Skip to contentSkip to conversation

firmware loader: use statically initialized data attribute

PATCH In review Dmitry Torokhov 2 participants 2 messages Mar 11, 2010 – Mar 13, 2010 Driver core

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 firmware loader: use statically initialized data attribute

There is no reason why we are using a template for binary attribute and copying it into per-firmware data before registering. Using the original works as well and gets rid of the following lockdep compaint:

platform microcode: firmware: requesting intel-ucode/06-0f-0b

 BUG: key ffff88011371be70 not in .data!
 ------------[ cut here ]------------
 WARNING: at kernel/lockdep.c:2706 lockdep_init_map+0x125/0x140()
 Hardware name: Latitude D630
 Modules linked in: ...
 Pid: 738, comm: modprobe Tainted: P           2.6.34-rc1 #237
 Call Trace:
  [<ffffffff81048fb6>] warn_slowpath_common+0x76/0xb0
  [<ffffffff81048fff>] warn_slowpath_null+0xf/0x20
  [<ffffffff8107c915>] lockdep_init_map+0x125/0x140
  [<ffffffff8118485a>] sysfs_add_file_mode+0x6a/0xc0
  [<ffffffff811848bc>] sysfs_add_file+0xc/0x10
  [<ffffffff811870e1>] sysfs_create_bin_file+0x21/0x30
  [<ffffffff8130dff1>] fw_setup_device+0x81/0x120
  [<ffffffff8130e14f>] _request_firmware+0xbf/0x270
  [<ffffffff8130e38e>] request_firmware+0xe/0x10
  [<ffffffffa0064be1>] request_microcode_fw+0x61/0xa0 [microcode]
  [<ffffffffa00643b8>] microcode_init_cpu+0xb8/0xd0 [microcode]
  [<ffffffffa0064426>] mc_sysdev_add+0x56/0x70 [microcode]
  [<ffffffff813032ae>] sysdev_driver_register+0x9e/0x130
  [<ffffffffa006a000>] ? microcode_init+0x0/0x12a [microcode]
  [<ffffffffa006a0bd>] microcode_init+0xbd/0x12a [microcode]
  [<ffffffff810001d7>] do_one_initcall+0x37/0x190
  [<ffffffff8108d178>] sys_init_module+0xd8/0x250
  [<ffffffff81002fab>] system_call_fastpath+0x16/0x1b
 ---[ end trace 93c9f72439beee7f ]---

---

 drivers/base/firmware_class.c |    8 +++-----
 1 files changed, 3 insertions(+), 5 deletions(-)
drivers/base/firmware_class.c +3 -5
index d0dc26a..3f50b2e 100644
@@ -41,7 +41,6 @@ static DEFINE_MUTEX(fw_lock);
 struct firmware_priv {
 	char *fw_id;
 	struct completion completion;
-	struct bin_attribute attr_data;
 	struct firmware *fw;
 	unsigned long status;
 	struct page **pages;
@@ -344,8 +343,8 @@ out:
 	return retval;
 }
 
-static struct bin_attribute firmware_attr_data_tmpl = {
-	.attr = {.name = "data", .mode = 0644},
+static struct bin_attribute firmware_attr_data = {
+	.attr = { .name = "data", .mode = 0644 },
 	.size = 0,
 	.read = firmware_data_read,
 	.write = firmware_data_write,
@@ -390,7 +389,6 @@ static int fw_register_device(struct device **dev_p, const char *fw_name,
 	}
 
 	init_completion(&fw_priv->completion);
-	fw_priv->attr_data = firmware_attr_data_tmpl;
 	fw_priv->fw_id = kstrdup(fw_name, GFP_KERNEL);
 	if (!fw_priv->fw_id) {
 		dev_err(device, "%s: Firmware name allocation failed\n",
@@ -442,7 +440,7 @@ static int fw_setup_device(struct firmware *fw, struct device **dev_p,
 	fw_priv = dev_get_drvdata(f_dev);
 
 	fw_priv->fw = fw;
-	retval = sysfs_create_bin_file(&f_dev->kobj, &fw_priv->attr_data);
+	retval = sysfs_create_bin_file(&f_dev->kobj, &firmware_attr_data);
 	if (retval) {
 		dev_err(device, "%s: sysfs_create_bin_file failed\n", __func__);
 		goto error_unreg;
Signature · 1 line
Dmitry
Acked-by on v1 1/1
#1 lore raw
Quoted 4 lines · Dmitry Torokhov <dmitry.torokhov <at> gmail.com> writes:

There is no reason why we are using a template for binary attribute and copying it into per-firmware data before registering. Using the original works as well and gets rid of the following lockdep compaint:

[...]

Quoted 1 line
Signed-off-by Dmitry Torokhov <dtor <at> mail.ru>