warning Event %d --------\n", index);
+			pr_err("event data:\n");
+			habv4_display_event(data, len);
+		}
+
 		len = sizeof(data);
 		index++;
 	}
diff --git a/include/hab.h b/include/hab.h
index 78c2b865ba..abfce18736 100644
--- a/include/hab.h
+++ b/include/hab.h
@@ -21,6 +21,7 @@
 #include <errno.h>
 
 #ifdef CONFIG_HABV4
+extern bool habv4_need_rng_software_self_test;
 int imx28_hab_get_status(void);
 int imx6_hab_get_status(void);
 #else
-- 
2.19.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH v2 2/2] i.MX: HABv4: always print HAB status at boot time

[Roland Hieber]

Currently, board code needs to call habv4_get_status() explicitely, but
there is no reason that it cannot be called automatically at startup
when HABv4 is enabled. This way the call cannot be forgotten and we can
make sure to report all potentially occuring HAB warnings and errors.

Signed-off-by: Roland Hieber <r.hieber@pengutronix.de>
---
v1 -> v2: no changes.
---
 drivers/hab/habv3.c |  4 ++++
 drivers/hab/habv4.c | 41 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 45 insertions(+)

diff --git a/drivers/hab/habv3.c b/drivers/hab/habv3.c
index 82ae245f8a..c190c78d40 100644
--- a/drivers/hab/habv3.c
+++ b/drivers/hab/habv3.c
@@ -78,5 +78,9 @@ int imx_habv3_get_status(uint32_t status)
 
 int imx25_hab_get_status(void)
 {
+	if (!cpu_is_mx25()) {
+		return 0;
+	}
 	return imx_habv3_get_status(readl(IOMEM(0x780018d4)));
 }
+postmmu_initcall(imx25_hab_get_status);
diff --git a/drivers/hab/habv4.c b/drivers/hab/habv4.c
index e33a8784a0..b7df7e4961 100644
--- a/drivers/hab/habv4.c
+++ b/drivers/hab/habv4.c
@@ -20,6 +20,7 @@
 
 #include <common.h>
 #include <hab.h>
+#include <init.h>
 #include <types.h>
 
 #include <mach/generic.h>
@@ -505,9 +506,49 @@ int imx6_hab_get_status(void)
 	return -EINVAL;
 }
 
+static int init_imx6_hab_get_status(void)
+{
+	int ret = 0;
+
+	if (!cpu_is_mx6())
+		/* can happen in multi-image builds and is not an error */
+		return 0;
+
+	ret = imx6_hab_get_status();
+
+	/* nobody will check the return value if there were HAB errors, but the
+	 * initcall will fail spectaculously with a strange error message. */
+	if (ret == -EPERM)
+		return 0;
+	return ret;
+}
+/* need to run before MMU setup because i.MX6 ROM code is mapped near 0x0,
+ * which will no longer be accessible when the MMU sets the zero page to
+ * faulting. */
+postconsole_initcall(init_imx6_hab_get_status);
+
 int imx28_hab_get_status(void)
 {
 	const struct habv4_rvt *rvt = (void *)HABV4_RVT_IMX28;
 
 	return habv4_get_status(rvt);
 }
+
+static int init_imx28_hab_get_status(void)
+{
+	int ret = 0;
+
+	if (!cpu_is_mx28())
+		/* can happen in multi-image builds and is not an error */
+		return 0;
+
+	ret = imx28_hab_get_status();
+
+	/* nobody will check the return value if there were HAB errors, but the
+	 * initcall will fail spectaculously with a strange error message. */
+	if (ret == -EPERM)
+		return 0;
+	return ret;
+}
+/* i.MX28 ROM code can be run after MMU setup to make use of caching */
+postmmu_initcall(init_imx28_hab_get_status);
-- 
2.19.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH v2 1/2] drivers: caam: add RNG software self-test

[Sascha Hauer]

Hi Roland,

Applied with several adjustments, see below.


On Thu, Nov 29, 2018 at 02:42:07PM +0100, Roland Hieber wrote:
> +/*
> + * construct_rng_self_test_jobdesc() - Implement destination address in RNG self test descriptors
> + * Returns zero on success, and negative on error.
> + */
> +static void construct_rng_self_test_jobdesc(u32 *desc, const u32 *rng_st_dsc, u8 *res_addr, int desc_size)
> +{
> +	int result_addr_idx = desc_size - 5;
> +	int i;
> +
> +	for (i = 0; i < desc_size; i++) {
> +		desc[i] = rng_st_dsc[i];
> +	}
> +
> +	/* Replace destination address in the descriptor */
> +	desc[result_addr_idx] = (u32)res_addr;
> +
> +	pr_vdebug("RNG SELF TEST DESCRIPTOR:\n");
> +	for (i = 0; i < desc_size; i++) {
> +		pr_vdebug("0x%08X\n", desc[i]);
> +	}

Dropped these for now. Should be a memory_display with suitable message
priority one we have it.

> +int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid, const u8 rngrev)
> +{
> +	int ret, i, desc_size = 0, result_size = 0, job_err = 0;
> +	const u32 *rng_st_dsc;
> +	const u8 *exp_result;
> +	u32 *desc;
> +	u8 *result;
> +
> +	pr_debug("got CAAM ERA %d, RNG Version ID %d, RNG revision %d\n", caam_era, rngvid, rngrev);
> +
> +	if (caam_era < 8 && rngvid == 4 && rngrev < 3) {
> +		// older affected i.MX chipsets have CAAM < 8 and have RNG4 < 4.3

No C++ comments please.

> +		rng_st_dsc = rng_dsc1;
> +		desc_size = ARRAY_SIZE(rng_dsc1);
> +		exp_result = rng_result1;
> +		result_size = ARRAY_SIZE(rng_result1);
> +	} else if (caam_era >= 8 || (rngvid >= 4 && rngrev >= 3)) {
> +		// newer affected chipsets have CAAM >= 8 or RNG4 >= 4.3
> +		rng_st_dsc = rng_dsc2;
> +		desc_size = ARRAY_SIZE(rng_dsc2);
> +		exp_result = rng_result2;
> +		result_size = ARRAY_SIZE(rng_result2);
> +	} else {
> +		pr_err("Error: Invalid CAAM ERA (%d) or RNG Version ID (%d) or RNG revision (%d)\n",
> +				caam_era, rngvid, rngrev);
> +		return -EINVAL;
> +	}
> +
> +	result = dma_alloc(sizeof(*result) * result_size);
> +	if (!result) {
> +		return -ENOMEM;
> +	}
> +
> +	desc = dma_alloc(sizeof(*desc) * desc_size);
> +	if (!desc) {
> +		return -ENOMEM;
> +	}

Loosing memory here.

> +
> +	construct_rng_self_test_jobdesc(desc, rng_st_dsc, result, desc_size);
> +
> +	dma_sync_single_for_device((unsigned long)desc,
> +			desc_size * sizeof(*desc), DMA_TO_DEVICE);
> +	dma_sync_single_for_device((unsigned long)result,
> +			result_size * sizeof(*result), DMA_FROM_DEVICE);
> +
> +	/* wait for job completion */
> +	ret = caam_jr_enqueue(dev, desc, rng_self_test_done, &job_err);
> +	if (ret) {
> +		pr_err("Error while running RNG self-test descriptor: %d %s\n",
> +				ret, strerror(ret));
> +		goto err;
> +	}
> +	if (job_err) {
> +		pr_err("Job Error:\n");
> +		caam_jr_strstatus(dev, job_err);
> +		goto err;

ret will be 0 here, not what we want.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH] fixup! drivers: caam: add RNG software self-test

[Roland Hieber]

Fix compile error and warning:

    drivers/crypto/caam/rng_self_test.c: In function 'caam_rng_self_test':
    drivers/crypto/caam/rng_self_test.c:167:2: error: expected ';' before '}' token
      }
      ^
    drivers/crypto/caam/rng_self_test.c:134:11: warning: unused variable 'i' [-Wunused-variable]
      int ret, i, desc_size = 0, result_size = 0, job_err = 0;
               ^

Signed-off-by: Roland Hieber <r.hieber@pengutronix.de>
---
Hi Sascha,

On Mon, Dec 03, 2018 at 09:01:56AM +0100, Sascha Hauer wrote:
> Hi Roland,
> 
> Applied with several adjustments, see below.

Right, thank you for those! I found two small problems left.

 - Roland
---
 drivers/crypto/caam/rng_self_test.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/crypto/caam/rng_self_test.c b/drivers/crypto/caam/rng_self_test.c
index 149b334f23..aab4fa2e47 100644
--- a/drivers/crypto/caam/rng_self_test.c
+++ b/drivers/crypto/caam/rng_self_test.c
@@ -131,7 +131,7 @@ static void rng_self_test_done(struct device_d *dev, u32 *desc, u32 err, void *a
  */
 int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid, const u8 rngrev)
 {
-	int ret, i, desc_size = 0, result_size = 0, job_err = 0;
+	int ret, desc_size = 0, result_size = 0, job_err = 0;
 	const u32 *rng_st_dsc;
 	const u8 *exp_result;
 	u32 *desc;
@@ -163,7 +163,7 @@ int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid,
 
 	if (!result || !desc) {
 		ret = -ENOMEM;
-		goto err
+		goto err;
 	}
 
 	construct_rng_self_test_jobdesc(desc, rng_st_dsc, result, desc_size);
-- 
2.19.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH v2 1/2] drivers: caam: add RNG software self-test

[Roland Hieber]

Hi Sascha,

On Thu, Nov 29, 2018 at 02:42:07PM +0100, Roland Hieber wrote:
> +/*
> + * caam_rng_self_test() - Perform RNG self test
> + * Returns zero on success, and negative on error.
> + */
> +int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid, const u8 rngrev)
> +{
> +	int ret, i, desc_size = 0, result_size = 0, job_err = 0;
> +	const u32 *rng_st_dsc;
> +	const u8 *exp_result;
> +	u32 *desc;
> +	u8 *result;
> +
> +	pr_debug("got CAAM ERA %d, RNG Version ID %d, RNG revision %d\n", caam_era, rngvid, rngrev);
> +
> +	if (caam_era < 8 && rngvid == 4 && rngrev < 3) {
> +		// older affected i.MX chipsets have CAAM < 8 and have RNG4 < 4.3
> +		rng_st_dsc = rng_dsc1;
> +		desc_size = ARRAY_SIZE(rng_dsc1);
> +		exp_result = rng_result1;
> +		result_size = ARRAY_SIZE(rng_result1);
> +	} else if (caam_era >= 8 || (rngvid >= 4 && rngrev >= 3)) {
> +		// newer affected chipsets have CAAM >= 8 or RNG4 >= 4.3
> +		rng_st_dsc = rng_dsc2;
> +		desc_size = ARRAY_SIZE(rng_dsc2);
> +		exp_result = rng_result2;
> +		result_size = ARRAY_SIZE(rng_result2);
> +	} else {
> +		pr_err("Error: Invalid CAAM ERA (%d) or RNG Version ID (%d) or RNG revision (%d)\n",
> +				caam_era, rngvid, rngrev);
> +		return -EINVAL;
> +	}
> +
> +	result = dma_alloc(sizeof(*result) * result_size);
> +	if (!result) {
> +		return -ENOMEM;
> +	}
> +
> +	desc = dma_alloc(sizeof(*desc) * desc_size);
> +	if (!desc) {
> +		return -ENOMEM;
> +	}
> +
> +	construct_rng_self_test_jobdesc(desc, rng_st_dsc, result, desc_size);
> +
> +	dma_sync_single_for_device((unsigned long)desc,
> +			desc_size * sizeof(*desc), DMA_TO_DEVICE);
> +	dma_sync_single_for_device((unsigned long)result,
> +			result_size * sizeof(*result), DMA_FROM_DEVICE);
> +
> +	/* wait for job completion */
> +	ret = caam_jr_enqueue(dev, desc, rng_self_test_done, &job_err);
> +	if (ret) {
> +		pr_err("Error while running RNG self-test descriptor: %d %s\n",
> +				ret, strerror(ret));
> +		goto err;
> +	}
> +	if (job_err) {
> +		pr_err("Job Error:\n");
> +		caam_jr_strstatus(dev, job_err);
> +		goto err;
> +	}
> +
> +	dma_sync_single_for_cpu((unsigned long)result, result_size * sizeof(*result),
> +			DMA_FROM_DEVICE);
> +
> +	pr_vdebug("Result\n");
> +	for (i = 0; i < result_size; i++) {
> +		pr_vdebug("%02X\n", result[i]);
> +	}
> +
> +	pr_vdebug("Expected Result:\n");
> +	for (i = 0; i < result_size; i++) {
> +		pr_vdebug("%02X\n", exp_result[i]);
> +	}
> +
> +	if (memcmp(result, exp_result, sizeof(*result) * result_size) != 0) {
> +		pr_err("RNG self-test failed with unexpected result\n");
> +		ret = -ERANGE;
> +		goto err;
> +	}
> +
> +	pr_notice("RNG software self-test passed\n");

I noticed that you downgraded this to a pr_info() when applying the
patch, is there a reason?  pr_warn()s have the "WARNING:" prefix,
pr_notice()s have a "NOTICE:" prefix, but pr_info()s don't have any
prefix. I think this one should be a pr_notice because it overrides the
pr_warn("RNG self-test failure detected...") from HAB and tells the user
that this warning is no longer relevant. With pr_info(), I think it will
get lost between all the other infos.

 - Roland

> +	ret = 0;
> +
> +err:
> +	dma_free(desc);
> +	dma_free(result);
> +	return ret;
> +}
> diff --git a/drivers/crypto/caam/rng_self_test.h b/drivers/crypto/caam/rng_self_test.h
[...]

-- 
Roland Hieber                     | r.hieber@pengutronix.de     |
Pengutronix e.K.                  | https://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim | Phone: +49-5121-206917-5086 |
Amtsgericht Hildesheim, HRA 2686  | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH] fixup! drivers: caam: add RNG software self-test

[Sascha Hauer]

On Mon, Dec 03, 2018 at 11:15:32AM +0100, Roland Hieber wrote:
> Fix compile error and warning:
> 
>     drivers/crypto/caam/rng_self_test.c: In function 'caam_rng_self_test':
>     drivers/crypto/caam/rng_self_test.c:167:2: error: expected ';' before '}' token
>       }
>       ^
>     drivers/crypto/caam/rng_self_test.c:134:11: warning: unused variable 'i' [-Wunused-variable]
>       int ret, i, desc_size = 0, result_size = 0, job_err = 0;
>                ^
> 
> Signed-off-by: Roland Hieber <r.hieber@pengutronix.de>
> ---
> Hi Sascha,
> 
> On Mon, Dec 03, 2018 at 09:01:56AM +0100, Sascha Hauer wrote:
> > Hi Roland,
> > 
> > Applied with several adjustments, see below.
> 
> Right, thank you for those! I found two small problems left.

Applied, thanks

Sascha

> 
>  - Roland
> ---
>  drivers/crypto/caam/rng_self_test.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/crypto/caam/rng_self_test.c b/drivers/crypto/caam/rng_self_test.c
> index 149b334f23..aab4fa2e47 100644
> --- a/drivers/crypto/caam/rng_self_test.c
> +++ b/drivers/crypto/caam/rng_self_test.c
> @@ -131,7 +131,7 @@ static void rng_self_test_done(struct device_d *dev, u32 *desc, u32 err, void *a
>   */
>  int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid, const u8 rngrev)
>  {
> -	int ret, i, desc_size = 0, result_size = 0, job_err = 0;
> +	int ret, desc_size = 0, result_size = 0, job_err = 0;
>  	const u32 *rng_st_dsc;
>  	const u8 *exp_result;
>  	u32 *desc;
> @@ -163,7 +163,7 @@ int caam_rng_self_test(struct device_d *dev, const u8 caam_era, const u8 rngvid,
>  
>  	if (!result || !desc) {
>  		ret = -ENOMEM;
> -		goto err
> +		goto err;
>  	}
>  
>  	construct_rng_self_test_jobdesc(desc, rng_st_dsc, result, desc_size);
> -- 
> 2.19.1
> 
> 
> _______________________________________________
> barebox mailing list
> barebox@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/barebox
> 

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH v2 1/2] drivers: caam: add RNG software self-test

[Sascha Hauer]

On Mon, Dec 03, 2018 at 11:52:26AM +0100, Roland Hieber wrote:
> Hi Sascha,
> 
> > +	pr_notice("RNG software self-test passed\n");
> 
> I noticed that you downgraded this to a pr_info() when applying the
> patch, is there a reason?  pr_warn()s have the "WARNING:" prefix,
> pr_notice()s have a "NOTICE:" prefix, but pr_info()s don't have any
> prefix. I think this one should be a pr_notice because it overrides the
> pr_warn("RNG self-test failure detected...") from HAB and tells the user
> that this warning is no longer relevant. With pr_info(), I think it will
> get lost between all the other infos.

I can follow this argumentation, but then I argue we should lower the
priority of the failure detected message to to pr_info or even pr_debug,
because the failure is pretty much expected, it is correctly being
worked around and there's nothing we can do about the failure.

People tend to bother about warnings and ask questions what is wrong.
When everything you can do is to tell that everything is fine then probably
the warning shouldn't be there in the first place.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH] fixup! drivers: caam: add RNG software self-test

[Roland Hieber]

Demote this warning to an info because the failure is expected and we
can work around it. The self-test will still error afterwards in case of
failure.

Signed-off-by: Roland Hieber <r.hieber@pengutronix.de>
---
On Tue, Dec 04, 2018 at 08:49:58AM +0100, Sascha Hauer wrote:
> On Mon, Dec 03, 2018 at 11:52:26AM +0100, Roland Hieber wrote:
> > Hi Sascha,
> > 
> > > + pr_notice("RNG software self-test passed\n");
> > 
> > I noticed that you downgraded this to a pr_info() when applying the
> > patch, is there a reason?  pr_warn()s have the "WARNING:" prefix,
> > pr_notice()s have a "NOTICE:" prefix, but pr_info()s don't have any
> > prefix. I think this one should be a pr_notice because it overrides the
> > pr_warn("RNG self-test failure detected...") from HAB and tells the user
> > that this warning is no longer relevant. With pr_info(), I think it will
> > get lost between all the other infos.
> 
> I can follow this argumentation, but then I argue we should lower the
> priority of the failure detected message to to pr_info or even pr_debug,
> because the failure is pretty much expected, it is correctly being
> worked around and there's nothing we can do about the failure.

Right, that argumentation seems more reasonable to me.

 - Roland

---
 drivers/hab/habv4.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/hab/habv4.c b/drivers/hab/habv4.c
index bab458ae32..cca46c1490 100644
--- a/drivers/hab/habv4.c
+++ b/drivers/hab/habv4.c
@@ -455,7 +455,7 @@ static int habv4_get_status(const struct habv4_rvt *rvt)
 		}
 
 		if (is_rng_fail_event) {
-			pr_warning("RNG self-test failure detected, will run software self-test\n");
+			pr_info("RNG self-test failure detected, will run software self-test\n");
 			habv4_need_rng_software_self_test = true;
 		} else {
 			pr_err("-------- HAB warning Event %d --------\n", index);
-- 
2.19.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox