All of lore.kernel.org
 help / color / mirror / Atom feed
From: ultracoolguy@tutanota.com
To: Pavel <pavel@ucw.cz>, Dmurphy <dmurphy@ti.com>
Cc: Linux Leds <linux-leds@vger.kernel.org>,
	Trivial <trivial@kernel.org>,
	linux-kernel@vger.kernel.org
Subject: [PATCH] leds: lm3697: Fix out-of-bound access
Date: Sat, 3 Oct 2020 15:02:51 +0200 (CEST)	[thread overview]
Message-ID: <MIiYgay--3-2@tutanota.com> (raw)

[-- Attachment #1: Type: text/plain, Size: 3495 bytes --]

Signed-off-by: Ultracoolguy <ultracoolguy@tutanota.com>
Hi, all. This is a patch fixing an out-of-bounds error due to lm3697_init expecting the device tree to use both control banks.  This fixes it by adding a new variable that will hold the number of used banks.

Panic caused by this bug:

<1>[    3.059763] Unable to handle kernel NULL pointer dereference at virtual address 00000000000001e5
<1>[    3.059779] Mem abort info:
<1>[    3.059788]   ESR = 0x96000004
<1>[    3.059799]   EC = 0x25: DABT (current EL), IL = 32 bits
<1>[    3.059807]   SET = 0, FnV = 0
<1>[    3.059816]   EA = 0, S1PTW = 0
<1>[    3.059824] Data abort info:
<1>[    3.059833]   ISV = 0, ISS = 0x00000004
<1>[    3.059841]   CM = 0, WnR = 0
<1>[    3.059850] [00000000000001e5] user address but active_mm is swapper
<0>[    3.059864] Internal error: Oops: 96000004 [#1] PREEMPT SMP
<7>[    3.059874] Modules linked in:
<7>[    3.059893] CPU: 3 PID: 1 Comm: swapper/0 Tainted: G        W         5.9.0-rc7-postmarketos-qcom-msm8953 #71
<7>[    3.059903] Hardware name: Motorola G7 Power (ocean) (DT)
<7>[    3.059916] pstate: a0000005 (NzCv daif -PAN -UAO BTYPE=--)
<7>[    3.059937] pc : regmap_write+0x1c/0x78
<7>[    3.059952] lr : ti_lmu_common_set_ramp+0x60/0x70
<7>[    3.059961] sp : ffff800010043ad0
<7>[    3.059970] x29: ffff800010043ad0 x28: ffff0000b98510d0 
 <7>[    3.059982] x27: ffff0000b9851288 x26: ffff800010906cc0 
<7>[    3.059995] x25: ffff8000108e5d85 x24: 0000000000000001 
<7>[    3.060008] x23: 0000000000000000 x22: ffff0000b9851080 
<7>[    3.060020] x21: 0000000000000001 x20: 000000000000000f 
<7>[    3.060032] x19: 0000000000000001 x18: 0000000000000000 
<7>[    3.060045] x17: 00000000f2f3902a x16: 00000000ee02cbfb 
<7>[    3.060058] x15: 000000000000000a x14: 0000000000000307 
<7>[    3.060070] x13: ffffffffffffffff x12: ffffffffffffffff 
<7>[    3.060083] x11: 0000000000000000 x10: 0000000000000950 
<7>[    3.060095] x9 : ffff8000105864e4 x8 : ffff0000b2f089b0 
<7>[    3.060108] x7 : 0000000000000004 x6 : 000000000000033c 
<7>[    3.060120] x5 : ffff0000b98c51d0 x4 : 0000000000000000 
<7>[    3.060132] x3 : ffff0000b2f08000 x2 : 00000000000000ff 
<7>[    3.060145] x1 : 0000000000000000 x0 : 0000000000000001
<7>[    3.060158] Call trace:
<7>[    3.060172]  regmap_write+0x1c/0x78
<7>[    3.060183]  ti_lmu_common_set_ramp+0x60/0x70
<7>[    3.060195]  lm3697_probe+0x4d4/0x534
<7>[    3.060211]  i2c_device_probe+0x22c/0x294
<7>[    3.060225]  really_probe+0x1e0/0x468
<7>[    3.060237]  driver_probe_device+0xfc/0x140
<7>[    3.060250]  device_driver_attach+0x48/0x70
<7>[    3.060262]  __driver_attach+0x134/0x14c
<7>[    3.060275]  bus_for_each_dev+0x64/0xbc
<7>[    3.060287]  driver_attach+0x28/0x30
<7>[    3.060299]  bus_add_driver+0x110/0x1fc
<7>[    3.060312]  driver_register+0xa8/0xec
<7>[    3.060325]  i2c_register_driver+0x94/0xac
<7>[    3.060341]  lm3697_driver_init+0x20/0x28
<7>[    3.060356]  do_one_initcall+0xc4/0x228
<7>[    3.060368]  kernel_init_freeable+0x1e4/0x24c
<7>[    3.060384]  kernel_init+0x18/0x110
<7>[    3.060397]  ret_from_fork+0x10/0x18
<0>[    3.060415] Code: 910003fd a90153f3 aa0003f3 f90013f5 (b941e400) 
<4>[    3.060439] ---[ end trace fcc24bd799273148 ]---


[-- Attachment #2: 0001-leds-lm3697-Fix-out-of-bound-access.patch --]
[-- Type: text/x-patch, Size: 2515 bytes --]

From 0dfd5ab647ccbc585c543d702b44d20f0e3fe436 Mon Sep 17 00:00:00 2001
From: Ultracoolguy <ultracoolguy@tutanota.com>
Date: Fri, 2 Oct 2020 18:27:00 -0400
Subject: [PATCH] leds:lm3697:Fix out-of-bound access

If both led banks aren't used in device tree,
an out-of-bounds condition in lm3697_init occurs
because of the for loop assuming that all the banks are used.
Fix it by adding a variable that contains the number of used banks.

Signed-off-by: Ultracoolguy <ultracoolguy@tutanota.com>
---
 drivers/leds/leds-lm3697.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/drivers/leds/leds-lm3697.c b/drivers/leds/leds-lm3697.c
index 024983088d59..a4ec2b6077e6 100644
--- a/drivers/leds/leds-lm3697.c
+++ b/drivers/leds/leds-lm3697.c
@@ -56,7 +56,7 @@ struct lm3697_led {
 	struct ti_lmu_bank lmu_data;
 	int control_bank;
 	int enabled;
-	int num_leds;
+	int num_led_strings;
 };

 /**
@@ -78,6 +78,7 @@ struct lm3697 {
 	struct mutex lock;

 	int bank_cfg;
+	int num_leds;

 	struct lm3697_led leds[];
 };
@@ -180,7 +181,7 @@ static int lm3697_init(struct lm3697 *priv)
 	if (ret)
 		dev_err(&priv->client->dev, "Cannot write OUTPUT config\n");

-	for (i = 0; i < LM3697_MAX_CONTROL_BANKS; i++) {
+	for (i = 0; i < priv->num_leds; i++) {
 		led = &priv->leds[i];
 		ret = ti_lmu_common_set_ramp(&led->lmu_data);
 		if (ret)
@@ -244,22 +245,22 @@ static int lm3697_probe_dt(struct lm3697 *priv)
 		led->lmu_data.lsb_brightness_reg = LM3697_CTRL_A_BRT_LSB +
 						   led->control_bank * 2;

-		led->num_leds = fwnode_property_count_u32(child, "led-sources");
-		if (led->num_leds > LM3697_MAX_LED_STRINGS) {
+		led->num_led_strings = fwnode_property_count_u32(child, "led-sources");
+		if (led->num_led_strings > LM3697_MAX_LED_STRINGS) {
 			dev_err(&priv->client->dev, "Too many LED strings defined\n");
 			continue;
 		}

 		ret = fwnode_property_read_u32_array(child, "led-sources",
 						    led->hvled_strings,
-						    led->num_leds);
+						    led->num_led_strings);
 		if (ret) {
 			dev_err(&priv->client->dev, "led-sources property missing\n");
 			fwnode_handle_put(child);
 			goto child_out;
 		}

-		for (j = 0; j < led->num_leds; j++)
+		for (j = 0; j < led->num_led_strings; j++)
 			priv->bank_cfg |=
 				(led->control_bank << led->hvled_strings[j]);

@@ -317,6 +318,8 @@ static int lm3697_probe(struct i2c_client *client,
 	if (!led)
 		return -ENOMEM;

+	led->num_leds = count;
+
 	mutex_init(&led->lock);
 	i2c_set_clientdata(client, led);

--
2.28.0


             reply	other threads:[~2020-10-03 13:02 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-03 13:02 ultracoolguy [this message]
2020-10-03 13:56 ` [PATCH] leds: lm3697: Fix out-of-bound access Pavel Machek
2020-10-03 14:43   ` ultracoolguy
2020-10-05 12:13 ` Marek Behun
2020-10-05 13:50   ` Pavel Machek
2020-10-05 13:57   ` ultracoolguy
2020-10-05 14:33     ` Dan Murphy
2020-10-05 14:37       ` Dan Murphy
2020-10-05 14:38       ` ultracoolguy
2020-10-05 14:41         ` Dan Murphy
2020-10-05 15:35           ` ultracoolguy
2020-10-05 16:05             ` Pavel Machek
2020-10-05 16:48             ` Alexander Dahl
2020-10-05 17:14               ` ultracoolguy
2020-10-05 17:32                 ` Pavel Machek
2020-10-05 18:29                   ` ultracoolguy
2020-10-05 18:31                     ` ultracoolguy
2020-10-05 18:39                       ` Pavel Machek
2020-10-05 18:48                         ` ultracoolguy
2020-10-06  7:33                   ` Marek Behun
2020-10-06 11:59                     ` ultracoolguy
2020-10-06 12:21                       ` Dan Murphy
2020-10-06 14:41                         ` Marek Behun
2020-10-06 14:57                           ` Dan Murphy
2020-10-06 15:14                             ` Marek Behun
2020-10-06 17:26                           ` Pavel Machek
2020-10-05 15:59       ` Pavel Machek

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=MIiYgay--3-2@tutanota.com \
    --to=ultracoolguy@tutanota.com \
    --cc=dmurphy@ti.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-leds@vger.kernel.org \
    --cc=pavel@ucw.cz \
    --cc=trivial@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.