[PATCH v1 0/3] Enable user helper interface for efi capsule update

[PATCH v1 0/3] Enable user helper interface for efi capsule update

[Kweh Hock Leong]

From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>

Hi Guys,

This patchset is created on top of "efi: Capsule update support" patch:
http://permalink.gmane.org/gmane.linux.kernel.efi/4837

It leverages the request_firmware_nowait() to expose the user helper
interface for user to upload the capsule binary and calling the
efi_capsule_update() API to pass the binary to EFI firmware.

Besides build in kernel, the design also cater for build as kernel driver
module. This patchset introduce a new API (request_firmware_abort()) at
firmware_class so that the driver module could be unloaded by calling
the API to properly stop user helper interface and release the device.

Thanks.


Kweh, Hock Leong (3):
  firmware loader: Introduce new API - request_firmware_abort()
  firmware loader: fix hung task warning dump
  efi: Capsule update with user helper interface

 drivers/base/firmware_class.c                  |   34 +++-
 drivers/firmware/efi/Kconfig                   |   13 ++
 drivers/firmware/efi/Makefile                  |    1 +
 drivers/firmware/efi/efi-capsule-user-helper.c |  246 ++++++++++++++++++++++++
 include/linux/firmware.h                       |    4 +
 5 files changed, 296 insertions(+), 2 deletions(-)
 create mode 100644 drivers/firmware/efi/efi-capsule-user-helper.c

-- 
1.7.9.5

[PATCH v1 0/3] Enable user helper interface for efi capsule update

[Kweh Hock Leong]

From: "Kweh, Hock Leong" <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

Hi Guys,

This patchset is created on top of "efi: Capsule update support" patch:
http://permalink.gmane.org/gmane.linux.kernel.efi/4837

It leverages the request_firmware_nowait() to expose the user helper
interface for user to upload the capsule binary and calling the
efi_capsule_update() API to pass the binary to EFI firmware.

Besides build in kernel, the design also cater for build as kernel driver
module. This patchset introduce a new API (request_firmware_abort()) at
firmware_class so that the driver module could be unloaded by calling
the API to properly stop user helper interface and release the device.

Thanks.


Kweh, Hock Leong (3):
  firmware loader: Introduce new API - request_firmware_abort()
  firmware loader: fix hung task warning dump
  efi: Capsule update with user helper interface

 drivers/base/firmware_class.c                  |   34 +++-
 drivers/firmware/efi/Kconfig                   |   13 ++
 drivers/firmware/efi/Makefile                  |    1 +
 drivers/firmware/efi/efi-capsule-user-helper.c |  246 ++++++++++++++++++++++++
 include/linux/firmware.h                       |    4 +
 5 files changed, 296 insertions(+), 2 deletions(-)
 create mode 100644 drivers/firmware/efi/efi-capsule-user-helper.c

-- 
1.7.9.5

[PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Kweh Hock Leong]

From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>

Besides aborting through user helper interface, a new API
request_firmware_abort() allows kernel driver module to abort the
request_firmware() / request_firmware_nowait() when they are no
longer needed. It is useful for cancelling an outstanding firmware
load if initiated from inside a module where that module is in the
process of being unloaded, since the unload function may not have
a handle to the struct firmware_buf.

Note for people who use request_firmware_nowait():
You are required to do your own synchronization after you call
request_firmware_abort() in order to continue unloading the
module. The example synchronization code shows below:

while (THIS_MODULE && module_refcount(THIS_MODULE))
	msleep(20);

Cc: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Kweh, Hock Leong <hock.leong.kweh@intel.com>
---
 drivers/base/firmware_class.c |   30 ++++++++++++++++++++++++++++++
 include/linux/firmware.h      |    4 ++++
 2 files changed, 34 insertions(+)

diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
index d276e33..b2fe40b 100644
--- a/drivers/base/firmware_class.c
+++ b/drivers/base/firmware_class.c
@@ -1291,6 +1291,36 @@ request_firmware_nowait(
 }
 EXPORT_SYMBOL(request_firmware_nowait);
 
+/**
+ * request_firmware_abort - abort the waiting of firmware request
+ * @fw_name: name of firmware file
+ *
+ *	@fw_name is the same name string while calling to request_firmware()
+ *		or request_firmware_nowait().
+ **/
+int request_firmware_abort(const char *fw_name)
+{
+	struct firmware_buf *buf;
+	struct firmware_buf *entry;
+
+	spin_lock(&fw_cache.lock);
+	buf = __fw_lookup_buf(fw_name);
+	spin_unlock(&fw_cache.lock);
+	if (!buf)
+		return -ENOENT;
+
+	mutex_lock(&fw_lock);
+	list_for_each_entry(entry, &pending_fw_head, pending_list) {
+		if (entry == buf) {
+			__fw_load_abort(buf);
+			break;
+		}
+	}
+	mutex_unlock(&fw_lock);
+	return 0;
+}
+EXPORT_SYMBOL(request_firmware_abort);
+
 #ifdef CONFIG_PM_SLEEP
 static ASYNC_DOMAIN_EXCLUSIVE(fw_cache_domain);
 
diff --git a/include/linux/firmware.h b/include/linux/firmware.h
index 5952933..ed90c8b 100644
--- a/include/linux/firmware.h
+++ b/include/linux/firmware.h
@@ -47,6 +47,7 @@ int request_firmware_nowait(
 	void (*cont)(const struct firmware *fw, void *context));
 
 void release_firmware(const struct firmware *fw);
+int request_firmware_abort(const char *fw_name);
 #else
 static inline int request_firmware(const struct firmware **fw,
 				   const char *name,
@@ -66,6 +67,9 @@ static inline void release_firmware(const struct firmware *fw)
 {
 }
 
+static inline int request_firmware_abort(const char *fw_name)
+{
+}
 #endif
 
 #ifdef CONFIG_FW_LOADER_USER_HELPER
-- 
1.7.9.5

[PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Kweh Hock Leong]

From: "Kweh, Hock Leong" <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

Besides aborting through user helper interface, a new API
request_firmware_abort() allows kernel driver module to abort the
request_firmware() / request_firmware_nowait() when they are no
longer needed. It is useful for cancelling an outstanding firmware
load if initiated from inside a module where that module is in the
process of being unloaded, since the unload function may not have
a handle to the struct firmware_buf.

Note for people who use request_firmware_nowait():
You are required to do your own synchronization after you call
request_firmware_abort() in order to continue unloading the
module. The example synchronization code shows below:

while (THIS_MODULE && module_refcount(THIS_MODULE))
	msleep(20);

Cc: Matt Fleming <matt.fleming-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Kweh, Hock Leong <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
 drivers/base/firmware_class.c |   30 ++++++++++++++++++++++++++++++
 include/linux/firmware.h      |    4 ++++
 2 files changed, 34 insertions(+)

diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
index d276e33..b2fe40b 100644
--- a/drivers/base/firmware_class.c
+++ b/drivers/base/firmware_class.c
@@ -1291,6 +1291,36 @@ request_firmware_nowait(
 }
 EXPORT_SYMBOL(request_firmware_nowait);
 
+/**
+ * request_firmware_abort - abort the waiting of firmware request
+ * @fw_name: name of firmware file
+ *
+ *	@fw_name is the same name string while calling to request_firmware()
+ *		or request_firmware_nowait().
+ **/
+int request_firmware_abort(const char *fw_name)
+{
+	struct firmware_buf *buf;
+	struct firmware_buf *entry;
+
+	spin_lock(&fw_cache.lock);
+	buf = __fw_lookup_buf(fw_name);
+	spin_unlock(&fw_cache.lock);
+	if (!buf)
+		return -ENOENT;
+
+	mutex_lock(&fw_lock);
+	list_for_each_entry(entry, &pending_fw_head, pending_list) {
+		if (entry == buf) {
+			__fw_load_abort(buf);
+			break;
+		}
+	}
+	mutex_unlock(&fw_lock);
+	return 0;
+}
+EXPORT_SYMBOL(request_firmware_abort);
+
 #ifdef CONFIG_PM_SLEEP
 static ASYNC_DOMAIN_EXCLUSIVE(fw_cache_domain);
 
diff --git a/include/linux/firmware.h b/include/linux/firmware.h
index 5952933..ed90c8b 100644
--- a/include/linux/firmware.h
+++ b/include/linux/firmware.h
@@ -47,6 +47,7 @@ int request_firmware_nowait(
 	void (*cont)(const struct firmware *fw, void *context));
 
 void release_firmware(const struct firmware *fw);
+int request_firmware_abort(const char *fw_name);
 #else
 static inline int request_firmware(const struct firmware **fw,
 				   const char *name,
@@ -66,6 +67,9 @@ static inline void release_firmware(const struct firmware *fw)
 {
 }
 
+static inline int request_firmware_abort(const char *fw_name)
+{
+}
 #endif
 
 #ifdef CONFIG_FW_LOADER_USER_HELPER
-- 
1.7.9.5

[PATCH v1 2/3] firmware loader: fix hung task warning dump

[Kweh Hock Leong]

From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>

When using request_firmware_nowait() with FW_ACTION_NOHOTPLUG param to
expose user helper interface, if the user do not react immediately, after
120 seconds there will be a hung task warning message dumped as below:

[ 3000.784235] INFO: task kworker/0:0:8259 blocked for more than 120 seconds.
[ 3000.791281]       Tainted: G            E 3.16.0-rc1-yocto-standard #41
[ 3000.798082] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 3000.806072] kworker/0:0     D cd0075c8     0  8259      2 0x00000000
[ 3000.812765] Workqueue: events request_firmware_work_func
[ 3000.818253]  cd375e18 00000046 0000000e cd0075c8 000000f0 cd40ea00 cd375fec 1b883e89
[ 3000.826374]  0000026b cd40ea00 80000000 00000001 cd0075c8 00000000 cd375de4 c119917f
[ 3000.834492]  cd563360 cd375df4 c119a0ab cd563360 00000000 cd375e24 c119a1d6 00000000
[ 3000.842616] Call Trace:
[ 3000.845252]  [<c119917f>] ? kernfs_next_descendant_post+0x3f/0x50
[ 3000.851543]  [<c119a0ab>] ? kernfs_activate+0x6b/0xc0
[ 3000.856790]  [<c119a1d6>] ? kernfs_add_one+0xd6/0x130
[ 3000.862047]  [<c15fdb02>] schedule+0x22/0x60
[ 3000.866548]  [<c15fd195>] schedule_timeout+0x175/0x1d0
[ 3000.871887]  [<c119b391>] ? __kernfs_create_file+0x71/0xa0
[ 3000.877574]  [<c119bb9a>] ? sysfs_add_file_mode_ns+0xaa/0x180
[ 3000.883533]  [<c15fe84f>] wait_for_completion+0x6f/0xb0
[ 3000.888961]  [<c1065200>] ? wake_up_process+0x40/0x40
[ 3000.894219]  [<c13cb600>] _request_firmware+0x750/0x9f0
[ 3000.899666]  [<c1382a7f>] ? n_tty_receive_buf2+0x1f/0x30
[ 3000.905200]  [<c13cba02>] request_firmware_work_func+0x22/0x50
[ 3000.911235]  [<c10550d2>] process_one_work+0x122/0x380
[ 3000.916571]  [<c1055859>] worker_thread+0xf9/0x470
[ 3000.921555]  [<c1055760>] ? create_and_start_worker+0x50/0x50
[ 3000.927497]  [<c1055760>] ? create_and_start_worker+0x50/0x50
[ 3000.933448]  [<c105a5ff>] kthread+0x9f/0xc0
[ 3000.937850]  [<c15ffd40>] ret_from_kernel_thread+0x20/0x30
[ 3000.943548]  [<c105a560>] ? kthread_worker_fn+0x100/0x100

This patch change the wait_for_completion() function call to
wait_for_completion_interruptible() function call for solving the issue.

Cc: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Kweh, Hock Leong <hock.leong.kweh@intel.com>
---
 drivers/base/firmware_class.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
index b2fe40b..c241454b 100644
--- a/drivers/base/firmware_class.c
+++ b/drivers/base/firmware_class.c
@@ -908,7 +908,7 @@ static int _request_firmware_load(struct firmware_priv *fw_priv,
 		kobject_uevent(&fw_priv->dev.kobj, KOBJ_ADD);
 	}
 
-	wait_for_completion(&buf->completion);
+	retval = wait_for_completion_interruptible(&buf->completion);
 
 	cancel_delayed_work_sync(&fw_priv->timeout_work);
 	if (!buf->data)
@@ -985,7 +985,7 @@ static int sync_cached_firmware_buf(struct firmware_buf *buf)
 			break;
 		}
 		mutex_unlock(&fw_lock);
-		wait_for_completion(&buf->completion);
+		ret = wait_for_completion_interruptible(&buf->completion);
 		mutex_lock(&fw_lock);
 	}
 	mutex_unlock(&fw_lock);
-- 
1.7.9.5

[PATCH v1 3/3] efi: Capsule update with user helper interface

[Kweh Hock Leong]

From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>

Introducing a kernel module to expose user helper interface for
user to upload capsule binaries. This module leverage the
request_firmware_nowait() to expose an interface to user.

Example steps to load the capsule binary:
1.) echo 1 > /sys/class/firmware/efi-capsule-file/loading
2.) cat capsule.bin > /sys/class/firmware/efi-capsule-file/data
3.) echo 0 > /sys/class/firmware/efi-capsule-file/loading

Whereas, this module does not allow the capsule binaries to be
obtained from the request_firmware library search path. If any
capsule binary loaded from firmware seach path, the module will
stop functioning.

Besides the request_firmware user helper interface, this module
also expose a 'capsule_loaded' file note for user to verify
the number of successfully uploaded capsule binaries. This
file note has the read only attribute.

Cc: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Kweh, Hock Leong <hock.leong.kweh@intel.com>
---
 drivers/firmware/efi/Kconfig                   |   13 ++
 drivers/firmware/efi/Makefile                  |    1 +
 drivers/firmware/efi/efi-capsule-user-helper.c |  246 ++++++++++++++++++++++++
 3 files changed, 260 insertions(+)
 create mode 100644 drivers/firmware/efi/efi-capsule-user-helper.c

diff --git a/drivers/firmware/efi/Kconfig b/drivers/firmware/efi/Kconfig
index f712d47..7dc814e 100644
--- a/drivers/firmware/efi/Kconfig
+++ b/drivers/firmware/efi/Kconfig
@@ -60,6 +60,19 @@ config EFI_RUNTIME_WRAPPERS
 config EFI_ARMSTUB
 	bool
 
+config EFI_CAPSULE_USER_HELPER
+	tristate "EFI capsule user mode helper"
+	depends on EFI
+	select FW_LOADER
+	select FW_LOADER_USER_HELPER
+	help
+	  This option exposes the user mode helper interface for user to load
+	  EFI capsule binary and update the EFI firmware after system reboot.
+	  This feature does not support auto locating capsule binaries at the
+	  firmware lib search path.
+
+	  If unsure, say N.
+
 endmenu
 
 config UEFI_CPER
diff --git a/drivers/firmware/efi/Makefile b/drivers/firmware/efi/Makefile
index 698846e..63f6910 100644
--- a/drivers/firmware/efi/Makefile
+++ b/drivers/firmware/efi/Makefile
@@ -8,3 +8,4 @@ obj-$(CONFIG_UEFI_CPER)			+= cper.o
 obj-$(CONFIG_EFI_RUNTIME_MAP)		+= runtime-map.o
 obj-$(CONFIG_EFI_RUNTIME_WRAPPERS)	+= runtime-wrappers.o
 obj-$(CONFIG_EFI_STUB)			+= libstub/
+obj-$(CONFIG_EFI_CAPSULE_USER_HELPER)	+= efi-capsule-user-helper.o
diff --git a/drivers/firmware/efi/efi-capsule-user-helper.c b/drivers/firmware/efi/efi-capsule-user-helper.c
new file mode 100644
index 0000000..84a1628
--- /dev/null
+++ b/drivers/firmware/efi/efi-capsule-user-helper.c
@@ -0,0 +1,246 @@
+/*
+ * EFI capsule user mode helper interface driver.
+ *
+ * Copyright 2014 Intel Corporation
+ *
+ * This file is part of the Linux kernel, and is made available under
+ * the terms of the GNU General Public License version 2.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/slab.h>
+#include <linux/mutex.h>
+#include <linux/reboot.h>
+#include <linux/efi.h>
+#include <linux/firmware.h>
+
+#define CAPSULE_NAME "efi-capsule-file"
+#define DEV_NAME "efi_capsule_user_helper"
+#define STRING_INTEGER_MAX_LENGTH 13
+
+static DEFINE_MUTEX(user_helper_lock);
+static int capsule_count;
+static int stop_request;
+static struct platform_device *efi_capsule_pdev;
+
+/*
+ * This function will store the capsule binary and pass it to
+ * efi_capsule_update() API in capsule.c
+ */
+static int efi_capsule_store(const struct firmware *fw)
+{
+	int i;
+	int ret;
+	int count = fw->size;
+	int copy_size = (fw->size > PAGE_SIZE) ? PAGE_SIZE : fw->size;
+	int pages_needed = ALIGN(fw->size, PAGE_SIZE) >> PAGE_SHIFT;
+	struct page **pages;
+	void *page_data;
+	efi_capsule_header_t *capsule = NULL;
+
+	pages = kmalloc_array(pages_needed, sizeof(void *), GFP_KERNEL);
+	if (!pages) {
+		pr_err("%s: kmalloc_array() failed\n", __func__);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < pages_needed; i++) {
+		pages[i] = alloc_page(GFP_KERNEL);
+		if (!pages[i]) {
+			pr_err("%s: alloc_page() failed\n", __func__);
+			--i;
+			ret = -ENOMEM;
+			goto failed;
+		}
+	}
+
+	for (i = 0; i < pages_needed; i++) {
+		page_data = kmap(pages[i]);
+		memcpy(page_data, fw->data + (i * PAGE_SIZE), copy_size);
+
+		if (i == 0)
+			capsule = (efi_capsule_header_t *)page_data;
+		else
+			kunmap(pages[i]);
+
+		count -= copy_size;
+		if (count < PAGE_SIZE)
+			copy_size = count;
+	}
+
+	ret = efi_capsule_update(capsule, pages);
+	if (ret) {
+		pr_err("%s: efi_capsule_update() failed\n", __func__);
+		--i;
+		goto failed;
+	}
+	++capsule_count;
+	kunmap(pages[0]);
+
+	/*
+	 * we cannot free the pages here due to reboot need that data
+	 * maintained.
+	 */
+	return 0;
+
+failed:
+	while (i >= 0)
+		__free_page(pages[i--]);
+	kfree(pages);
+	return ret;
+}
+
+/*
+ * This callback function will be called by request_firmware_nowait() when
+ * user has loaded the capsule binary or aborted user helper interface
+ */
+static void callbackfn_efi_capsule(const struct firmware *fw, void *context)
+{
+	int ret = 0;
+
+	if (fw) {
+		/*
+		 * Binary not getting from user helper interface, fw->pages
+		 * is equal to NULL
+		 */
+		if (!fw->pages) {
+			pr_err("%s: ERROR: Capsule binary '%s' at %s\n",
+			       __func__, CAPSULE_NAME,
+			       "firmware lib search path are not supported");
+			pr_err("user helper interface disabled\n");
+			stop_request = 1;
+		} else {
+			efi_capsule_store(fw);
+		}
+		release_firmware(fw);
+	}
+
+	mutex_lock(&user_helper_lock);
+	if (!stop_request) {
+		ret = request_firmware_nowait(THIS_MODULE,
+					      FW_ACTION_NOHOTPLUG,
+					      CAPSULE_NAME,
+					      &efi_capsule_pdev->dev,
+					      GFP_KERNEL, NULL,
+					      callbackfn_efi_capsule);
+		if (ret) {
+			pr_err("%s: request_firmware_nowait() failed\n",
+			       __func__);
+		}
+	}
+	mutex_unlock(&user_helper_lock);
+}
+
+static ssize_t capsule_loaded_show(struct device *dev,
+				   struct device_attribute *attr, char *buf)
+{
+	return snprintf(buf, STRING_INTEGER_MAX_LENGTH, "%d\n", capsule_count);
+}
+
+static DEVICE_ATTR_RO(capsule_loaded);
+
+/* stop user helper interface for reboot or module unload */
+static void capsule_stop_user_helper(void)
+{
+	mutex_lock(&user_helper_lock);
+	stop_request = 1;
+	mutex_unlock(&user_helper_lock);
+	request_firmware_abort(CAPSULE_NAME);
+}
+
+/* reboot notifier for avoid deadlock with usermode_lock */
+static int capsule_shutdown_notify(struct notifier_block *nb,
+				   unsigned long sys_state,
+				   void *reboot_cmd)
+{
+	capsule_stop_user_helper();
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block capsule_shutdown_nb = {
+	.notifier_call = capsule_shutdown_notify,
+	/*
+	 * In order to reboot properly, it is required to ensure the priority
+	 * here is higher than firmware_class fw_shutdown_nb priority
+	 */
+	.priority = 1,
+};
+
+/*
+ * Use request_firmware_nowait() exposing an user helper interface to obtain
+ * capsule binary from user space
+ */
+static int __init efi_capsule_user_helper_init(void)
+{
+	int ret;
+
+	register_reboot_notifier(&capsule_shutdown_nb);
+
+	efi_capsule_pdev = platform_device_register_simple(DEV_NAME,
+							   PLATFORM_DEVID_NONE,
+							   NULL, 0);
+	if (IS_ERR(efi_capsule_pdev)) {
+		pr_err("%s: platform_device_register_simple() failed\n",
+		       __func__);
+		return PTR_ERR(efi_capsule_pdev);
+	}
+
+	/*
+	 * create this file node for user to check the number of binaries that
+	 * are successfully loaded
+	 */
+	ret = device_create_file(&efi_capsule_pdev->dev,
+				 &dev_attr_capsule_loaded);
+	if (ret) {
+		pr_err("%s: device_create_file() failed\n", __func__);
+		return ret;
+	}
+
+	ret = request_firmware_nowait(THIS_MODULE, FW_ACTION_NOHOTPLUG,
+				      CAPSULE_NAME, &efi_capsule_pdev->dev,
+				      GFP_KERNEL, NULL, callbackfn_efi_capsule);
+	if (ret) {
+		pr_err("%s: request_firmware_nowait() failed\n", __func__);
+		goto out;
+	}
+
+	return 0;
+
+out:
+	device_remove_file(&efi_capsule_pdev->dev, &dev_attr_capsule_loaded);
+	unregister_reboot_notifier(&capsule_shutdown_nb);
+	return ret;
+}
+module_init(efi_capsule_user_helper_init);
+
+/*
+ * rmmod app itself will stop you to remove the module if the user
+ * helper interface is still exposing. In order to remove this driver,
+ * you are require to do the command as below:
+ * rmmod -f efi_capsule_user_helper.ko
+ */
+static void __exit efi_capsule_user_helper_exit(void)
+{
+	unregister_reboot_notifier(&capsule_shutdown_nb);
+
+	capsule_stop_user_helper();
+	/*
+	 * synchronization is needed to make sure request_firmware is fully
+	 * aborted
+	 */
+	while (efi_capsule_pdev->dev.kobj.kref.refcount.counter > 3)
+		msleep(20); /* avoid busy waiting for cooperative kernel */
+
+	device_remove_file(&efi_capsule_pdev->dev, &dev_attr_capsule_loaded);
+	platform_device_unregister(efi_capsule_pdev);
+}
+module_exit(efi_capsule_user_helper_exit);
+
+MODULE_DESCRIPTION("EFI Capsule user helper binary load utility");
+MODULE_LICENSE("GPL v2");
-- 
1.7.9.5

Re: [PATCH v1 2/3] firmware loader: fix hung task warning dump

[Ming Lei]

On Fri, Oct 24, 2014 at 2:18 PM, Kweh Hock Leong
<hock.leong.kweh@intel.com> wrote:
> From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>
>
> When using request_firmware_nowait() with FW_ACTION_NOHOTPLUG param to
> expose user helper interface, if the user do not react immediately, after
> 120 seconds there will be a hung task warning message dumped as below:
>
> [ 3000.784235] INFO: task kworker/0:0:8259 blocked for more than 120 seconds.
> [ 3000.791281]       Tainted: G            E 3.16.0-rc1-yocto-standard #41
> [ 3000.798082] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> [ 3000.806072] kworker/0:0     D cd0075c8     0  8259      2 0x00000000
> [ 3000.812765] Workqueue: events request_firmware_work_func
> [ 3000.818253]  cd375e18 00000046 0000000e cd0075c8 000000f0 cd40ea00 cd375fec 1b883e89
> [ 3000.826374]  0000026b cd40ea00 80000000 00000001 cd0075c8 00000000 cd375de4 c119917f
> [ 3000.834492]  cd563360 cd375df4 c119a0ab cd563360 00000000 cd375e24 c119a1d6 00000000
> [ 3000.842616] Call Trace:
> [ 3000.845252]  [<c119917f>] ? kernfs_next_descendant_post+0x3f/0x50
> [ 3000.851543]  [<c119a0ab>] ? kernfs_activate+0x6b/0xc0
> [ 3000.856790]  [<c119a1d6>] ? kernfs_add_one+0xd6/0x130
> [ 3000.862047]  [<c15fdb02>] schedule+0x22/0x60
> [ 3000.866548]  [<c15fd195>] schedule_timeout+0x175/0x1d0
> [ 3000.871887]  [<c119b391>] ? __kernfs_create_file+0x71/0xa0
> [ 3000.877574]  [<c119bb9a>] ? sysfs_add_file_mode_ns+0xaa/0x180
> [ 3000.883533]  [<c15fe84f>] wait_for_completion+0x6f/0xb0
> [ 3000.888961]  [<c1065200>] ? wake_up_process+0x40/0x40
> [ 3000.894219]  [<c13cb600>] _request_firmware+0x750/0x9f0
> [ 3000.899666]  [<c1382a7f>] ? n_tty_receive_buf2+0x1f/0x30
> [ 3000.905200]  [<c13cba02>] request_firmware_work_func+0x22/0x50
> [ 3000.911235]  [<c10550d2>] process_one_work+0x122/0x380
> [ 3000.916571]  [<c1055859>] worker_thread+0xf9/0x470
> [ 3000.921555]  [<c1055760>] ? create_and_start_worker+0x50/0x50
> [ 3000.927497]  [<c1055760>] ? create_and_start_worker+0x50/0x50
> [ 3000.933448]  [<c105a5ff>] kthread+0x9f/0xc0
> [ 3000.937850]  [<c15ffd40>] ret_from_kernel_thread+0x20/0x30
> [ 3000.943548]  [<c105a560>] ? kthread_worker_fn+0x100/0x100
>
> This patch change the wait_for_completion() function call to
> wait_for_completion_interruptible() function call for solving the issue.
>
> Cc: Matt Fleming <matt.fleming@intel.com>
> Signed-off-by: Kweh, Hock Leong <hock.leong.kweh@intel.com>

Acked-by: Ming Lei <ming.lei@canonical.com>

Thanks,

Re: [PATCH v1 2/3] firmware loader: fix hung task warning dump

[Ming Lei]

On Fri, Oct 24, 2014 at 2:18 PM, Kweh Hock Leong
<hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> wrote:
> From: "Kweh, Hock Leong" <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
>
> When using request_firmware_nowait() with FW_ACTION_NOHOTPLUG param to
> expose user helper interface, if the user do not react immediately, after
> 120 seconds there will be a hung task warning message dumped as below:
>
> [ 3000.784235] INFO: task kworker/0:0:8259 blocked for more than 120 seconds.
> [ 3000.791281]       Tainted: G            E 3.16.0-rc1-yocto-standard #41
> [ 3000.798082] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> [ 3000.806072] kworker/0:0     D cd0075c8     0  8259      2 0x00000000
> [ 3000.812765] Workqueue: events request_firmware_work_func
> [ 3000.818253]  cd375e18 00000046 0000000e cd0075c8 000000f0 cd40ea00 cd375fec 1b883e89
> [ 3000.826374]  0000026b cd40ea00 80000000 00000001 cd0075c8 00000000 cd375de4 c119917f
> [ 3000.834492]  cd563360 cd375df4 c119a0ab cd563360 00000000 cd375e24 c119a1d6 00000000
> [ 3000.842616] Call Trace:
> [ 3000.845252]  [<c119917f>] ? kernfs_next_descendant_post+0x3f/0x50
> [ 3000.851543]  [<c119a0ab>] ? kernfs_activate+0x6b/0xc0
> [ 3000.856790]  [<c119a1d6>] ? kernfs_add_one+0xd6/0x130
> [ 3000.862047]  [<c15fdb02>] schedule+0x22/0x60
> [ 3000.866548]  [<c15fd195>] schedule_timeout+0x175/0x1d0
> [ 3000.871887]  [<c119b391>] ? __kernfs_create_file+0x71/0xa0
> [ 3000.877574]  [<c119bb9a>] ? sysfs_add_file_mode_ns+0xaa/0x180
> [ 3000.883533]  [<c15fe84f>] wait_for_completion+0x6f/0xb0
> [ 3000.888961]  [<c1065200>] ? wake_up_process+0x40/0x40
> [ 3000.894219]  [<c13cb600>] _request_firmware+0x750/0x9f0
> [ 3000.899666]  [<c1382a7f>] ? n_tty_receive_buf2+0x1f/0x30
> [ 3000.905200]  [<c13cba02>] request_firmware_work_func+0x22/0x50
> [ 3000.911235]  [<c10550d2>] process_one_work+0x122/0x380
> [ 3000.916571]  [<c1055859>] worker_thread+0xf9/0x470
> [ 3000.921555]  [<c1055760>] ? create_and_start_worker+0x50/0x50
> [ 3000.927497]  [<c1055760>] ? create_and_start_worker+0x50/0x50
> [ 3000.933448]  [<c105a5ff>] kthread+0x9f/0xc0
> [ 3000.937850]  [<c15ffd40>] ret_from_kernel_thread+0x20/0x30
> [ 3000.943548]  [<c105a560>] ? kthread_worker_fn+0x100/0x100
>
> This patch change the wait_for_completion() function call to
> wait_for_completion_interruptible() function call for solving the issue.
>
> Cc: Matt Fleming <matt.fleming-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> Signed-off-by: Kweh, Hock Leong <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>

Acked-by: Ming Lei <ming.lei-Z7WLFzj8eWMS+FvcfC7Uqw@public.gmane.org>

Thanks,

Re: [PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Ming Lei]

On Fri, Oct 24, 2014 at 2:18 PM, Kweh Hock Leong
<hock.leong.kweh@intel.com> wrote:
> From: "Kweh, Hock Leong" <hock.leong.kweh@intel.com>
>
> Besides aborting through user helper interface, a new API
> request_firmware_abort() allows kernel driver module to abort the
> request_firmware() / request_firmware_nowait() when they are no
> longer needed. It is useful for cancelling an outstanding firmware
> load if initiated from inside a module where that module is in the
> process of being unloaded, since the unload function may not have
> a handle to the struct firmware_buf.
>
> Note for people who use request_firmware_nowait():
> You are required to do your own synchronization after you call
> request_firmware_abort() in order to continue unloading the
> module. The example synchronization code shows below:
>
> while (THIS_MODULE && module_refcount(THIS_MODULE))
>         msleep(20);
>
> Cc: Matt Fleming <matt.fleming@intel.com>
> Signed-off-by: Kweh, Hock Leong <hock.leong.kweh@intel.com>
> ---
>  drivers/base/firmware_class.c |   30 ++++++++++++++++++++++++++++++
>  include/linux/firmware.h      |    4 ++++
>  2 files changed, 34 insertions(+)
>
> diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
> index d276e33..b2fe40b 100644
> --- a/drivers/base/firmware_class.c
> +++ b/drivers/base/firmware_class.c
> @@ -1291,6 +1291,36 @@ request_firmware_nowait(
>  }
>  EXPORT_SYMBOL(request_firmware_nowait);
>
> +/**
> + * request_firmware_abort - abort the waiting of firmware request
> + * @fw_name: name of firmware file
> + *
> + *     @fw_name is the same name string while calling to request_firmware()
> + *             or request_firmware_nowait().
> + **/
> +int request_firmware_abort(const char *fw_name)
> +{
> +       struct firmware_buf *buf;
> +       struct firmware_buf *entry;
> +
> +       spin_lock(&fw_cache.lock);
> +       buf = __fw_lookup_buf(fw_name);
> +       spin_unlock(&fw_cache.lock);

You can call fw_lookup_buf() directly, otherwise feel free to add:

Acked-by: Ming Lei <ming.lei@canonical.com>

> +       if (!buf)
> +               return -ENOENT;
> +
> +       mutex_lock(&fw_lock);
> +       list_for_each_entry(entry, &pending_fw_head, pending_list) {
> +               if (entry == buf) {
> +                       __fw_load_abort(buf);
> +                       break;
> +               }
> +       }
> +       mutex_unlock(&fw_lock);
> +       return 0;
> +}
> +EXPORT_SYMBOL(request_firmware_abort);
> +
>  #ifdef CONFIG_PM_SLEEP
>  static ASYNC_DOMAIN_EXCLUSIVE(fw_cache_domain);
>
> diff --git a/include/linux/firmware.h b/include/linux/firmware.h
> index 5952933..ed90c8b 100644
> --- a/include/linux/firmware.h
> +++ b/include/linux/firmware.h
> @@ -47,6 +47,7 @@ int request_firmware_nowait(
>         void (*cont)(const struct firmware *fw, void *context));
>
>  void release_firmware(const struct firmware *fw);
> +int request_firmware_abort(const char *fw_name);
>  #else
>  static inline int request_firmware(const struct firmware **fw,
>                                    const char *name,
> @@ -66,6 +67,9 @@ static inline void release_firmware(const struct firmware *fw)
>  {
>  }
>
> +static inline int request_firmware_abort(const char *fw_name)
> +{
> +}
>  #endif
>
>  #ifdef CONFIG_FW_LOADER_USER_HELPER
> --
> 1.7.9.5
>

Re: [PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Ming Lei]

On Fri, Oct 24, 2014 at 2:18 PM, Kweh Hock Leong
<hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> wrote:
> From: "Kweh, Hock Leong" <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
>
> Besides aborting through user helper interface, a new API
> request_firmware_abort() allows kernel driver module to abort the
> request_firmware() / request_firmware_nowait() when they are no
> longer needed. It is useful for cancelling an outstanding firmware
> load if initiated from inside a module where that module is in the
> process of being unloaded, since the unload function may not have
> a handle to the struct firmware_buf.
>
> Note for people who use request_firmware_nowait():
> You are required to do your own synchronization after you call
> request_firmware_abort() in order to continue unloading the
> module. The example synchronization code shows below:
>
> while (THIS_MODULE && module_refcount(THIS_MODULE))
>         msleep(20);
>
> Cc: Matt Fleming <matt.fleming-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> Signed-off-by: Kweh, Hock Leong <hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> ---
>  drivers/base/firmware_class.c |   30 ++++++++++++++++++++++++++++++
>  include/linux/firmware.h      |    4 ++++
>  2 files changed, 34 insertions(+)
>
> diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
> index d276e33..b2fe40b 100644
> --- a/drivers/base/firmware_class.c
> +++ b/drivers/base/firmware_class.c
> @@ -1291,6 +1291,36 @@ request_firmware_nowait(
>  }
>  EXPORT_SYMBOL(request_firmware_nowait);
>
> +/**
> + * request_firmware_abort - abort the waiting of firmware request
> + * @fw_name: name of firmware file
> + *
> + *     @fw_name is the same name string while calling to request_firmware()
> + *             or request_firmware_nowait().
> + **/
> +int request_firmware_abort(const char *fw_name)
> +{
> +       struct firmware_buf *buf;
> +       struct firmware_buf *entry;
> +
> +       spin_lock(&fw_cache.lock);
> +       buf = __fw_lookup_buf(fw_name);
> +       spin_unlock(&fw_cache.lock);

You can call fw_lookup_buf() directly, otherwise feel free to add:

Acked-by: Ming Lei <ming.lei-Z7WLFzj8eWMS+FvcfC7Uqw@public.gmane.org>

> +       if (!buf)
> +               return -ENOENT;
> +
> +       mutex_lock(&fw_lock);
> +       list_for_each_entry(entry, &pending_fw_head, pending_list) {
> +               if (entry == buf) {
> +                       __fw_load_abort(buf);
> +                       break;
> +               }
> +       }
> +       mutex_unlock(&fw_lock);
> +       return 0;
> +}
> +EXPORT_SYMBOL(request_firmware_abort);
> +
>  #ifdef CONFIG_PM_SLEEP
>  static ASYNC_DOMAIN_EXCLUSIVE(fw_cache_domain);
>
> diff --git a/include/linux/firmware.h b/include/linux/firmware.h
> index 5952933..ed90c8b 100644
> --- a/include/linux/firmware.h
> +++ b/include/linux/firmware.h
> @@ -47,6 +47,7 @@ int request_firmware_nowait(
>         void (*cont)(const struct firmware *fw, void *context));
>
>  void release_firmware(const struct firmware *fw);
> +int request_firmware_abort(const char *fw_name);
>  #else
>  static inline int request_firmware(const struct firmware **fw,
>                                    const char *name,
> @@ -66,6 +67,9 @@ static inline void release_firmware(const struct firmware *fw)
>  {
>  }
>
> +static inline int request_firmware_abort(const char *fw_name)
> +{
> +}
>  #endif
>
>  #ifdef CONFIG_FW_LOADER_USER_HELPER
> --
> 1.7.9.5
>

RE: [PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Kweh, Hock Leong]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8", Size: 746 bytes --]

> -----Original Message-----
> From: Ming Lei [mailto:ming.lei@canonical.com]
> Sent: Monday, October 27, 2014 12:00 AM
> 
> You can call fw_lookup_buf() directly, otherwise feel free to add:
> 
> Acked-by: Ming Lei <ming.lei@canonical.com>
> 

Hi Ming Lei,

The fw_lookup_buf() is defined inside the conditional preprocessor directive for CONFIG_PM_SLEEP.
Since the request_firmware_abort() may not only be used in PM_SLEEP, could I move the fw_lookup_buf()
out from the CONFIG_PM_SLEEP block if we want to call fw_lookup_buf() instead of __fw_lookup_buf()?

Thanks.


Regards,
Wilson
ÿôèº{.nÇ+‰·Ÿ®‰­†+%ŠËÿ±éݶ\x17¥Šwÿº{.nÇ+‰·¥Š{±þG«éÿŠ{ayº\x1dʇڙë,j\a­¢f£¢·hšïêÿ‘êçz_è®\x03(­éšŽŠÝ¢j"ú\x1a¶^[m§ÿÿ¾\a«þG«éÿ¢¸?™¨è­Ú&£ø§~á¶iO•æ¬z·švØ^\x14\x04\x1a¶^[m§ÿÿÃ\fÿ¶ìÿ¢¸?–I¥

RE: [PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Kweh, Hock Leong]

> -----Original Message-----
> From: Ming Lei [mailto:ming.lei@canonical.com]
> Sent: Monday, October 27, 2014 12:00 AM
> 
> You can call fw_lookup_buf() directly, otherwise feel free to add:
> 
> Acked-by: Ming Lei <ming.lei@canonical.com>
> 

Hi Ming Lei,

The fw_lookup_buf() is defined inside the conditional preprocessor directive for CONFIG_PM_SLEEP.
Since the request_firmware_abort() may not only be used in PM_SLEEP, could I move the fw_lookup_buf()
out from the CONFIG_PM_SLEEP block if we want to call fw_lookup_buf() instead of __fw_lookup_buf()?

Thanks.


Regards,
Wilson

Re: [PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Ming Lei]

On Wed, Oct 29, 2014 at 11:04 AM, Kweh, Hock Leong
<hock.leong.kweh@intel.com> wrote:
>> -----Original Message-----
>> From: Ming Lei [mailto:ming.lei@canonical.com]
>> Sent: Monday, October 27, 2014 12:00 AM
>>
>> You can call fw_lookup_buf() directly, otherwise feel free to add:
>>
>> Acked-by: Ming Lei <ming.lei@canonical.com>
>>
>
> Hi Ming Lei,
>
> The fw_lookup_buf() is defined inside the conditional preprocessor directive for CONFIG_PM_SLEEP.
> Since the request_firmware_abort() may not only be used in PM_SLEEP, could I move the fw_lookup_buf()
> out from the CONFIG_PM_SLEEP block if we want to call fw_lookup_buf() instead of __fw_lookup_buf()?

Sure.

Thanks,

Re: [PATCH v1 1/3] firmware loader: Introduce new API - request_firmware_abort()

[Ming Lei]

On Wed, Oct 29, 2014 at 11:04 AM, Kweh, Hock Leong
<hock.leong.kweh-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> wrote:
>> -----Original Message-----
>> From: Ming Lei [mailto:ming.lei-Z7WLFzj8eWMS+FvcfC7Uqw@public.gmane.org]
>> Sent: Monday, October 27, 2014 12:00 AM
>>
>> You can call fw_lookup_buf() directly, otherwise feel free to add:
>>
>> Acked-by: Ming Lei <ming.lei-Z7WLFzj8eWMS+FvcfC7Uqw@public.gmane.org>
>>
>
> Hi Ming Lei,
>
> The fw_lookup_buf() is defined inside the conditional preprocessor directive for CONFIG_PM_SLEEP.
> Since the request_firmware_abort() may not only be used in PM_SLEEP, could I move the fw_lookup_buf()
> out from the CONFIG_PM_SLEEP block if we want to call fw_lookup_buf() instead of __fw_lookup_buf()?

Sure.

Thanks,