Skip to contentSkip to conversation

driver core fixes against 2.6.25-rc2 git

PATCH In review Greg KH 3 participants 16 messages Feb 21, 2008 – Feb 22, 2008

Subsystems (11)
Documentation PowerPC ACPI PCI +7 more subsystems
lore
Share & tools
lore

Greg KH

author
#0 lore raw

PATCH 1/1 driver core fixes against 2.6.25-rc2 git

Here are a few driver core fixes against your current git tree that fix some more problems that have cropped up:
- shutdown problem due to logic problem with cpufreq usage of kobjects.
- build fix for powerpc due to previous kobject changes.
- runtime errors when CONFIG_SYSFS=n
- UIO code now works properly from my previous messups
- proper encoding of the ja_JP stable_kernel_rules.txt file
- updates to the stable_kernel_rules.txt file
- mark ide=reverse as obsolete in preparation of 2.6.26 (Bart wanted this to go in through my tree as I have the .26 patches pending.)
- other minor fixes.

Please pull from:
master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6.git/

Patches will be sent as a follow-on to this message to lkml for people to see.

thanks,

greg k-h

------------

 Documentation/ja_JP/stable_kernel_rules.txt |  106 +++++++++++++-------------
 Documentation/stable_kernel_rules.txt       |    7 +-
 Documentation/vm/slabinfo.c                 |    2 +-
 arch/powerpc/platforms/pseries/power.c      |    2 +-
 drivers/base/bus.c                          |    9 +-
 drivers/base/driver.c                       |    3 +
 drivers/base/power/main.c                   |    2 +-
 drivers/cpufreq/cpufreq.c                   |    8 --
 drivers/ide/ide.c                           |    2 +-
 drivers/uio/uio.c                           |   54 +++++++++----
 kernel/module.c                             |    7 +-
 11 files changed, 109 insertions(+), 93 deletions(-)

---------------

Balaji Rao (1):
cpufreq: fix kobject reference count handling

Bartlomiej Zolnierkiewicz (1):
ide: mark "ide=reverse" option as obsolete

Brandon Philips (1):
UIO: fix Greg's stupid changes

Christoph Lameter (1):
slabinfo: fall back from /sys/kernel/slab to /sys/slab

Cornelia Huck (1):
Driver core: Fix error handling in bus_add_driver().

Kay Sievers (1):
modules: do not try to add sysfs attributes if !CONFIG_SYSFS

Rafael J. Wysocki (1):
PM: Remove unbalanced mutex_unlock() from dpm_resume()

Randy Dunlap (1):
driver-core: fix kernel-doc function parameters

Stefan Richter (1):
stable_kernel_rules: fix must already be in mainline

Stephen Rothwell (1):
POWERPC: fix typo in pseries/power.c

Tsugikazu Shibata (1):
Fix broken utf-8 encodings in ja_JP translation of stable_kernel_rules.txt

#1 lore raw

PATCH 1/11 Fix broken utf-8 encodings in ja_JP translation of stable_kernel_rules.txt

From: Tsugikazu Shibata <[email protected]>

Fix the UTF-8 broken encodings in the ja_JP version of stable_kernel_rules.txt

From: Tsugikazu Shibata <[email protected]>

---
 Documentation/ja_JP/stable_kernel_rules.txt |  106 +++++++++++++-------------
 1 files changed, 53 insertions(+), 53 deletions(-)
Documentation/ja_JP/stable_kernel_rules.txt +53 -53
index 17d8751..b3ffe87 100644
@@ -11,69 +11,69 @@ comment or update of this file, please try to update Original(English)
 file at first.
 
 ==================================
-これは、
+これは、
 linux-2.6.24/Documentation/stable_kernel_rules.txt
-の和訳です。
+の和訳です。
 
-翻訳団体: JF プロジェクト < http://www.linux.or.jp/JF/ >
-翻訳日: 2007/12/30
-翻訳者: Tsugikazu Shibata <tshibata at ab dot jp dot nec dot com>
-校正者: 武井伸光さん、<takei at webmasters dot gr dot jp>
-         かねこさん (Seiji Kaneko) <skaneko at a2 dot mbn dot or dot jp>
-         小林 雅典さん (Masanori Kobayasi) <zap03216 at nifty dot ne dot jp>
-         野口さん (Kenji Noguchi) <tokyo246 at gmail dot com>
-         神宮信太郎さん <jin at libjingu dot jp>
+翻訳団体: JF プロジェクト < http://www.linux.or.jp/JF/ >
+翻訳日: 2007/12/30
+翻訳者: Tsugikazu Shibata <tshibata at ab dot jp dot nec dot com>
+校正者: 武井伸光さん、<takei at webmasters dot gr dot jp>
+         かねこさん (Seiji Kaneko) <skaneko at a2 dot mbn dot or dot jp>
+         小林 雅典さん (Masanori Kobayasi) <zap03216 at nifty dot ne dot jp>
+         野口さん (Kenji Noguchi) <tokyo246 at gmail dot com>
+         神宮信太郎さん <jin at libjingu dot jp>
 ==================================
 
-ずっと知りたかった Linux 2.6 -stable リリースの全て
+ずっと知りたかった Linux 2.6 -stable リリースの全て
 
-"-stable" ツリーにどのような種類のパッチが受け入れられるか、どのような
-ものが受け入れられないか、についての規則-
+"-stable" ツリーにどのような種類のパッチが受け入れられるか、どのような
+ものが受け入れられないか、についての規則-
 
- - 明らかに正しく、テストされているものでなければならない。
- - 文脈(変更行の前後)を含めて 100 行より大きくてはいけない。
- - ただ一個のことだけを修正しているべき。
- - 皆を悩ませている本物のバグを修正しなければならない。("これはバグで
-   あるかもしれないが..." のようなものではない)
- - ビルドエラー(CONFIG_BROKENになっているものを除く), oops, ハング、デー
-   タ破壊、現実のセキュリティ問題、その他 "ああ、これはダメだね"という
-   ようなものを修正しなければならない。短く言えば、重大な問題。
- - どのように競合状態が発生するかの説明も一緒に書かれていない限り、
-   "理論的には競合状態になる"ようなものは不可。
- - いかなる些細な修正も含めることはできない。(スペルの修正、空白のクリー
-   ンアップなど)
- - 対応するサブシステムメンテナが受け入れたものでなければならない。
- - Documentation/SubmittingPatches の規則に従ったものでなければならない。
+ - 明らかに正しく、テストされているものでなければならない。
+ - 文脈(変更行の前後)を含めて 100 行より大きくてはいけない。
+ - ただ一個のことだけを修正しているべき。
+ - 皆を悩ませている本物のバグを修正しなければならない。("これはバグで
+   あるかもしれないが..." のようなものではない)
+ - ビルドエラー(CONFIG_BROKENになっているものを除く), oops, ハング、デー
+   タ破壊、現実のセキュリティ問題、その他 "ああ、これはダメだね"という
+   ようなものを修正しなければならない。短く言えば、重大な問題。
+ - どのように競合状態が発生するかの説明も一緒に書かれていない限り、
+   "理論的には競合状態になる"ようなものは不可。
+ - いかなる些細な修正も含めることはできない。(スペルの修正、空白のクリー
+   ンアップなど)
+ - 対応するサブシステムメンテナが受け入れたものでなければならない。
+ - Documentation/SubmittingPatches の規則に従ったものでなければならない。
 
--stable ツリーにパッチを送付する手続き-
+-stable ツリーにパッチを送付する手続き-
 
- - 上記の規則に従っているかを確認した後に、[email protected] にパッチ
-   を送る。
- - 送信者はパッチがキューに受け付けられた際には ACK を、却下された場合
-   には NAK を受け取る。この反応は開発者たちのスケジュールによって、数
-   日かかる場合がある。
- - もし受け取られたら、パッチは他の開発者たちのレビューのために
-   -stable キューに追加される。
- - セキュリティパッチはこのエイリアス ([email protected]) に送られるべ
-   きではなく、代わりに [email protected] のアドレスに送られる。
+ - 上記の規則に従っているかを確認した後に、[email protected] にパッチ
+   を送る。
+ - 送信者はパッチがキューに受け付けられた際には ACK を、却下された場合
+   には NAK を受け取る。この反応は開発者たちのスケジュールによって、数
+   日かかる場合がある。
+ - もし受け取られたら、パッチは他の開発者たちのレビューのために
+   -stable キューに追加される。
+ - セキュリティパッチはこのエイリアス ([email protected]) に送られるべ
+   きではなく、代わりに [email protected] のアドレスに送られる。
 
-レビューサイクル-
+レビューサイクル-
 
- - -stable メンテナがレビューサイクルを決めるとき、パッチはレビュー委
-   員会とパッチが影響する領域のメンテナ(提供者がその領域のメンテナで無
-   い限り)に送られ、linux-kernel メーリングリストにCCされる。
- - レビュー委員会は 48時間の間に ACK か NAK を出す。
- - もしパッチが委員会のメンバから却下れるか、メンテナ達やメンバが気付
-   かなかった問題が持ちあがり、linux-kernel メンバがパッチに異議を唱え
-   た場合には、パッチはキューから削除される。
- - レビューサイクルの最後に、ACK を受けたパッチは最新の -stable リリー
-   スに追加され、その後に新しい -stable リリースが行われる。
- - セキュリティパッチは、通常のレビューサイクルを通らず、セキュリティ
-   カーネルチームから直接 -stable ツリーに受け付けられる。
-   この手続きの詳細については kernel security チームに問い合わせること。
+ - -stable メンテナがレビューサイクルを決めるとき、パッチはレビュー委
+   員会とパッチが影響する領域のメンテナ(提供者がその領域のメンテナで無
+   い限り)に送られ、linux-kernel メーリングリストにCCされる。
+ - レビュー委員会は 48時間の間に ACK か NAK を出す。
+ - もしパッチが委員会のメンバから却下されるか、メンテナ達やメンバが気付
+   かなかった問題が持ちあがり、linux-kernel メンバがパッチに異議を唱え
+   た場合には、パッチはキューから削除される。
+ - レビューサイクルの最後に、ACK を受けたパッチは最新の -stable リリー
+   スに追加され、その後に新しい -stable リリースが行われる。
+ - セキュリティパッチは、通常のレビューサイクルを通らず、セキュリティ
+   カーネルチームから直接 -stable ツリーに受け付けられる。
+   この手続きの詳細については kernel security チームに問い合わせること。
 
-レビュー委員会-
+レビュー委員会-
 
- - この委員会は、このタスクについて活動する多くのボランティアと、少数の
-   非ボランティアのカーネル開発者達で構成されている。
+ - この委員会は、このタスクについて活動する多くのボランティアと、少数の
+   非ボランティアのカーネル開発者達で構成されている。
 
Signature · 1 line
1.5.4
#2 lore raw

PATCH 2/11 slabinfo: fall back from /sys/kernel/slab to /sys/slab

In reply to Greg KH

From: Christoph Lameter <[email protected]>

I keep running upstream and mm kernels and the location of the slab directory is different since upstream still uses /sys/slab. This patch makes slabinfo check /sys/slab if /sys/kernel/slab is not there. Makes slabinfo work on any kernel.

---
 Documentation/vm/slabinfo.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
Documentation/vm/slabinfo.c +1 -1
index 7123fee..22d7e3e 100644
@@ -1123,7 +1123,7 @@ void read_slab_dir(void)
 	char *t;
 	int count;
 
-	if (chdir("/sys/kernel/slab"))
+	if (chdir("/sys/kernel/slab") && chdir("/sys/slab"))
 		fatal("SYSFS support for SLUB not active\n");
 
 	dir = opendir(".");
Signature · 1 line
1.5.4
#3 lore raw

PATCH 3/11 cpufreq: fix kobject reference count handling

In reply to Greg KH

From: Balaji Rao <[email protected]>

The cpufreq core should not take an extra kobject reference count for no reason, and then refuse to release it. This has been reported as keeping machines from properly powering down all the way.

Cc Dave Jones <[email protected]>
Cc Yi Yang <[email protected]>
Cc Alan Stern <[email protected]>
Cc Frans Pop <[email protected]>
Cc Yinghai Lu <[email protected]>
Cc Ingo Molnar <[email protected]>
---
 drivers/cpufreq/cpufreq.c |    8 --------
 1 files changed, 0 insertions(+), 8 deletions(-)
drivers/cpufreq/cpufreq.c +0 -8
index 64926aa..89a29cd 100644
@@ -1006,14 +1006,6 @@ static int __cpufreq_remove_dev (struct sys_device * sys_dev)
 	}
 #endif
 
-
-	if (!kobject_get(&data->kobj)) {
-		spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
-		cpufreq_debug_enable_ratelimit();
-		unlock_policy_rwsem_write(cpu);
-		return -EFAULT;
-	}
-
 #ifdef CONFIG_SMP
 
 #ifdef CONFIG_HOTPLUG_CPU
Signature · 1 line
1.5.4
#4 lore raw

Re: [PATCH 03/11] cpufreq: fix kobject reference count handling

Quoted 7 lines · Greg Kroah-Hartman

From: Balaji Rao <[email protected]>

The cpufreq core should not take an extra kobject reference count for no reason, and then refuse to release it. This has been reported as keeping machines from properly powering down all the way.

Quoted 8 lines
Cc Dave Jones <[email protected]>
Cc Yi Yang <[email protected]>
Cc Alan Stern <[email protected]>
Cc Frans Pop <[email protected]>
Cc Yinghai Lu <[email protected]>
Cc Ingo Molnar <[email protected]>

ACK. This needs to go in soon, so I can stop hearing about it :-)

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#5 lore raw

PATCH 4/11 driver-core: fix kernel-doc function parameters

In reply to Greg KH

From: Randy Dunlap <[email protected]>

Fix drivers/base/ missing kernel-doc parameters:
Warning(linux-2.6.24-git12//drivers/base/driver.c:133): No description found for parameter 'drv' Warning(linux-2.6.24-git12//drivers/base/driver.c:133): No description found for parameter 'kobj' Warning(linux-2.6.24-git12//drivers/base/driver.c:133): No description found for parameter 'fmt' Warning(linux-2.6.24-git12//drivers/base/power/main.c:530): No description found for parameter 'state'

---
 drivers/base/driver.c     |    3 +++
 drivers/base/power/main.c |    1 +
 2 files changed, 4 insertions(+), 0 deletions(-)
drivers/base/driver.c +3 -0
index ba75184..bf31a01 100644
@@ -120,6 +120,9 @@ EXPORT_SYMBOL_GPL(driver_remove_file);
 
 /**
  * driver_add_kobj - add a kobject below the specified driver
+ * @drv: requesting device driver
+ * @kobj: kobject to add below this driver
+ * @fmt: format string that names the kobject
  *
  * You really don't want to do this, this is only here due to one looney
  * iseries driver, go poke those developers if you are annoyed about
drivers/base/power/main.c +1 -0
index bdc03f7..d609560 100644
@@ -523,6 +523,7 @@ static void lock_all_devices(void)
 
 /**
  *	device_suspend - Save state and stop all devices in system.
+ *	@state: new power management state
  *
  *	Prevent new devices from being registered, then lock all devices
  *	and suspend them.
Signature · 1 line
1.5.4
#6 lore raw

PATCH 5/11 Driver core: Fix error handling in bus_add_driver().

In reply to Greg KH

From: Cornelia Huck <[email protected]>

- If the allocation of ->priv fails, the reference on the bus must be dropped.
- If adding the kobject fails, kobject_put must be called to clean things up.

---
 drivers/base/bus.c |    9 +++++----
 1 files changed, 5 insertions(+), 4 deletions(-)
drivers/base/bus.c +5 -4
index 055989e..2d207ad 100644
@@ -658,9 +658,10 @@ int bus_add_driver(struct device_driver *drv)
 	pr_debug("bus: '%s': add driver %s\n", bus->name, drv->name);
 
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
-
+	if (!priv) {
+		error = -ENOMEM;
+		goto out_put_bus;
+	}
 	klist_init(&priv->klist_devices, NULL, NULL);
 	priv->driver = drv;
 	drv->p = priv;
@@ -668,7 +669,7 @@ int bus_add_driver(struct device_driver *drv)
 	error = kobject_init_and_add(&priv->kobj, &driver_ktype, NULL,
 				     "%s", drv->name);
 	if (error)
-		goto out_put_bus;
+		goto out_unregister;
 
 	if (drv->bus->p->drivers_autoprobe) {
 		error = driver_attach(drv);
Signature · 1 line
1.5.4
#7 lore raw

PATCH 6/11 ide: mark "ide=reverse" option as obsolete

In reply to Greg KH

From: Bartlomiej Zolnierkiewicz <[email protected]>

- it is valid only if "Probe IDE PCI devices in the PCI bus order (DEPRECATED)" config option is used

- Greg needs to remove pci_get_device_reverse() for PCI core changes

---
 drivers/ide/ide.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/ide/ide.c +1 -1
index 4a8952a..477833f 100644
@@ -1229,7 +1229,7 @@ static int __init ide_setup(char *s)
 	if (!strcmp(s, "ide=reverse")) {
 		ide_scan_direction = 1;
 		printk(" : Enabled support for IDE inverse scan order.\n");
-		return 1;
+		goto obsolete_option;
 	}
 #endif
 
Signature · 1 line
1.5.4
#8 lore raw

PATCH 7/11 stable_kernel_rules: fix must already be in mainline

In reply to Greg KH

From: Stefan Richter <[email protected]>

stable_kernel_rules: fix must already be in mainline

---
 Documentation/stable_kernel_rules.txt |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)
Documentation/stable_kernel_rules.txt +5 -2
index c815c52..4cfc788 100644
@@ -16,8 +16,9 @@ Rules on what kind of patches are accepted, and which ones are not, into the
    race can be exploited is also provided.
  - It cannot contain any "trivial" fixes in it (spelling changes,
    whitespace cleanups, etc).
- - It must be accepted by the relevant subsystem maintainer.
  - It must follow the Documentation/SubmittingPatches rules.
+ - It or an equivalent fix must already exist in Linus' tree.  Quote the
+   respective commit ID in Linus' tree in your patch submission to -stable.
 
 
 Procedure for submitting patches to the -stable tree:
@@ -28,7 +29,9 @@ Procedure for submitting patches to the -stable tree:
    queue, or a NAK if the patch is rejected.  This response might take a few
    days, according to the developer's schedules.
  - If accepted, the patch will be added to the -stable queue, for review by
-   other developers.
+   other developers and by the relevant subsystem maintainer.
+ - If the [email protected] address is added to a patch, when it goes into
+   Linus's tree it will automatically be emailed to the stable team.
  - Security patches should not be sent to this alias, but instead to the
    documented [email protected] address.
 
Signature · 1 line
1.5.4
#9 lore raw

PATCH 8/11 UIO: fix Greg's stupid changes

In reply to Greg KH

From: Brandon Philips <[email protected]>

This fixes two bugs with UIO that cropped up recently in -rc1

1) WARNING: at fs/sysfs/file.c:334 sysfs_open_file when trying to open a map addr/size file - complaining about missing sysfs_ops for ktype

2) Permission denied when reading uio/uio0/maps/map0/{addr,size} when files are mode S_IRUGO

Also fix a typo: attr_attribute -> addr_attribute

---
 drivers/uio/uio.c |   54 ++++++++++++++++++++++++++++++++++++----------------
 1 files changed, 37 insertions(+), 17 deletions(-)
drivers/uio/uio.c +37 -17
index 2a77e9d..e8a01f2 100644
@@ -57,29 +57,29 @@ struct uio_map {
 };
 #define to_map(map) container_of(map, struct uio_map, kobj)
 
-
-static ssize_t map_attr_show(struct kobject *kobj, struct kobj_attribute *attr,
-			     char *buf)
+static ssize_t map_addr_show(struct uio_mem *mem, char *buf)
 {
-	struct uio_map *map = to_map(kobj);
-	struct uio_mem *mem = map->mem;
-
-	if (strncmp(attr->attr.name, "addr", 4) == 0)
-		return sprintf(buf, "0x%lx\n", mem->addr);
-
-	if (strncmp(attr->attr.name, "size", 4) == 0)
-		return sprintf(buf, "0x%lx\n", mem->size);
+	return sprintf(buf, "0x%lx\n", mem->addr);
+}
 
-	return -ENODEV;
+static ssize_t map_size_show(struct uio_mem *mem, char *buf)
+{
+	return sprintf(buf, "0x%lx\n", mem->size);
 }
 
-static struct kobj_attribute attr_attribute =
-	__ATTR(addr, S_IRUGO, map_attr_show, NULL);
-static struct kobj_attribute size_attribute =
-	__ATTR(size, S_IRUGO, map_attr_show, NULL);
+struct uio_sysfs_entry {
+	struct attribute attr;
+	ssize_t (*show)(struct uio_mem *, char *);
+	ssize_t (*store)(struct uio_mem *, const char *, size_t);
+};
+
+static struct uio_sysfs_entry addr_attribute =
+	__ATTR(addr, S_IRUGO, map_addr_show, NULL);
+static struct uio_sysfs_entry size_attribute =
+	__ATTR(size, S_IRUGO, map_size_show, NULL);
 
 static struct attribute *attrs[] = {
-	&attr_attribute.attr,
+	&addr_attribute.attr,
 	&size_attribute.attr,
 	NULL,	/* need to NULL terminate the list of attributes */
 };
@@ -90,8 +90,28 @@ static void map_release(struct kobject *kobj)
 	kfree(map);
 }
 
+static ssize_t map_type_show(struct kobject *kobj, struct attribute *attr,
+			     char *buf)
+{
+	struct uio_map *map = to_map(kobj);
+	struct uio_mem *mem = map->mem;
+	struct uio_sysfs_entry *entry;
+
+	entry = container_of(attr, struct uio_sysfs_entry, attr);
+
+	if (!entry->show)
+		return -EIO;
+
+	return entry->show(mem, buf);
+}
+
+static struct sysfs_ops uio_sysfs_ops = {
+	.show = map_type_show,
+};
+
 static struct kobj_type map_attr_type = {
 	.release	= map_release,
+	.sysfs_ops	= &uio_sysfs_ops,
 	.default_attrs	= attrs,
 };
 
Signature · 1 line
1.5.4
#10 lore raw

PATCH 9/11 PM: Remove unbalanced mutex_unlock() from dpm_resume()

In reply to Greg KH

From: Rafael J. Wysocki <[email protected]>

Remove an unnecessary unlocking of dpm_list_mtx in the error path in drivers/base/power/main.c:dpm_suspend() .

---
 drivers/base/power/main.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)
drivers/base/power/main.c +0 -1
index d609560..cea5ed3 100644
@@ -479,7 +479,6 @@ static int dpm_suspend(pm_message_t state)
 			mutex_lock(&dpm_list_mtx);
 			if (list_empty(&dev->power.entry))
 				list_add(&dev->power.entry, &dpm_locked);
-			mutex_unlock(&dpm_list_mtx);
 			break;
 		}
 		mutex_lock(&dpm_list_mtx);
Signature · 1 line
1.5.4
#11 lore raw

PATCH 10/11 POWERPC: fix typo in pseries/power.c

In reply to Greg KH

From: Stephen Rothwell <[email protected]>

Introduced by commit 79393fc46ede43451a500a132e5de9856f5a4c83 ("kobject:
convert pseries/power.c to kobj_attr interface").

sys_create_file takes a "struct attrbute *" not a "struct kobj_addribute *".

arch/powerpc/platforms/pseries/power.c: In function 'apo_pm_init':
arch/powerpc/platforms/pseries/power.c:78: warning: passing argument 2 of 'sysfs_create_file' from incompatible pointer type

---
 arch/powerpc/platforms/pseries/power.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
arch/powerpc/platforms/pseries/power.c +1 -1
index e95fc15..6d62662 100644
@@ -75,7 +75,7 @@ core_initcall(pm_init);
 #else
 static int __init apo_pm_init(void)
 {
-	return (sysfs_create_file(power_kobj, &auto_poweron_attr));
+	return (sysfs_create_file(power_kobj, &auto_poweron_attr.attr));
 }
 __initcall(apo_pm_init);
 #endif
Signature · 1 line
1.5.4
#12 lore raw

PATCH 11/11 modules: do not try to add sysfs attributes if !CONFIG_SYSFS

In reply to Greg KH

From: Kay Sievers <[email protected]>

Thanks to Alexey for the testing and the fix of the fix.

---
 kernel/module.c |    7 ++-----
 1 files changed, 2 insertions(+), 5 deletions(-)
kernel/module.c +2 -5
index 92595ba..901cd6a 100644
@@ -987,12 +987,11 @@ static unsigned long resolve_symbol(Elf_Shdr *sechdrs,
 	return ret;
 }
 
-
 /*
  * /sys/module/foo/sections stuff
  * J. Corbet <[email protected]>
  */
-#ifdef CONFIG_KALLSYMS
+#if defined(CONFIG_KALLSYMS) && defined(CONFIG_SYSFS)
 static ssize_t module_sect_show(struct module_attribute *mattr,
 				struct module *mod, char *buf)
 {
@@ -1188,7 +1187,7 @@ static inline void add_notes_attrs(struct module *mod, unsigned int nsect,
 static inline void remove_notes_attrs(struct module *mod)
 {
 }
-#endif /* CONFIG_KALLSYMS */
+#endif
 
 #ifdef CONFIG_SYSFS
 int module_add_modinfo_attrs(struct module *mod)
@@ -1231,9 +1230,7 @@ void module_remove_modinfo_attrs(struct module *mod)
 	}
 	kfree(mod->modinfo_attrs);
 }
-#endif
 
-#ifdef CONFIG_SYSFS
 int mod_sysfs_init(struct module *mod)
 {
 	int err;
Signature · 1 line
1.5.4
#13 lore raw
In reply to Greg KH

Hi Greg.

Quoted 5 lines · Greg KH

Here are a few driver core fixes against your current git tree that fix some more problems that have cropped up:
- shutdown problem due to logic problem with cpufreq usage of kobjects.
- build fix for powerpc due to previous kobject changes.

Quoted 8 lines

- runtime errors when CONFIG_SYSFS=n
- UIO code now works properly from my previous messups
- proper encoding of the ja_JP stable_kernel_rules.txt file
- updates to the stable_kernel_rules.txt file
- mark ide=reverse as obsolete in preparation of 2.6.26 (Bart wanted this to go in through my tree as I have the .26 patches pending.)
- other minor fixes.

Do we have any outstanding issues with section mismatch warnings in PCI Core? I am as always optimistic and hope to get down to zero warnings soon.

Sam

author
#14 lore raw
Quoted 18 lines · Sam Ravnborg

Hi Greg.

Quoted 5 lines · Greg KH

Here are a few driver core fixes against your current git tree that fix some more problems that have cropped up:
- shutdown problem due to logic problem with cpufreq usage of kobjects.
- build fix for powerpc due to previous kobject changes.

Quoted 8 lines

- runtime errors when CONFIG_SYSFS=n
- UIO code now works properly from my previous messups
- proper encoding of the ja_JP stable_kernel_rules.txt file
- updates to the stable_kernel_rules.txt file
- mark ide=reverse as obsolete in preparation of 2.6.26 (Bart wanted this to go in through my tree as I have the .26 patches pending.)
- other minor fixes.

Quoted 2 lines

Do we have any outstanding issues with section mismatch warnings in PCI Core? I am as always optimistic and hope to get down to zero warnings soon.

I do not know, I saw your patches earlier and thought that they were going in through another tree for some reason. If you still see any remaining, please send me the needed patches again.

thanks,

greg k-h

#15 lore raw
Quoted 19 lines · Greg KH
Quoted 18 lines · Sam Ravnborg

Hi Greg.

Quoted 13 lines · Greg KH

Here are a few driver core fixes against your current git tree that fix some more problems that have cropped up:
- shutdown problem due to logic problem with cpufreq usage of kobjects.
- build fix for powerpc due to previous kobject changes.
- runtime errors when CONFIG_SYSFS=n
- UIO code now works properly from my previous messups
- proper encoding of the ja_JP stable_kernel_rules.txt file
- updates to the stable_kernel_rules.txt file
- mark ide=reverse as obsolete in preparation of 2.6.26 (Bart wanted this to go in through my tree as I have the .26 patches pending.)
- other minor fixes.

Quoted 6 lines
Quoted 2 lines

Do we have any outstanding issues with section mismatch warnings in PCI Core? I am as always optimistic and hope to get down to zero warnings soon.

I do not know, I saw your patches earlier and thought that they were going in through another tree for some reason. If you still see any remaining, please send me the needed patches again.

I looked and notice there were actually only one. Andrew picked it up but I have attached it here again. Then I leave the decision to you.

Sam

Quoted 1 line

From 426451c7c5d9447384b129ee3a750f66c0413a13 Mon Sep 17 00:00:00 2001

From: Sam Ravnborg <[email protected]>
Date: Sun, 17 Feb 2008 10:45:28 +0100
Subject: [PATCH 25/27] pci: fix section mismatch warning in pci_scan_child_bus

Fix following warning:
WARNING: vmlinux.o(.text+0x47bdb1): Section mismatch in reference from the function pci_scan_child_bus() to the function .devinit.text:pcibios_fixup_bus()

We had plenty of functions that could be annotated __devinit but due to the former restriction that annotated symbols could not be annotated they were not so.
So annotate these function and fix the references from the pci/hotplug/* code to silence the resuting warnings.

---
 drivers/pci/hotplug-pci.c              |    2 +-
 drivers/pci/hotplug/acpiphp_glue.c     |    2 +-
 drivers/pci/hotplug/cpci_hotplug_pci.c |    2 +-
 drivers/pci/hotplug/pciehp_pci.c       |    2 +-
 drivers/pci/hotplug/shpchp_pci.c       |    2 +-
 drivers/pci/probe.c                    |    8 ++++----
 6 files changed, 9 insertions(+), 9 deletions(-)
drivers/pci/hotplug-pci.c +1 -1
index a590ef6..4d4a644 100644
@@ -4,7 +4,7 @@
 #include "pci.h"
 
 
-unsigned int pci_do_scan_bus(struct pci_bus *bus)
+unsigned int __devinit pci_do_scan_bus(struct pci_bus *bus)
 {
 	unsigned int max;
 
drivers/pci/hotplug/acpiphp_glue.c +1 -1
index cf22f9e..5e50008 100644
@@ -1085,7 +1085,7 @@ static int acpiphp_bus_trim(acpi_handle handle)
  * This function should be called per *physical slot*,
  * not per each slot object in ACPI namespace.
  */
-static int enable_device(struct acpiphp_slot *slot)
+static int __ref enable_device(struct acpiphp_slot *slot)
 {
 	struct pci_dev *dev;
 	struct pci_bus *bus = slot->bridge->pci_bus;
drivers/pci/hotplug/cpci_hotplug_pci.c +1 -1
index 5e9be44..dddac96 100644
@@ -250,7 +250,7 @@ int cpci_led_off(struct slot* slot)
  * Device configuration functions
  */
 
-int cpci_configure_slot(struct slot* slot)
+int __ref cpci_configure_slot(struct slot* slot)
 {
 	struct pci_bus *parent;
 	int fn;
drivers/pci/hotplug/pciehp_pci.c +1 -1
index dd50713..9372a84 100644
@@ -167,7 +167,7 @@ static void program_fw_provided_values(struct pci_dev *dev)
 	}
 }
 
-static int pciehp_add_bridge(struct pci_dev *dev)
+static int __ref pciehp_add_bridge(struct pci_dev *dev)
 {
 	struct pci_bus *parent = dev->bus;
 	int pass, busnr, start = parent->secondary;
drivers/pci/hotplug/shpchp_pci.c +1 -1
index 0a6b25e..a69a215 100644
@@ -96,7 +96,7 @@ static void program_fw_provided_values(struct pci_dev *dev)
 	}
 }
 
-int shpchp_configure_device(struct slot *p_slot)
+int __ref shpchp_configure_device(struct slot *p_slot)
 {
 	struct pci_dev *dev;
 	struct pci_bus *parent = p_slot->ctrl->pci_dev->subordinate;
drivers/pci/probe.c +4 -4
index 4d23b9f..5ffd610 100644
@@ -286,7 +286,7 @@ static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
 	}
 }
 
-void pci_read_bridge_bases(struct pci_bus *child)
+void __devinit pci_read_bridge_bases(struct pci_bus *child)
 {
 	struct pci_dev *dev = child->self;
 	u8 io_base_lo, io_limit_lo;
@@ -472,7 +472,7 @@ static void pci_fixup_parent_subordinate_busnr(struct pci_bus *child, int max)
  * them, we proceed to assigning numbers to the remaining buses in
  * order to avoid overlaps between old and new bus numbers.
  */
-int pci_scan_bridge(struct pci_bus *bus, struct pci_dev * dev, int max, int pass)
+int __devinit pci_scan_bridge(struct pci_bus *bus, struct pci_dev * dev, int max, int pass)
 {
 	struct pci_bus *child;
 	int is_cardbus = (dev->hdr_type == PCI_HEADER_TYPE_CARDBUS);
@@ -1008,7 +1008,7 @@ int pci_scan_slot(struct pci_bus *bus, int devfn)
 	return nr;
 }
 
-unsigned int pci_scan_child_bus(struct pci_bus *bus)
+unsigned int __devinit pci_scan_child_bus(struct pci_bus *bus)
 {
 	unsigned int devfn, pass, max = bus->secondary;
 	struct pci_dev *dev;
@@ -1116,7 +1116,7 @@ err_out:
 	return NULL;
 }
 
-struct pci_bus *pci_scan_bus_parented(struct device *parent,
+struct pci_bus * __devinit pci_scan_bus_parented(struct device *parent,
 		int bus, struct pci_ops *ops, void *sysdata)
 {
 	struct pci_bus *b;
Signature · 1 line
1.5.4.rc3.14.g44397