[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[Vincent Legoll]

[-- Attachment #1: Type: text/plain, Size: 79 bytes --]

I checked that the CVE patches have been applied upstream.

-- 
Vincent Legoll

[-- Attachment #2: 0001-gnu-bluez-Update-to-5.54.patch --]
[-- Type: text/x-patch, Size: 8915 bytes --]

From a84f040fd88b02b556a0fbd207f8edce1a940924 Mon Sep 17 00:00:00 2001
From: Vincent Legoll <vincent.legoll@gmail.com>
Date: Sun, 15 Mar 2020 19:07:57 +0100
Subject: [PATCH] gnu: bluez: Update to 5.54.

* gnu/packages/linux.scm (bluez) : Update to 5.54.
(bluez/fixed): remove variable.
* gnu/packages/patches/bluez-CVE-2020-0556.patch: Remove file
---
 gnu/packages/linux.scm                        |  13 +-
 .../patches/bluez-CVE-2020-0556.patch         | 180 ------------------
 2 files changed, 2 insertions(+), 191 deletions(-)
 delete mode 100644 gnu/packages/patches/bluez-CVE-2020-0556.patch

diff --git a/gnu/packages/linux.scm b/gnu/packages/linux.scm
index 688d9eefaf..78c6d48474 100644
--- a/gnu/packages/linux.scm
+++ b/gnu/packages/linux.scm
@@ -4009,8 +4009,7 @@ Bluetooth audio output devices like headphones or loudspeakers.")
 (define-public bluez
   (package
     (name "bluez")
-    (replacement bluez/fixed)
-    (version "5.53")
+    (version "5.54")
     (source (origin
               (method url-fetch)
               (uri (string-append
@@ -4018,7 +4017,7 @@ Bluetooth audio output devices like headphones or loudspeakers.")
                     version ".tar.xz"))
               (sha256
                (base32
-                "1g1qg6dz6hl3csrmz75ixr12lwv836hq3ckb259svvrg62l2vaiq"))))
+                "1p2ncvjz6alr9n3l5wvq2arqgc7xjs6dqyar1l9jp0z8cfgapkb8"))))
     (build-system gnu-build-system)
     (arguments
      `(#:configure-flags
@@ -4075,14 +4074,6 @@ Bluetooth audio output devices like headphones or loudspeakers.")
 is flexible, efficient and uses a modular implementation.")
     (license license:gpl2+)))
 
-(define bluez/fixed
-  (package
-    (inherit bluez)
-    (source (origin
-              (inherit (package-source bluez))
-              (patches (append (origin-patches (package-source bluez))
-                               (search-patches "bluez-CVE-2020-0556.patch")))))))
-
 (define-public fuse-exfat
   (package
     (name "fuse-exfat")
diff --git a/gnu/packages/patches/bluez-CVE-2020-0556.patch b/gnu/packages/patches/bluez-CVE-2020-0556.patch
deleted file mode 100644
index 7c34459a3a..0000000000
--- a/gnu/packages/patches/bluez-CVE-2020-0556.patch
+++ /dev/null
@@ -1,180 +0,0 @@
-Fix CVE-2020-0556:
-
-https://lore.kernel.org/linux-bluetooth/20200310023516.209146-1-alainm@chromium.org/
-https://www.intel.com/content/www/us/en/security-center/advisory/intel-sa-00352.html
-http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2020-0556
-
-Patches copied from upstream source repository:
-
-https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=3cccdbab2324086588df4ccf5f892fb3ce1f1787
-https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=8cdbd3b09f29da29374e2f83369df24228da0ad1
-
-From 3cccdbab2324086588df4ccf5f892fb3ce1f1787 Mon Sep 17 00:00:00 2001
-From: Alain Michaud <alainm@chromium.org>
-Date: Tue, 10 Mar 2020 02:35:18 +0000
-Subject: [PATCH] HID accepts bonded device connections only.
-
-This change adds a configuration for platforms to choose a more secure
-posture for the HID profile.  While some older mice are known to not
-support pairing or encryption, some platform may choose a more secure
-posture by requiring the device to be bonded  and require the
-connection to be encrypted when bonding is required.
-
-Reference:
-https://www.intel.com/content/www/us/en/security-center/advisory/intel-sa-00352.html
----
- profiles/input/device.c   | 23 ++++++++++++++++++++++-
- profiles/input/device.h   |  1 +
- profiles/input/input.conf |  8 ++++++++
- profiles/input/manager.c  | 13 ++++++++++++-
- 4 files changed, 43 insertions(+), 2 deletions(-)
-
-diff --git a/profiles/input/device.c b/profiles/input/device.c
-index 2cb3811c8..d89da2d7c 100644
---- a/profiles/input/device.c
-+++ b/profiles/input/device.c
-@@ -92,6 +92,7 @@ struct input_device {
- 
- static int idle_timeout = 0;
- static bool uhid_enabled = false;
-+static bool classic_bonded_only = false;
- 
- void input_set_idle_timeout(int timeout)
- {
-@@ -103,6 +104,11 @@ void input_enable_userspace_hid(bool state)
- 	uhid_enabled = state;
- }
- 
-+void input_set_classic_bonded_only(bool state)
-+{
-+	classic_bonded_only = state;
-+}
-+
- static void input_device_enter_reconnect_mode(struct input_device *idev);
- static int connection_disconnect(struct input_device *idev, uint32_t flags);
- 
-@@ -970,8 +976,18 @@ static int hidp_add_connection(struct input_device *idev)
- 	if (device_name_known(idev->device))
- 		device_get_name(idev->device, req->name, sizeof(req->name));
- 
-+	/* Make sure the device is bonded if required */
-+	if (classic_bonded_only && !device_is_bonded(idev->device,
-+				btd_device_get_bdaddr_type(idev->device))) {
-+		error("Rejected connection from !bonded device %s", dst_addr);
-+		goto cleanup;
-+	}
-+
- 	/* Encryption is mandatory for keyboards */
--	if (req->subclass & 0x40) {
-+	/* Some platforms may choose to require encryption for all devices */
-+	/* Note that this only matters for pre 2.1 devices as otherwise the */
-+	/* device is encrypted by default by the lower layers */
-+	if (classic_bonded_only || req->subclass & 0x40) {
- 		if (!bt_io_set(idev->intr_io, &gerr,
- 					BT_IO_OPT_SEC_LEVEL, BT_IO_SEC_MEDIUM,
- 					BT_IO_OPT_INVALID)) {
-@@ -1203,6 +1219,11 @@ static void input_device_enter_reconnect_mode(struct input_device *idev)
- 	DBG("path=%s reconnect_mode=%s", idev->path,
- 				reconnect_mode_to_string(idev->reconnect_mode));
- 
-+	/* Make sure the device is bonded if required */
-+	if (classic_bonded_only && !device_is_bonded(idev->device,
-+				btd_device_get_bdaddr_type(idev->device)))
-+		return;
-+
- 	/* Only attempt an auto-reconnect when the device is required to
- 	 * accept reconnections from the host.
- 	 */
-diff --git a/profiles/input/device.h b/profiles/input/device.h
-index 51a9aee18..3044db673 100644
---- a/profiles/input/device.h
-+++ b/profiles/input/device.h
-@@ -29,6 +29,7 @@ struct input_conn;
- 
- void input_set_idle_timeout(int timeout);
- void input_enable_userspace_hid(bool state);
-+void input_set_classic_bonded_only(bool state);
- 
- int input_device_register(struct btd_service *service);
- void input_device_unregister(struct btd_service *service);
-diff --git a/profiles/input/input.conf b/profiles/input/input.conf
-index 3e1d65aae..166aff4a4 100644
---- a/profiles/input/input.conf
-+++ b/profiles/input/input.conf
-@@ -11,3 +11,11 @@
- # Enable HID protocol handling in userspace input profile
- # Defaults to false (HIDP handled in HIDP kernel module)
- #UserspaceHID=true
-+
-+# Limit HID connections to bonded devices
-+# The HID Profile does not specify that devices must be bonded, however some
-+# platforms may want to make sure that input connections only come from bonded
-+# device connections. Several older mice have been known for not supporting
-+# pairing/encryption.
-+# Defaults to false to maximize device compatibility.
-+#ClassicBondedOnly=true
-diff --git a/profiles/input/manager.c b/profiles/input/manager.c
-index 1d31b0652..5cd27b839 100644
---- a/profiles/input/manager.c
-+++ b/profiles/input/manager.c
-@@ -96,7 +96,7 @@ static int input_init(void)
- 	config = load_config_file(CONFIGDIR "/input.conf");
- 	if (config) {
- 		int idle_timeout;
--		gboolean uhid_enabled;
-+		gboolean uhid_enabled, classic_bonded_only;
- 
- 		idle_timeout = g_key_file_get_integer(config, "General",
- 							"IdleTimeout", &err);
-@@ -114,6 +114,17 @@ static int input_init(void)
- 			input_enable_userspace_hid(uhid_enabled);
- 		} else
- 			g_clear_error(&err);
-+
-+		classic_bonded_only = g_key_file_get_boolean(config, "General",
-+						"ClassicBondedOnly", &err);
-+
-+		if (!err) {
-+			DBG("input.conf: ClassicBondedOnly=%s",
-+					classic_bonded_only ? "true" : "false");
-+			input_set_classic_bonded_only(classic_bonded_only);
-+		} else
-+			g_clear_error(&err);
-+
- 	}
- 
- 	btd_profile_register(&input_profile);
--- 
-2.25.1
-
-From 8cdbd3b09f29da29374e2f83369df24228da0ad1 Mon Sep 17 00:00:00 2001
-From: Alain Michaud <alainm@chromium.org>
-Date: Tue, 10 Mar 2020 02:35:16 +0000
-Subject: [PATCH] HOGP must only accept data from bonded devices.
-
-HOGP 1.0 Section 6.1 establishes that the HOGP must require bonding.
-
-Reference:
-https://www.intel.com/content/www/us/en/security-center/advisory/intel-sa-00352.htm
----
- profiles/input/hog.c | 4 ++++
- 1 file changed, 4 insertions(+)
-
-diff --git a/profiles/input/hog.c b/profiles/input/hog.c
-index 83c017dcb..dfac68921 100644
---- a/profiles/input/hog.c
-+++ b/profiles/input/hog.c
-@@ -186,6 +186,10 @@ static int hog_accept(struct btd_service *service)
- 			return -EINVAL;
- 	}
- 
-+	/* HOGP 1.0 Section 6.1 requires bonding */
-+	if (!device_is_bonded(device, btd_device_get_bdaddr_type(device)))
-+		return -ECONNREFUSED;
-+
- 	/* TODO: Replace GAttrib with bt_gatt_client */
- 	bt_hog_attach(dev->hog, attrib);
- 
--- 
-2.25.1
-
-- 
2.25.1

[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[Vincent Legoll]

But I have not tested the new version,
I'm in a VM with no bluetooth HW...

-- 
Vincent Legoll

[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[Leo Famulari]

On Sun, Mar 15, 2020 at 07:13:17PM +0100, Vincent Legoll wrote:
> But I have not tested the new version,
> I'm in a VM with no bluetooth HW...

That's okay, I think we can expect it to "just work".

bug#40081: [PATCH] gnu: bluez: Update to 5.54.

[Leo Famulari]

On Sun, Mar 15, 2020 at 07:11:35PM +0100, Vincent Legoll wrote:
> From a84f040fd88b02b556a0fbd207f8edce1a940924 Mon Sep 17 00:00:00 2001
> From: Vincent Legoll <vincent.legoll@gmail.com>
> Date: Sun, 15 Mar 2020 19:07:57 +0100
> Subject: [PATCH] gnu: bluez: Update to 5.54.
> 
> * gnu/packages/linux.scm (bluez) : Update to 5.54.
> (bluez/fixed): remove variable.
> * gnu/packages/patches/bluez-CVE-2020-0556.patch: Remove file

Thanks!

I also removed the patch from 'gnu/local.mk' and made some small edits
to the commit message.

Pushed as f24aaa81de8c709adfda2e89271c562a5ca8d959

[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[Marius Bakke]

[-- Attachment #1: Type: text/plain, Size: 1023 bytes --]

Leo Famulari <leo@famulari.name> writes:

> On Sun, Mar 15, 2020 at 07:11:35PM +0100, Vincent Legoll wrote:
>> From a84f040fd88b02b556a0fbd207f8edce1a940924 Mon Sep 17 00:00:00 2001
>> From: Vincent Legoll <vincent.legoll@gmail.com>
>> Date: Sun, 15 Mar 2020 19:07:57 +0100
>> Subject: [PATCH] gnu: bluez: Update to 5.54.
>> 
>> * gnu/packages/linux.scm (bluez) : Update to 5.54.
>> (bluez/fixed): remove variable.
>> * gnu/packages/patches/bluez-CVE-2020-0556.patch: Remove file
>
> Thanks!
>
> I also removed the patch from 'gnu/local.mk' and made some small edits
> to the commit message.
>
> Pushed as f24aaa81de8c709adfda2e89271c562a5ca8d959

BlueZ causes 1000+ rebuilds according to 'guix refresh -l bluez'.
Berlin already failed to build pulseaudio on armhf, leaving ARM users
without substitutes until someone is able to restart all the builds:

https://ci.guix.gnu.org/eval/11627?status=failed
https://ci.guix.gnu.org/build/2390393/details

Should we revert it, or turn it into a graft?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[Leo Famulari]

On Mon, Mar 16, 2020 at 12:45:36AM +0100, Marius Bakke wrote:
> BlueZ causes 1000+ rebuilds according to 'guix refresh -l bluez'.
> Berlin already failed to build pulseaudio on armhf, leaving ARM users
> without substitutes until someone is able to restart all the builds:
> 
> https://ci.guix.gnu.org/eval/11627?status=failed
> https://ci.guix.gnu.org/build/2390393/details
> 
> Should we revert it, or turn it into a graft?

Oops! Reverted...

[bug#40081] [PATCH] gnu: bluez: Update to 5.54.

[Leo Famulari]

On Mon, Mar 16, 2020 at 12:45:36AM +0100, Marius Bakke wrote:
> BlueZ causes 1000+ rebuilds according to 'guix refresh -l bluez'.
> Berlin already failed to build pulseaudio on armhf, leaving ARM users
> without substitutes until someone is able to restart all the builds:
> 
> https://ci.guix.gnu.org/eval/11627?status=failed
> https://ci.guix.gnu.org/build/2390393/details
> 
> Should we revert it, or turn it into a graft?

I pushed the original change to a new 'staging' branch on Savannah.