Skip to contentSkip to conversation

gfs2: assign rgrp glock before compute_bitstructs

PATCH New Sasha Levin 1 participant 25 messages Apr 7, 2022

Subsystems (18)
lore
Share & tools
lore
author maintainer
#1 lore raw

PATCH 2/25 rtc: fix use-after-free on device removal

From: Vincent Whitchurch <[email protected]>

[ Upstream commit c8fa17d9f08a448184f03d352145099b5beb618e ]

If the irqwork is still scheduled or running while the RTC device is removed, a use-after-free occurs in rtc_timer_do_work(). Cleanup the timerqueue and ensure the work is stopped to fix this.

 BUG: KASAN: use-after-free in mutex_lock+0x94/0x110
 Write of size 8 at addr ffffff801d846338 by task kworker/3:1/41

 Workqueue: events rtc_timer_do_work
 Call trace:
  mutex_lock+0x94/0x110
  rtc_timer_do_work+0xec/0x630
  process_one_work+0x5fc/0x1344
  ...

 Allocated by task 551:
  kmem_cache_alloc_trace+0x384/0x6e0
  devm_rtc_allocate_device+0xf0/0x574
  devm_rtc_device_register+0x2c/0x12c
  ...

 Freed by task 572:
  kfree+0x114/0x4d0
  rtc_device_release+0x64/0x80
  device_release+0x8c/0x1f4
  kobject_put+0x1c4/0x4b0
  put_device+0x20/0x30
  devm_rtc_release_device+0x1c/0x30
  devm_action_release+0x54/0x90
  release_nodes+0x124/0x310
  devres_release_group+0x170/0x240
  i2c_device_remove+0xd8/0x314
  ...

 Last potentially related work creation:
  insert_work+0x5c/0x330
  queue_work_on+0xcc/0x154
  rtc_set_time+0x188/0x5bc
  rtc_dev_ioctl+0x2ac/0xbd0
  ...
---
 drivers/rtc/class.c | 9 +++++++++
 1 file changed, 9 insertions(+)
drivers/rtc/class.c +9 -0
index 7c88d190c51f..625effe6cb65 100644
@@ -26,6 +26,15 @@ struct class *rtc_class;
 static void rtc_device_release(struct device *dev)
 {
 	struct rtc_device *rtc = to_rtc_device(dev);
+	struct timerqueue_head *head = &rtc->timerqueue;
+	struct timerqueue_node *node;
+
+	mutex_lock(&rtc->ops_lock);
+	while ((node = timerqueue_getnext(head)))
+		timerqueue_del(head, node);
+	mutex_unlock(&rtc->ops_lock);
+
+	cancel_work_sync(&rtc->irqwork);
 
 	ida_simple_remove(&rtc_ida, rtc->id);
 	kfree(rtc);
Signature · 1 line
2.35.1
author maintainer
#2 lore raw

PATCH 3/25 rtc: pcf2127: fix bug when reading alarm registers

In reply to Sasha Levin

From: Hugo Villeneuve <[email protected]>

[ Upstream commit 73ce05302007eece23a6acb7dc124c92a2209087 ]

The first bug is that reading the 5 alarm registers results in a read operation of 20 bytes. The reason is because the destination buffer is defined as an array of "unsigned int", and we use the sizeof() operator on this array to define the bulk read count.

The second bug is that the read value is invalid, because we are indexing the destination buffer as integers (4 bytes), instead of indexing it as u8.

Changing the destination buffer type to u8 fixes both problems.

---
 drivers/rtc/rtc-pcf2127.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/rtc/rtc-pcf2127.c +2 -1
index f0a6861ff3ae..715513311ece 100644
@@ -366,7 +366,8 @@ static int pcf2127_watchdog_init(struct device *dev, struct pcf2127 *pcf2127)
 static int pcf2127_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
 {
 	struct pcf2127 *pcf2127 = dev_get_drvdata(dev);
-	unsigned int buf[5], ctrl2;
+	u8 buf[5];
+	unsigned int ctrl2;
 	int ret;
 
 	ret = regmap_read(pcf2127->regmap, PCF2127_REG_CTRL2, &ctrl2);
Signature · 1 line
2.35.1
author maintainer
#3 lore raw

PATCH 4/25 um: Cleanup syscall_handler_t definition/cast, fix warning

In reply to Sasha Levin

From: David Gow <[email protected]>

[ Upstream commit f4f03f299a56ce4d73c5431e0327b3b6cb55ebb9 ]

The syscall_handler_t type for x86_64 was defined as 'long (*)(void)', but always cast to 'long (*)(long, long, long, long, long, long)' before use. This now triggers a warning (see below).

Define syscall_handler_t as the latter instead, and remove the cast. This simplifies the code, and fixes the warning.

Warning:
In file included from ../arch/um/include/asm/processor-generic.h:13

                 from ../arch/x86/um/asm/processor.h:41
                 from ../include/linux/rcupdate.h:30
                 from ../include/linux/rculist.h:11
                 from ../include/linux/pid.h:5
                 from ../include/linux/sched.h:14
                 from ../include/linux/ptrace.h:6
                 from ../arch/um/kernel/skas/syscall.c:7:

../arch/um/kernel/skas/syscall.c: In function ‘handle_syscall’:
../arch/x86/um/shared/sysdep/syscalls_64.h:18:11: warning: cast between incompatible function types from ‘long int (*)(void)’ to ‘long int (*)(long int, long int, long int, long int, long int, long int)’ [ -Wcast-function-type]
18 | (((long (*)(long, long, long, long, long, long)) \
| ^
../arch/x86/um/asm/ptrace.h:36:62: note: in definition of macro ‘PT_REGS_SET_SYSCALL_RETURN’
36 | #define PT_REGS_SET_SYSCALL_RETURN(r, res) (PT_REGS_AX(r) = (res))
| ^~~
../arch/um/kernel/skas/syscall.c:46:33: note: in expansion of macro ‘EXECUTE_SYSCALL’

   46 |                                 EXECUTE_SYSCALL(syscall, regs));
      |                                 ^~~~~~~~~~~~~~~
---
 arch/x86/um/shared/sysdep/syscalls_64.h | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
arch/x86/um/shared/sysdep/syscalls_64.h +2 -3
index 8a7d5e1da98e..1e6875b4ffd8 100644
@@ -10,13 +10,12 @@
 #include <linux/msg.h>
 #include <linux/shm.h>
 
-typedef long syscall_handler_t(void);
+typedef long syscall_handler_t(long, long, long, long, long, long);
 
 extern syscall_handler_t *sys_call_table[];
 
 #define EXECUTE_SYSCALL(syscall, regs) \
-	(((long (*)(long, long, long, long, long, long)) \
-	  (*sys_call_table[syscall]))(UPT_SYSCALL_ARG1(&regs->regs), \
+	(((*sys_call_table[syscall]))(UPT_SYSCALL_ARG1(&regs->regs), \
 		 		      UPT_SYSCALL_ARG2(&regs->regs), \
 				      UPT_SYSCALL_ARG3(&regs->regs), \
 				      UPT_SYSCALL_ARG4(&regs->regs), \
Signature · 7 lines
2.35.1


_______________________________________________
linux-um mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-um
author maintainer
#4 lore raw

PATCH 5/25 um: port_user: Improve error handling when port-helper is not found

In reply to Sasha Levin

From: Glenn Washburn <[email protected]>

[ Upstream commit 3cb5a7f167c620a8b0e38b0446df2e024d2243dc ]

Check if port-helper exists and is executable. If not, write an error message to the kernel log with information to help the user diagnose the issue and exit with an error. If UML_PORT_HELPER was not set, write a message suggesting that the user set it. This makes it easier to understand why telneting to the UML instance is failing and what can be done to fix it.

---
 arch/um/drivers/port_user.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/um/drivers/port_user.c +12 -0
index 5b5b64cb1071..133ca7bf2d91 100644
@@ -5,6 +5,7 @@
 
 #include <stdio.h>
 #include <stdlib.h>
+#include <string.h>
 #include <errno.h>
 #include <termios.h>
 #include <unistd.h>
@@ -175,6 +176,17 @@ int port_connection(int fd, int *socket, int *pid_out)
 	if (new < 0)
 		return -errno;
 
+	err = os_access(argv[2], X_OK);
+	if (err < 0) {
+		printk(UM_KERN_ERR "port_connection : error accessing port-helper "
+		       "executable at %s: %s\n", argv[2], strerror(-err));
+		if (env == NULL)
+			printk(UM_KERN_ERR "Set UML_PORT_HELPER environment "
+				"variable to path to uml-utilities port-helper "
+				"binary\n");
+		goto out_close;
+	}
+
 	err = os_pipe(socket, 0, 0);
 	if (err < 0)
 		goto out_close;
Signature · 7 lines
2.35.1


_______________________________________________
linux-um mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-um
author maintainer
#5 lore raw

PATCH 6/25 Input: add bounds checking to input_set_capability()

In reply to Sasha Levin

From: Jeff LaBundy <[email protected]>

[ Upstream commit 409353cbe9fe48f6bc196114c442b1cff05a39bc ]

Update input_set_capability() to prevent kernel panic in case the event code exceeds the bitmap for the given event type.

---
 drivers/input/input.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
drivers/input/input.c +19 -0
index ff9dc37eff34..ce628a8446ea 100644
@@ -47,6 +47,17 @@ static DEFINE_MUTEX(input_mutex);
 
 static const struct input_value input_value_sync = { EV_SYN, SYN_REPORT, 1 };
 
+static const unsigned int input_max_code[EV_CNT] = {
+	[EV_KEY] = KEY_MAX,
+	[EV_REL] = REL_MAX,
+	[EV_ABS] = ABS_MAX,
+	[EV_MSC] = MSC_MAX,
+	[EV_SW] = SW_MAX,
+	[EV_LED] = LED_MAX,
+	[EV_SND] = SND_MAX,
+	[EV_FF] = FF_MAX,
+};
+
 static inline int is_event_supported(unsigned int code,
 				     unsigned long *bm, unsigned int max)
 {
@@ -1976,6 +1987,14 @@ EXPORT_SYMBOL(input_get_timestamp);
  */
 void input_set_capability(struct input_dev *dev, unsigned int type, unsigned int code)
 {
+	if (type < EV_CNT && input_max_code[type] &&
+	    code > input_max_code[type]) {
+		pr_err("%s: invalid code %u for type %u\n", __func__, code,
+		       type);
+		dump_stack();
+		return;
+	}
+
 	switch (type) {
 	case EV_KEY:
 		__set_bit(code, dev->keybit);
Signature · 1 line
2.35.1
author maintainer
#6 lore raw

PATCH 7/25 Input: stmfts - fix reference leak in stmfts_input_open

In reply to Sasha Levin

From: Zheng Yongjun <[email protected]>

[ Upstream commit 26623eea0da3476446909af96c980768df07bbd9 ]

pm_runtime_get_sync() will increment pm usage counter even it failed. Forgetting to call pm_runtime_put_noidle will result in reference leak in stmfts_input_open, so we should fix it.

---
 drivers/input/touchscreen/stmfts.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
drivers/input/touchscreen/stmfts.c +5 -3
index 9a64e1dbc04a..64b690a72d10 100644
@@ -339,11 +339,11 @@ static int stmfts_input_open(struct input_dev *dev)
 
 	err = pm_runtime_get_sync(&sdata->client->dev);
 	if (err < 0)
-		return err;
+		goto out;
 
 	err = i2c_smbus_write_byte(sdata->client, STMFTS_MS_MT_SENSE_ON);
 	if (err)
-		return err;
+		goto out;
 
 	mutex_lock(&sdata->mutex);
 	sdata->running = true;
@@ -366,7 +366,9 @@ static int stmfts_input_open(struct input_dev *dev)
 				 "failed to enable touchkey\n");
 	}
 
-	return 0;
+out:
+	pm_runtime_put_noidle(&sdata->client->dev);
+	return err;
 }
 
 static void stmfts_input_close(struct input_dev *dev)
Signature · 1 line
2.35.1
author maintainer
#7 lore raw

PATCH 8/25 nvme-pci: add quirks for Samsung X5 SSDs

In reply to Sasha Levin

From: Monish Kumar R <[email protected]>

[ Upstream commit bc360b0b1611566e1bd47384daf49af6a1c51837 ]

Add quirks to not fail the initialization and to have quick resume latency after cold/warm reboot.

---
 drivers/nvme/host/pci.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/nvme/host/pci.c +4 -1
index 97afeb898b25..ba067c1a6cf6 100644
@@ -3262,7 +3262,10 @@ static const struct pci_device_id nvme_id_table[] = {
 				NVME_QUIRK_128_BYTES_SQES |
 				NVME_QUIRK_SHARED_TAGS |
 				NVME_QUIRK_SKIP_CID_GEN },
-
+	{ PCI_DEVICE(0x144d, 0xa808),   /* Samsung X5 */
+		.driver_data =  NVME_QUIRK_DELAY_BEFORE_CHK_RDY|
+				NVME_QUIRK_NO_DEEPEST_PS |
+				NVME_QUIRK_IGNORE_DEV_SUBNQN, },
 	{ PCI_DEVICE_CLASS(PCI_CLASS_STORAGE_EXPRESS, 0xffffff) },
 	{ 0, }
 };
Signature · 1 line
2.35.1
author maintainer
#9 lore raw

PATCH 10/25 rtc: sun6i: Fix time overflow handling

In reply to Sasha Levin

From: Andre Przywara <[email protected]>

[ Upstream commit 9f6cd82eca7e91a0d0311242a87c6aa3c2737968 ]

Using "unsigned long" for UNIX timestamps is never a good idea, and comparing the value of such a variable against U32_MAX does not do anything useful on 32-bit systems.

Use the proper time64_t type when dealing with timestamps, and avoid cutting down the time range unnecessarily. This also fixes the flawed check for the alarm time being too far into the future.

The check for this condition is actually somewhat theoretical, as the RTC counts till 2033 only anyways, and 2^32 seconds from now is not before the year 2157 - at which point I hope nobody will be using this hardware anymore.

---
 drivers/rtc/rtc-sun6i.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)
drivers/rtc/rtc-sun6i.c +5 -9
index f2818cdd11d8..52b36b7c6129 100644
@@ -138,7 +138,7 @@ struct sun6i_rtc_dev {
 	const struct sun6i_rtc_clk_data *data;
 	void __iomem *base;
 	int irq;
-	unsigned long alarm;
+	time64_t alarm;
 
 	struct clk_hw hw;
 	struct clk_hw *int_osc;
@@ -510,10 +510,8 @@ static int sun6i_rtc_setalarm(struct device *dev, struct rtc_wkalrm *wkalrm)
 	struct sun6i_rtc_dev *chip = dev_get_drvdata(dev);
 	struct rtc_time *alrm_tm = &wkalrm->time;
 	struct rtc_time tm_now;
-	unsigned long time_now = 0;
-	unsigned long time_set = 0;
-	unsigned long time_gap = 0;
-	int ret = 0;
+	time64_t time_now, time_set;
+	int ret;
 
 	ret = sun6i_rtc_gettime(dev, &tm_now);
 	if (ret < 0) {
@@ -528,9 +526,7 @@ static int sun6i_rtc_setalarm(struct device *dev, struct rtc_wkalrm *wkalrm)
 		return -EINVAL;
 	}
 
-	time_gap = time_set - time_now;
-
-	if (time_gap > U32_MAX) {
+	if ((time_set - time_now) > U32_MAX) {
 		dev_err(dev, "Date too far in the future\n");
 		return -EINVAL;
 	}
@@ -539,7 +535,7 @@ static int sun6i_rtc_setalarm(struct device *dev, struct rtc_wkalrm *wkalrm)
 	writel(0, chip->base + SUN6I_ALRM_COUNTER);
 	usleep_range(100, 300);
 
-	writel(time_gap, chip->base + SUN6I_ALRM_COUNTER);
+	writel(time_set - time_now, chip->base + SUN6I_ALRM_COUNTER);
 	chip->alarm = time_set;
 
 	sun6i_rtc_setaie(wkalrm->enabled, chip);
Signature · 1 line
2.35.1
author maintainer
#10 lore raw

PATCH 11/25 crypto: stm32 - fix reference leak in stm32_crc_remove

In reply to Sasha Levin

From: Zheng Yongjun <[email protected]>

[ Upstream commit e9a36feecee0ee5845f2e0656f50f9942dd0bed3 ]

pm_runtime_get_sync() will increment pm usage counter even it failed. Forgetting to call pm_runtime_put_noidle will result in reference leak in stm32_crc_remove, so we should fix it.

---
 drivers/crypto/stm32/stm32-crc32.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/crypto/stm32/stm32-crc32.c +3 -1
index be1bf39a317d..90a920e7f664 100644
@@ -384,8 +384,10 @@ static int stm32_crc_remove(struct platform_device *pdev)
 	struct stm32_crc *crc = platform_get_drvdata(pdev);
 	int ret = pm_runtime_get_sync(crc->dev);
 
-	if (ret < 0)
+	if (ret < 0) {
+		pm_runtime_put_noidle(crc->dev);
 		return ret;
+	}
 
 	spin_lock(&crc_list.lock);
 	list_del(&crc->list);
Signature · 1 line
2.35.1
author maintainer
#11 lore raw

PATCH 12/25 crypto: x86/chacha20 - Avoid spurious jumps to other functions

In reply to Sasha Levin

From: Peter Zijlstra <[email protected]>

[ Upstream commit 4327d168515fd8b5b92fa1efdf1d219fb6514460 ]

The chacha_Nblock_xor_avx512vl() functions all have their own, identical, .LdoneN label, however in one particular spot {2,4} jump to the 8 version instead of their own. Resulting in:

arch/x86/crypto/chacha-x86_64.o: warning: objtool: chacha_2block_xor_avx512vl() falls through to next function chacha_8block_xor_avx512vl() arch/x86/crypto/chacha-x86_64.o: warning: objtool: chacha_4block_xor_avx512vl() falls through to next function chacha_8block_xor_avx512vl()

Make each function consistently use its own done label.

---
 arch/x86/crypto/chacha-avx512vl-x86_64.S | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/crypto/chacha-avx512vl-x86_64.S +2 -2
index bb193fde123a..8713c16c2501 100644
@@ -172,7 +172,7 @@ SYM_FUNC_START(chacha_2block_xor_avx512vl)
 	# xor remaining bytes from partial register into output
 	mov		%rcx,%rax
 	and		$0xf,%rcx
-	jz		.Ldone8
+	jz		.Ldone2
 	mov		%rax,%r9
 	and		$~0xf,%r9
 
@@ -438,7 +438,7 @@ SYM_FUNC_START(chacha_4block_xor_avx512vl)
 	# xor remaining bytes from partial register into output
 	mov		%rcx,%rax
 	and		$0xf,%rcx
-	jz		.Ldone8
+	jz		.Ldone4
 	mov		%rax,%r9
 	and		$~0xf,%r9
 
Signature · 1 line
2.35.1
author maintainer
#12 lore raw

PATCH 13/25 ALSA: hda/realtek: Enable headset mic on Lenovo P360

In reply to Sasha Levin

From: Kai-Heng Feng <[email protected]>

[ Upstream commit 5a8738571747c1e275a40b69a608657603867b7e ]

Lenovo P360 is another platform equipped with ALC897, and it needs ALC897_FIXUP_HEADSET_MIC_PIN quirk to make its headset mic work.

---
 sound/pci/hda/patch_realtek.c | 1 +
 1 file changed, 1 insertion(+)
sound/pci/hda/patch_realtek.c +1 -0
index 3bd37c02ce0e..91042c445c8c 100644
@@ -10861,6 +10861,7 @@ static const struct snd_pci_quirk alc662_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x144d, 0xc051, "Samsung R720", ALC662_FIXUP_IDEAPAD),
 	SND_PCI_QUIRK(0x14cd, 0x5003, "USI", ALC662_FIXUP_USI_HEADSET_MODE),
 	SND_PCI_QUIRK(0x17aa, 0x1036, "Lenovo P520", ALC662_FIXUP_LENOVO_MULTI_CODECS),
+	SND_PCI_QUIRK(0x17aa, 0x1057, "Lenovo P360", ALC897_FIXUP_HEADSET_MIC_PIN),
 	SND_PCI_QUIRK(0x17aa, 0x32ca, "Lenovo ThinkCentre M80", ALC897_FIXUP_HEADSET_MIC_PIN),
 	SND_PCI_QUIRK(0x17aa, 0x32cb, "Lenovo ThinkCentre M70", ALC897_FIXUP_HEADSET_MIC_PIN),
 	SND_PCI_QUIRK(0x17aa, 0x32cf, "Lenovo ThinkCentre M950", ALC897_FIXUP_HEADSET_MIC_PIN),
Signature · 1 line
2.35.1
author maintainer
#13 lore raw

PATCH 14/25 s390/pci: improve zpci_dev reference counting

In reply to Sasha Levin

From: Niklas Schnelle <[email protected]>

[ Upstream commit c122383d221dfa2f41cfe5e672540595de986fde ]

Currently zpci_dev uses kref based reference counting but only accounts for one original reference plus one reference from an added pci_dev to its underlying zpci_dev. Counting just the original reference worked until the pci_dev reference was added in commit 2a671f77ee49 ("s390/pci:
fix use after free of zpci_dev") because once a zpci_dev goes away, i.e. enters the reserved state, it would immediately get released. However with the pci_dev reference this is no longer the case and the zpci_dev may still appear in multiple availability events indicating that it was reserved. This was solved by detecting when the zpci_dev is already on its way out but still hanging around. This has however shown some light on how unusual our zpci_dev reference counting is.

Improve upon this by modelling zpci_dev reference counting on pci_dev. Analogous to pci_get_slot() increment the reference count in get_zdev_by_fid(). Thus all users of get_zdev_by_fid() must drop the reference once they are done with the zpci_dev.

Similar to pci_scan_single_device(), zpci_create_device() returns the device with an initial count of 1 and the device added to the zpci_list (analogous to the PCI bus' device_list). In turn users of zpci_create_device() must only drop the reference once the device is gone from the point of view of the zPCI subsystem, it might still be referenced by the common PCI subsystem though.

---
 arch/s390/pci/pci.c       | 1 +
 arch/s390/pci/pci_bus.h   | 3 ++-
 arch/s390/pci/pci_clp.c   | 9 +++++++--
 arch/s390/pci/pci_event.c | 7 ++++++-
 4 files changed, 16 insertions(+), 4 deletions(-)
arch/s390/pci/pci.c +1 -0
index e14e4a3a647a..74799439b259 100644
@@ -69,6 +69,7 @@ struct zpci_dev *get_zdev_by_fid(u32 fid)
 	list_for_each_entry(tmp, &zpci_list, entry) {
 		if (tmp->fid == fid) {
 			zdev = tmp;
+			zpci_zdev_get(zdev);
 			break;
 		}
 	}
arch/s390/pci/pci_bus.h +2 -1
index 55c9488e504c..8d2fcd091ca7 100644
@@ -13,7 +13,8 @@ void zpci_bus_device_unregister(struct zpci_dev *zdev);
 void zpci_release_device(struct kref *kref);
 static inline void zpci_zdev_put(struct zpci_dev *zdev)
 {
-	kref_put(&zdev->kref, zpci_release_device);
+	if (zdev)
+		kref_put(&zdev->kref, zpci_release_device);
 }
 
 static inline void zpci_zdev_get(struct zpci_dev *zdev)
arch/s390/pci/pci_clp.c +7 -2
index 0a0e8b8293be..d1a5c80a41cb 100644
@@ -22,6 +22,8 @@
 #include <asm/clp.h>
 #include <uapi/asm/clp.h>
 
+#include "pci_bus.h"
+
 bool zpci_unique_uid;
 
 void update_uid_checking(bool new)
@@ -372,8 +374,11 @@ static void __clp_add(struct clp_fh_list_entry *entry, void *data)
 		return;
 
 	zdev = get_zdev_by_fid(entry->fid);
-	if (!zdev)
-		zpci_create_device(entry->fid, entry->fh, entry->config_state);
+	if (zdev) {
+		zpci_zdev_put(zdev);
+		return;
+	}
+	zpci_create_device(entry->fid, entry->fh, entry->config_state);
 }
 
 int clp_scan_pci_devices(void)
arch/s390/pci/pci_event.c +6 -1
index b7cfde7e80a8..6ced44b5be8a 100644
@@ -61,10 +61,12 @@ static void __zpci_event_error(struct zpci_ccdf_err *ccdf)
 	       pdev ? pci_name(pdev) : "n/a", ccdf->pec, ccdf->fid);
 
 	if (!pdev)
-		return;
+		goto no_pdev;
 
 	pdev->error_state = pci_channel_io_perm_failure;
 	pci_dev_put(pdev);
+no_pdev:
+	zpci_zdev_put(zdev);
 }
 
 void zpci_event_error(void *data)
@@ -76,6 +78,7 @@ void zpci_event_error(void *data)
 static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 {
 	struct zpci_dev *zdev = get_zdev_by_fid(ccdf->fid);
+	bool existing_zdev = !!zdev;
 	enum zpci_state state;
 	struct pci_dev *pdev;
 	int ret;
@@ -161,6 +164,8 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 	default:
 		break;
 	}
+	if (existing_zdev)
+		zpci_zdev_put(zdev);
 }
 
 void zpci_event_availability(void *data)
Signature · 1 line
2.35.1
author maintainer
#14 lore raw

PATCH 15/25 vhost_vdpa: don't setup irq offloading when irq_num < 0

In reply to Sasha Levin

From: Zhu Lingshan <[email protected]>

[ Upstream commit cce0ab2b2a39072d81f98017f7b076f3410ef740 ]

When irq number is negative(e.g., -EINVAL), the virtqueue may be disabled or the virtqueues are sharing a device irq. In such case, we should not setup irq offloading for a virtqueue.

---
 drivers/vhost/vdpa.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/vhost/vdpa.c +4 -1
index e4d60009d908..04578aa87e4d 100644
@@ -97,8 +97,11 @@ static void vhost_vdpa_setup_vq_irq(struct vhost_vdpa *v, u16 qid)
 		return;
 
 	irq = ops->get_vq_irq(vdpa, qid);
+	if (irq < 0)
+		return;
+
 	irq_bypass_unregister_producer(&vq->call_ctx.producer);
-	if (!vq->call_ctx.ctx || irq < 0)
+	if (!vq->call_ctx.ctx)
 		return;
 
 	vq->call_ctx.producer.token = vq->call_ctx.ctx;
Signature · 6 lines
2.35.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
author maintainer
#15 lore raw

PATCH 16/25 tools/virtio: compile with -pthread

In reply to Sasha Levin

From: "Michael S. Tsirkin" <[email protected]>

[ Upstream commit f03560a57c1f60db6ac23ffd9714e1c69e2f95c7 ]

When using pthreads, one has to compile and link with -lpthread, otherwise e.g. glibc is not guaranteed to be reentrant.

This replaces -lpthread.

---
 tools/virtio/Makefile | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
tools/virtio/Makefile +2 -1
index 0d7bbe49359d..1b25cc7c64bb 100644
@@ -5,7 +5,8 @@ virtio_test: virtio_ring.o virtio_test.o
 vringh_test: vringh_test.o vringh.o virtio_ring.o
 
 CFLAGS += -g -O2 -Werror -Wno-maybe-uninitialized -Wall -I. -I../include/ -I ../../usr/include/ -Wno-pointer-sign -fno-strict-overflow -fno-strict-aliasing -fno-common -MMD -U_FORTIFY_SOURCE -include ../../include/linux/kconfig.h
-LDFLAGS += -lpthread
+CFLAGS += -pthread
+LDFLAGS += -pthread
 vpath %.c ../../drivers/virtio ../../drivers/vhost
 mod:
 	${MAKE} -C `pwd`/../.. M=`pwd`/vhost_test V=${V}
Signature · 6 lines
2.35.1

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
author maintainer
#16 lore raw

PATCH 17/25 nvme-multipath: fix hang when disk goes live over reconnect

In reply to Sasha Levin

From: Anton Eidelman <[email protected]>

[ Upstream commit a4a6f3c8f61c3cfbda4998ad94596059ad7e4332 ]

nvme_mpath_init_identify() invoked from nvme_init_identify() fetches a fresh ANA log from the ctrl. This is essential to have an up to date path states for both existing namespaces and for those scan_work may discover once the ctrl is up.

This happens in the following cases:
1) A new ctrl is being connected.
2) An existing ctrl is successfully reconnected.
3) An existing ctrl is being reset.

While in (1) ctrl->namespaces is empty, (2 & 3) may have namespaces, and nvme_read_ana_log() may call nvme_update_ns_ana_state().

This result in a hang when the ANA state of an existing namespace changes and makes the disk live: nvme_mpath_set_live() issues IO to the namespace through the ctrl, which does NOT have IO queues yet.

See sample hang below.

Solution:
- nvme_update_ns_ana_state() to call set_live only if ctrl is live
- nvme_read_ana_log() call from nvme_mpath_init_identify() therefore only fetches and parses the ANA log;
any erros in this process will fail the ctrl setup as appropriate;
- a separate function nvme_mpath_update()
is called in nvme_start_ctrl();
this parses the ANA log without fetching it.
At this point the ctrl is live,
therefore, disks can be set live normally.

Sample failure:

    nvme nvme0: starting error recovery
    nvme nvme0: Reconnecting in 10 seconds...
    block nvme0n6: no usable path - requeuing I/O
    INFO: task kworker/u8:3:312 blocked for more than 122 seconds.
          Tainted: G            E     5.14.5-1.el7.elrepo.x86_64 #1
    Workqueue: nvme-wq nvme_tcp_reconnect_ctrl_work [nvme_tcp]
    Call Trace:
     __schedule+0x2a2/0x7e0
     schedule+0x4e/0xb0
     io_schedule+0x16/0x40
     wait_on_page_bit_common+0x15c/0x3e0
     do_read_cache_page+0x1e0/0x410
     read_cache_page+0x12/0x20
     read_part_sector+0x46/0x100
     read_lba+0x121/0x240
     efi_partition+0x1d2/0x6a0
     bdev_disk_changed.part.0+0x1df/0x430
     bdev_disk_changed+0x18/0x20
     blkdev_get_whole+0x77/0xe0
     blkdev_get_by_dev+0xd2/0x3a0
     __device_add_disk+0x1ed/0x310
     device_add_disk+0x13/0x20
     nvme_mpath_set_live+0x138/0x1b0 [nvme_core]
     nvme_update_ns_ana_state+0x2b/0x30 [nvme_core]
     nvme_update_ana_state+0xca/0xe0 [nvme_core]
     nvme_parse_ana_log+0xac/0x170 [nvme_core]
     nvme_read_ana_log+0x7d/0xe0 [nvme_core]
     nvme_mpath_init_identify+0x105/0x150 [nvme_core]
     nvme_init_identify+0x2df/0x4d0 [nvme_core]
     nvme_init_ctrl_finish+0x8d/0x3b0 [nvme_core]
     nvme_tcp_setup_ctrl+0x337/0x390 [nvme_tcp]
     nvme_tcp_reconnect_ctrl_work+0x24/0x40 [nvme_tcp]
     process_one_work+0x1bd/0x360
     worker_thread+0x50/0x3d0
---
 drivers/nvme/host/core.c      |  1 +
 drivers/nvme/host/multipath.c | 25 +++++++++++++++++++++++--
 drivers/nvme/host/nvme.h      |  4 ++++
 3 files changed, 28 insertions(+), 2 deletions(-)
drivers/nvme/host/core.c +1 -0
index 71c85c99e86c..3a3c28c639df 100644
@@ -4409,6 +4409,7 @@ void nvme_start_ctrl(struct nvme_ctrl *ctrl)
 	if (ctrl->queue_count > 1) {
 		nvme_queue_scan(ctrl);
 		nvme_start_queues(ctrl);
+		nvme_mpath_update(ctrl);
 	}
 }
 EXPORT_SYMBOL_GPL(nvme_start_ctrl);
drivers/nvme/host/multipath.c +23 -2
index 18a756444d5a..a9e15c8f907b 100644
@@ -484,8 +484,17 @@ static void nvme_update_ns_ana_state(struct nvme_ana_group_desc *desc,
 	ns->ana_grpid = le32_to_cpu(desc->grpid);
 	ns->ana_state = desc->state;
 	clear_bit(NVME_NS_ANA_PENDING, &ns->flags);
-
-	if (nvme_state_is_live(ns->ana_state))
+	/*
+	 * nvme_mpath_set_live() will trigger I/O to the multipath path device
+	 * and in turn to this path device.  However we cannot accept this I/O
+	 * if the controller is not live.  This may deadlock if called from
+	 * nvme_mpath_init_identify() and the ctrl will never complete
+	 * initialization, preventing I/O from completing.  For this case we
+	 * will reprocess the ANA log page in nvme_mpath_update() once the
+	 * controller is ready.
+	 */
+	if (nvme_state_is_live(ns->ana_state) &&
+	    ns->ctrl->state == NVME_CTRL_LIVE)
 		nvme_mpath_set_live(ns);
 }
 
@@ -572,6 +581,18 @@ static void nvme_ana_work(struct work_struct *work)
 	nvme_read_ana_log(ctrl);
 }
 
+void nvme_mpath_update(struct nvme_ctrl *ctrl)
+{
+	u32 nr_change_groups = 0;
+
+	if (!ctrl->ana_log_buf)
+		return;
+
+	mutex_lock(&ctrl->ana_lock);
+	nvme_parse_ana_log(ctrl, &nr_change_groups, nvme_update_ana_state);
+	mutex_unlock(&ctrl->ana_lock);
+}
+
 static void nvme_anatt_timeout(struct timer_list *t)
 {
 	struct nvme_ctrl *ctrl = from_timer(ctrl, t, anatt_timer);
drivers/nvme/host/nvme.h +4 -0
index 5dd1dd8021ba..58e3bb672ce3 100644
@@ -707,6 +707,7 @@ void nvme_mpath_add_disk(struct nvme_ns *ns, struct nvme_id_ns *id);
 void nvme_mpath_remove_disk(struct nvme_ns_head *head);
 int nvme_mpath_init_identify(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id);
 void nvme_mpath_init_ctrl(struct nvme_ctrl *ctrl);
+void nvme_mpath_update(struct nvme_ctrl *ctrl);
 void nvme_mpath_uninit(struct nvme_ctrl *ctrl);
 void nvme_mpath_stop(struct nvme_ctrl *ctrl);
 bool nvme_mpath_clear_current_path(struct nvme_ns *ns);
@@ -793,6 +794,9 @@ static inline int nvme_mpath_init_identify(struct nvme_ctrl *ctrl,
 "Please enable CONFIG_NVME_MULTIPATH for full support of multi-port devices.\n");
 	return 0;
 }
+static inline void nvme_mpath_update(struct nvme_ctrl *ctrl)
+{
+}
 static inline void nvme_mpath_uninit(struct nvme_ctrl *ctrl)
 {
 }
Signature · 1 line
2.35.1
author maintainer
#17 lore raw

PATCH 18/25 rtc: mc146818-lib: Fix the AltCentury for AMD platforms

In reply to Sasha Levin

From: Mario Limonciello <[email protected]>

[ Upstream commit 3ae8fd41573af4fb3a490c9ed947fc936ba87190 ]

Setting the century forward has been failing on AMD platforms. There was a previous attempt at fixing this for family 0x17 as part of commit 7ad295d5196a ("rtc: Fix the AltCentury value on AMD/Hygon platform") but this was later reverted due to some problems reported that appeared to stem from an FW bug on a family 0x17 desktop system.

The same comments mentioned in the previous commit continue to apply to the newer platforms as well.

```
MC146818 driver use function mc146818_set_time() to set register
RTC_FREQ_SELECT(RTC_REG_A)'s bit4-bit6 field which means divider stage
reset value on Intel platform to 0x7.

While AMD/Hygon RTC_REG_A(0Ah)'s bit4 is defined as DV0 [Reference]:
DV0 = 0 selects Bank 0, DV0 = 1 selects Bank 1. Bit5-bit6 is defined
as reserved.

DV0 is set to 1, it will select Bank 1, which will disable AltCentury
register(0x32) access. As UEFI pass acpi_gbl_FADT.century 0x32
(AltCentury), the CMOS write will be failed on code:
CMOS_WRITE(century, acpi_gbl_FADT.century).

Correct RTC_REG_A bank select bit(DV0) to 0 on AMD/Hygon CPUs, it will
enable AltCentury(0x32) register writing and finally setup century as
expected.
```

However in closer examination the change previously submitted was also modifying bits 5 & 6 which are declared reserved in the AMD documentation. So instead modify just the DV0 bank selection bit.

Being cognizant that there was a failure reported before, split the code change out to a static function that can also be used for exclusions if any regressions such as Mikhail's pop up again.

---
 drivers/rtc/rtc-mc146818-lib.c | 16 +++++++++++++++-
 include/linux/mc146818rtc.h    |  2 ++
 2 files changed, 17 insertions(+), 1 deletion(-)
drivers/rtc/rtc-mc146818-lib.c +15 -1
index 5add637c9ad2..b036ff33fbe6 100644
@@ -99,6 +99,17 @@ unsigned int mc146818_get_time(struct rtc_time *time)
 }
 EXPORT_SYMBOL_GPL(mc146818_get_time);
 
+/* AMD systems don't allow access to AltCentury with DV1 */
+static bool apply_amd_register_a_behavior(void)
+{
+#ifdef CONFIG_X86
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD ||
+	    boot_cpu_data.x86_vendor == X86_VENDOR_HYGON)
+		return true;
+#endif
+	return false;
+}
+
 /* Set the current date and time in the real time clock. */
 int mc146818_set_time(struct rtc_time *time)
 {
@@ -172,7 +183,10 @@ int mc146818_set_time(struct rtc_time *time)
 	save_control = CMOS_READ(RTC_CONTROL);
 	CMOS_WRITE((save_control|RTC_SET), RTC_CONTROL);
 	save_freq_select = CMOS_READ(RTC_FREQ_SELECT);
-	CMOS_WRITE((save_freq_select|RTC_DIV_RESET2), RTC_FREQ_SELECT);
+	if (apply_amd_register_a_behavior())
+		CMOS_WRITE((save_freq_select & ~RTC_AMD_BANK_SELECT), RTC_FREQ_SELECT);
+	else
+		CMOS_WRITE((save_freq_select|RTC_DIV_RESET2), RTC_FREQ_SELECT);
 
 #ifdef CONFIG_MACH_DECSTATION
 	CMOS_WRITE(real_yrs, RTC_DEC_YEAR);
include/linux/mc146818rtc.h +2 -0
index 0661af17a758..1e0205811394 100644
@@ -86,6 +86,8 @@ struct cmos_rtc_board_info {
    /* 2 values for divider stage reset, others for "testing purposes only" */
 #  define RTC_DIV_RESET1	0x60
 #  define RTC_DIV_RESET2	0x70
+   /* In AMD BKDG bit 5 and 6 are reserved, bit 4 is for select dv0 bank */
+#  define RTC_AMD_BANK_SELECT	0x10
   /* Periodic intr. / Square wave rate select. 0=none, 1=32.8kHz,... 15=2Hz */
 # define RTC_RATE_SELECT 	0x0F
 
Signature · 1 line
2.35.1
author maintainer
#18 lore raw

PATCH 19/25 fs: fix an infinite loop in iomap_fiemap

In reply to Sasha Levin

From: Guo Xuenan <[email protected]>

[ Upstream commit 49df34221804cfd6384135b28b03c9461a31d024 ]

when get fiemap starting from MAX_LFS_FILESIZE, (maxbytes - *len) < start will always true , then *len set zero. because of start offset is beyond file size, for erofs filesystem it will always return iomap.length with zero,iomap iterate will enter infinite loop. it is necessary cover this corner case to avoid this situation.

------------[ cut here ]------------
WARNING: CPU: 7 PID: 905 at fs/iomap/iter.c:35 iomap_iter+0x97f/0xc70
Modules linked in: xfs erofs
CPU: 7 PID: 905 Comm: iomap Tainted: G        W         5.17.0-rc8 #27
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014
RIP: 0010:iomap_iter+0x97f/0xc70
Code: 85 a1 fc ff ff e8 71 be 9c ff 0f 1f 44 00 00 e9 92 fc ff ff e8 62 be 9c ff 0f 0b b8 fb ff ff ff e9 fc f8 ff ff e8 51 be 9c ff <0f> 0b e9 2b fc ff ff e8 45 be 9c ff 0f 0b e9 e1 fb ff ff e8 39 be
RSP: 0018:ffff888060a37ab0 EFLAGS: 00010293
RAX: 0000000000000000 RBX: ffff888060a37bb0 RCX: 0000000000000000
RDX: ffff88807e19a900 RSI: ffffffff81a7da7f RDI: ffff888060a37be0
RBP: 7fffffffffffffff R08: 0000000000000000 R09: ffff888060a37c20
R10: ffff888060a37c67 R11: ffffed100c146f8c R12: 7fffffffffffffff
R13: 0000000000000000 R14: ffff888060a37bd8 R15: ffff888060a37c20
FS:  00007fd3cca01540(0000) GS:ffff888108780000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000020010820 CR3: 0000000054b92000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 <TASK>
 iomap_fiemap+0x1c9/0x2f0
 erofs_fiemap+0x64/0x90 [erofs]
 do_vfs_ioctl+0x40d/0x12e0
 __x64_sys_ioctl+0xaa/0x1c0
 do_syscall_64+0x35/0x80
 entry_SYSCALL_64_after_hwframe+0x44/0xae
 </TASK>
---[ end trace 0000000000000000 ]---
watchdog: BUG: soft lockup - CPU#7 stuck for 26s! [iomap:905]
---
 fs/ioctl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ioctl.c +1 -1
index 4e6cc0a7d69c..7bcc60091287 100644
@@ -170,7 +170,7 @@ int fiemap_prep(struct inode *inode, struct fiemap_extent_info *fieinfo,
 
 	if (*len == 0)
 		return -EINVAL;
-	if (start > maxbytes)
+	if (start >= maxbytes)
 		return -EFBIG;
 
 	/*
Signature · 1 line
2.35.1
author maintainer
#19 lore raw

PATCH 20/25 MIPS: lantiq: check the return value of kzalloc()

In reply to Sasha Levin

From: Xiaoke Wang <[email protected]>

[ Upstream commit 34123208bbcc8c884a0489f543a23fe9eebb5514 ]

kzalloc() is a memory allocation function which can return NULL when some internal memory errors happen. So it is better to check the return value of it to prevent potential wrong memory access or memory leak.

---
 arch/mips/lantiq/falcon/sysctrl.c |  2 ++
 arch/mips/lantiq/xway/gptu.c      |  2 ++
 arch/mips/lantiq/xway/sysctrl.c   | 46 ++++++++++++++++++++-----------
 3 files changed, 34 insertions(+), 16 deletions(-)
arch/mips/lantiq/falcon/sysctrl.c +2 -0
index 42222f849bd2..446a2536999b 100644
@@ -167,6 +167,8 @@ static inline void clkdev_add_sys(const char *dev, unsigned int module,
 {
 	struct clk *clk = kzalloc(sizeof(struct clk), GFP_KERNEL);
 
+	if (!clk)
+		return;
 	clk->cl.dev_id = dev;
 	clk->cl.con_id = NULL;
 	clk->cl.clk = clk;
arch/mips/lantiq/xway/gptu.c +2 -0
index 3d5683e75cf1..200fe9ff641d 100644
@@ -122,6 +122,8 @@ static inline void clkdev_add_gptu(struct device *dev, const char *con,
 {
 	struct clk *clk = kzalloc(sizeof(struct clk), GFP_KERNEL);
 
+	if (!clk)
+		return;
 	clk->cl.dev_id = dev_name(dev);
 	clk->cl.con_id = con;
 	clk->cl.clk = clk;
arch/mips/lantiq/xway/sysctrl.c +30 -16

75 lines not on this page · raw message

Signature · 1 line
2.35.1
author maintainer
#20 lore raw

PATCH 21/25 drbd: remove usage of list iterator variable after loop

In reply to Sasha Levin

From: Jakob Koschel <[email protected]>

[ Upstream commit 901aeda62efa21f2eae937bccb71b49ae531be06 ]

In preparation to limit the scope of a list iterator to the list traversal loop, use a dedicated pointer to iterate through the list [1].

Since that variable should not be used past the loop iteration, a separate variable is used to 'remember the current location within the loop'.

To either continue iterating from that position or skip the iteration (if the previous iteration was complete) list_prepare_entry() is used.

---
 drivers/block/drbd/drbd_main.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/block/drbd/drbd_main.c +5 -2
index 65b95aef8dbc..3cdbd81f983f 100644
@@ -184,7 +184,7 @@ void tl_release(struct drbd_connection *connection, unsigned int barrier_nr,
 		unsigned int set_size)
 {
 	struct drbd_request *r;
-	struct drbd_request *req = NULL;
+	struct drbd_request *req = NULL, *tmp = NULL;
 	int expect_epoch = 0;
 	int expect_size = 0;
 
@@ -238,8 +238,11 @@ void tl_release(struct drbd_connection *connection, unsigned int barrier_nr,
 	 * to catch requests being barrier-acked "unexpectedly".
 	 * It usually should find the same req again, or some READ preceding it. */
 	list_for_each_entry(req, &connection->transfer_log, tl_requests)
-		if (req->epoch == expect_epoch)
+		if (req->epoch == expect_epoch) {
+			tmp = req;
 			break;
+		}
+	req = list_prepare_entry(tmp, &connection->transfer_log, tl_requests);
 	list_for_each_entry_safe_from(req, r, &connection->transfer_log, tl_requests) {
 		if (req->epoch != expect_epoch)
 			break;
Signature · 1 line
2.35.1
author maintainer
#21 lore raw

PATCH 22/25 platform/chrome: cros_ec_debugfs: detach log reader wq from devm

In reply to Sasha Levin

From: Tzung-Bi Shih <[email protected]>

[ Upstream commit 0e8eb5e8acbad19ac2e1856b2fb2320184299b33 ]

Debugfs console_log uses devm memory (e.g. debug_info in cros_ec_console_log_poll()). However, lifecycles of device and debugfs are independent. An use-after-free issue is observed if userland program operates the debugfs after the memory has been freed.

The call trace:
do_raw_spin_lock
_raw_spin_lock_irqsave
remove_wait_queue
ep_unregister_pollwait
ep_remove
do_epoll_ctl

A Python example to reproduce the issue:
... import select
... p = select.epoll()
... f = open('/sys/kernel/debug/cros_scp/console_log') ... p.register(f, select.POLLIN)
... p.poll(1)
[(4, 1)] # 4=fd, 1=select.POLLIN

[ shutdown cros_scp at the point ]

... p.poll(1)
[(4, 16)] # 4=fd, 16=select.POLLHUP
... p.unregister(f)

An use-after-free issue raises here. It called epoll_ctl with EPOLL_CTL_DEL which in turn to use the workqueue in the devm (i.e. log_wq).

Detaches log reader's workqueue from devm to make sure it is persistent even if the device has been removed.

---
 drivers/platform/chrome/cros_ec_debugfs.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/platform/chrome/cros_ec_debugfs.c +6 -6
index 272c89837d74..0dbceee87a4b 100644
@@ -25,6 +25,9 @@
 
 #define CIRC_ADD(idx, size, value)	(((idx) + (value)) & ((size) - 1))
 
+/* waitqueue for log readers */
+static DECLARE_WAIT_QUEUE_HEAD(cros_ec_debugfs_log_wq);
+
 /**
  * struct cros_ec_debugfs - EC debugging information.
  *
@@ -33,7 +36,6 @@
  * @log_buffer: circular buffer for console log information
  * @read_msg: preallocated EC command and buffer to read console log
  * @log_mutex: mutex to protect circular buffer
- * @log_wq: waitqueue for log readers
  * @log_poll_work: recurring task to poll EC for new console log data
  * @panicinfo_blob: panicinfo debugfs blob
  */
@@ -44,7 +46,6 @@ struct cros_ec_debugfs {
 	struct circ_buf log_buffer;
 	struct cros_ec_command *read_msg;
 	struct mutex log_mutex;
-	wait_queue_head_t log_wq;
 	struct delayed_work log_poll_work;
 	/* EC panicinfo */
 	struct debugfs_blob_wrapper panicinfo_blob;
@@ -107,7 +108,7 @@ static void cros_ec_console_log_work(struct work_struct *__work)
 			buf_space--;
 		}
 
-		wake_up(&debug_info->log_wq);
+		wake_up(&cros_ec_debugfs_log_wq);
 	}
 
 	mutex_unlock(&debug_info->log_mutex);
@@ -141,7 +142,7 @@ static ssize_t cros_ec_console_log_read(struct file *file, char __user *buf,
 
 		mutex_unlock(&debug_info->log_mutex);
 
-		ret = wait_event_interruptible(debug_info->log_wq,
+		ret = wait_event_interruptible(cros_ec_debugfs_log_wq,
 					CIRC_CNT(cb->head, cb->tail, LOG_SIZE));
 		if (ret < 0)
 			return ret;
@@ -173,7 +174,7 @@ static __poll_t cros_ec_console_log_poll(struct file *file,
 	struct cros_ec_debugfs *debug_info = file->private_data;
 	__poll_t mask = 0;
 
-	poll_wait(file, &debug_info->log_wq, wait);
+	poll_wait(file, &cros_ec_debugfs_log_wq, wait);
 
 	mutex_lock(&debug_info->log_mutex);
 	if (CIRC_CNT(debug_info->log_buffer.head,
@@ -377,7 +378,6 @@ static int cros_ec_create_console_log(struct cros_ec_debugfs *debug_info)
 	debug_info->log_buffer.tail = 0;
 
 	mutex_init(&debug_info->log_mutex);
-	init_waitqueue_head(&debug_info->log_wq);
 
 	debugfs_create_file("console_log", S_IFREG | 0444, debug_info->dir,
 			    debug_info, &cros_ec_console_log_fops);
Signature · 1 line
2.35.1
author maintainer
#22 lore raw

PATCH 23/25 ARM: 9191/1: arm/stacktrace, kasan: Silence KASAN warnings in unwind_frame()

In reply to Sasha Levin

From: linyujun <[email protected]>

[ Upstream commit 9be4c88bb7924f68f88cfd47d925c2d046f51a73 ]

The following KASAN warning is detected by QEMU.

==================================================================

BUG: KASAN: stack-out-of-bounds in unwind_frame+0x508/0x870
Read of size 4 at addr c36bba90 by task cat/163

CPU: 1 PID: 163 Comm: cat Not tainted 5.10.0-rc1 #40
Hardware name: ARM-Versatile Express
[<c0113fac>] (unwind_backtrace) from [<c010e71c>] (show_stack+0x10/0x14)
[<c010e71c>] (show_stack) from [<c0b805b4>] (dump_stack+0x98/0xb0)
[<c0b805b4>] (dump_stack) from [<c0b7d658>] (print_address_description.constprop.0+0x58/0x4bc)
[<c0b7d658>] (print_address_description.constprop.0) from [<c031435c>] (kasan_report+0x154/0x170)
[<c031435c>] (kasan_report) from [<c0113c44>] (unwind_frame+0x508/0x870)
[<c0113c44>] (unwind_frame) from [<c010e298>] (__save_stack_trace+0x110/0x134)
[<c010e298>] (__save_stack_trace) from [<c01ce0d8>] (stack_trace_save+0x8c/0xb4)
[<c01ce0d8>] (stack_trace_save) from [<c0313520>] (kasan_set_track+0x38/0x60)
[<c0313520>] (kasan_set_track) from [<c0314cb8>] (kasan_set_free_info+0x20/0x2c)
[<c0314cb8>] (kasan_set_free_info) from [<c0313474>] (__kasan_slab_free+0xec/0x120)
[<c0313474>] (__kasan_slab_free) from [<c0311e20>] (kmem_cache_free+0x7c/0x334)
[<c0311e20>] (kmem_cache_free) from [<c01c35dc>] (rcu_core+0x390/0xccc)
[<c01c35dc>] (rcu_core) from [<c01013a8>] (__do_softirq+0x180/0x518)
[<c01013a8>] (__do_softirq) from [<c0135214>] (irq_exit+0x9c/0xe0)
[<c0135214>] (irq_exit) from [<c01a40e4>] (__handle_domain_irq+0xb0/0x110)
[<c01a40e4>] (__handle_domain_irq) from [<c0691248>] (gic_handle_irq+0xa0/0xb8)
[<c0691248>] (gic_handle_irq) from [<c0100b0c>] (__irq_svc+0x6c/0x94)
Exception stack(0xc36bb928 to 0xc36bb970)

b920: c36bb9c0 00000000 c0126919 c0101228 c36bb9c0 b76d7730
b940: c36b8000 c36bb9a0 c3335b00 c01ce0d8 00000003 c36bba3c c36bb940 c36bb978
b960: c010e298 c011373c 60000013 ffffffff
[<c0100b0c>] (__irq_svc) from [<c011373c>] (unwind_frame+0x0/0x870) [<c011373c>] (unwind_frame) from [<00000000>] (0x0)

The buggy address belongs to the page:
page:(ptrval) refcount:0 mapcount:0 mapping:00000000 index:0x0 pfn:0x636bb
flags: 0x0()
raw: 00000000 00000000 ef867764 00000000 00000000 00000000 ffffffff 00000000
page dumped because: kasan: bad access detected

addr c36bba90 is located in stack of task cat/163 at offset 48 in frame:
stack_trace_save+0x0/0xb4

this frame has 1 object:
[32, 48) 'trace'

Memory state around the buggy address:
 c36bb980: f1 f1 f1 f1 00 04 f2 f2 00 00 f3 f3 00 00 00 00
 c36bba00: 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1
>c36bba80: 00 00 f3 f3 00 00 00 00 00 00 00 00 00 00 00 00
                 ^
 c36bbb00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
 c36bbb80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
==================================================================

There is a same issue on x86 and has been resolved by the commit f7d27c35ddff ("x86/mm, kasan: Silence KASAN warnings in get_wchan()").
The solution could be applied to arm architecture too.

---
 arch/arm/kernel/stacktrace.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
arch/arm/kernel/stacktrace.c +5 -5
index db798eac7431..824774999825 100644
@@ -53,17 +53,17 @@ int notrace unwind_frame(struct stackframe *frame)
 		return -EINVAL;
 
 	frame->sp = frame->fp;
-	frame->fp = *(unsigned long *)(fp);
-	frame->pc = *(unsigned long *)(fp + 4);
+	frame->fp = READ_ONCE_NOCHECK(*(unsigned long *)(fp));
+	frame->pc = READ_ONCE_NOCHECK(*(unsigned long *)(fp + 4));
 #else
 	/* check current frame pointer is within bounds */
 	if (fp < low + 12 || fp > high - 4)
 		return -EINVAL;
 
 	/* restore the registers from the stack frame */
-	frame->fp = *(unsigned long *)(fp - 12);
-	frame->sp = *(unsigned long *)(fp - 8);
-	frame->pc = *(unsigned long *)(fp - 4);
+	frame->fp = READ_ONCE_NOCHECK(*(unsigned long *)(fp - 12));
+	frame->sp = READ_ONCE_NOCHECK(*(unsigned long *)(fp - 8));
+	frame->pc = READ_ONCE_NOCHECK(*(unsigned long *)(fp - 4));
 #endif
 
 	return 0;
Signature · 1 line
2.35.1
author maintainer
#23 lore raw

PATCH 24/25 nilfs2: fix lockdep warnings in page operations for btree nodes

In reply to Sasha Levin

From: Ryusuke Konishi <[email protected]>

[ Upstream commit e897be17a441fa637cd166fc3de1445131e57692 ]

Patch series "nilfs2 lockdep warning fixes".

The first two are to resolve the lockdep warning issue, and the last one is the accompanying cleanup and low priority.

Based on your comment, this series solves the issue by separating inode object as needed. Since I was worried about the impact of the object composition changes, I tested the series carefully not to cause regressions especially for delicate functions such like disk space reclamation and snapshots.

This patch (of 3):

If CONFIG_LOCKDEP is enabled, nilfs2 hits lockdep warnings at inode_to_wb() during page/folio operations for btree nodes:

  WARNING: CPU: 0 PID: 6575 at include/linux/backing-dev.h:269 inode_to_wb include/linux/backing-dev.h:269 [inline]
  WARNING: CPU: 0 PID: 6575 at include/linux/backing-dev.h:269 folio_account_dirtied mm/page-writeback.c:2460 [inline]
  WARNING: CPU: 0 PID: 6575 at include/linux/backing-dev.h:269 __folio_mark_dirty+0xa7c/0xe30 mm/page-writeback.c:2509
  Modules linked in:
  ...
  RIP: 0010:inode_to_wb include/linux/backing-dev.h:269 [inline]
  RIP: 0010:folio_account_dirtied mm/page-writeback.c:2460 [inline]
  RIP: 0010:__folio_mark_dirty+0xa7c/0xe30 mm/page-writeback.c:2509
  ...
  Call Trace:
    __set_page_dirty include/linux/pagemap.h:834 [inline]
    mark_buffer_dirty+0x4e6/0x650 fs/buffer.c:1145
    nilfs_btree_propagate_p fs/nilfs2/btree.c:1889 [inline]
    nilfs_btree_propagate+0x4ae/0xea0 fs/nilfs2/btree.c:2085
    nilfs_bmap_propagate+0x73/0x170 fs/nilfs2/bmap.c:337
    nilfs_collect_dat_data+0x45/0xd0 fs/nilfs2/segment.c:625
    nilfs_segctor_apply_buffers+0x14a/0x470 fs/nilfs2/segment.c:1009
    nilfs_segctor_scan_file+0x47a/0x700 fs/nilfs2/segment.c:1048
    nilfs_segctor_collect_blocks fs/nilfs2/segment.c:1224 [inline]
    nilfs_segctor_collect fs/nilfs2/segment.c:1494 [inline]
    nilfs_segctor_do_construct+0x14f3/0x6c60 fs/nilfs2/segment.c:2036
    nilfs_segctor_construct+0x7a7/0xb30 fs/nilfs2/segment.c:2372
    nilfs_segctor_thread_construct fs/nilfs2/segment.c:2480 [inline]
    nilfs_segctor_thread+0x3c3/0xf90 fs/nilfs2/segment.c:2563
    kthread+0x405/0x4f0 kernel/kthread.c:327
    ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:295

This is because nilfs2 uses two page caches for each inode and inode->i_mapping never points to one of them, the btree node cache.

This causes inode_to_wb(inode) to refer to a different page cache than the caller page/folio operations such like __folio_start_writeback(), __folio_end_writeback(), or __folio_mark_dirty() acquired the lock.

This patch resolves the issue by allocating and using an additional inode to hold the page cache of btree nodes. The inode is attached one-to-one to the traditional nilfs2 inode if it requires a block mapping with b-tree. This setup change is in memory only and does not affect the disk format.

---
 fs/nilfs2/btnode.c  |  23 ++++++++--
 fs/nilfs2/btnode.h  |   1 +
 fs/nilfs2/btree.c   |  27 ++++++++----
 fs/nilfs2/gcinode.c |   7 +--
 fs/nilfs2/inode.c   | 104 ++++++++++++++++++++++++++++++++++++++------
 fs/nilfs2/mdt.c     |   7 +--
 fs/nilfs2/nilfs.h   |  14 +++---
 fs/nilfs2/page.c    |   7 ++-
 fs/nilfs2/segment.c |   9 ++--
 fs/nilfs2/super.c   |   5 +--
 10 files changed, 154 insertions(+), 50 deletions(-)
fs/nilfs2/btnode.c +20 -3
index 4391fd3abd8f..e00e184b1261 100644
@@ -20,6 +20,23 @@
 #include "page.h"
 #include "btnode.h"
 
+
+/**
+ * nilfs_init_btnc_inode - initialize B-tree node cache inode
+ * @btnc_inode: inode to be initialized
+ *
+ * nilfs_init_btnc_inode() sets up an inode for B-tree node cache.
+ */
+void nilfs_init_btnc_inode(struct inode *btnc_inode)
+{
+	struct nilfs_inode_info *ii = NILFS_I(btnc_inode);
+
+	btnc_inode->i_mode = S_IFREG;
+	ii->i_flags = 0;
+	memset(&ii->i_bmap_data, 0, sizeof(struct nilfs_bmap));
+	mapping_set_gfp_mask(btnc_inode->i_mapping, GFP_NOFS);
+}
+
 void nilfs_btnode_cache_clear(struct address_space *btnc)
 {
 	invalidate_mapping_pages(btnc, 0, -1);
@@ -29,7 +46,7 @@ void nilfs_btnode_cache_clear(struct address_space *btnc)
 struct buffer_head *
 nilfs_btnode_create_block(struct address_space *btnc, __u64 blocknr)
 {
-	struct inode *inode = NILFS_BTNC_I(btnc);
+	struct inode *inode = btnc->host;
 	struct buffer_head *bh;
 
 	bh = nilfs_grab_buffer(inode, btnc, blocknr, BIT(BH_NILFS_Node));
@@ -57,7 +74,7 @@ int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr,
 			      struct buffer_head **pbh, sector_t *submit_ptr)
 {
 	struct buffer_head *bh;
-	struct inode *inode = NILFS_BTNC_I(btnc);
+	struct inode *inode = btnc->host;
 	struct page *page;
 	int err;
 
@@ -157,7 +174,7 @@ int nilfs_btnode_prepare_change_key(struct address_space *btnc,
 				    struct nilfs_btnode_chkey_ctxt *ctxt)
 {
 	struct buffer_head *obh, *nbh;
-	struct inode *inode = NILFS_BTNC_I(btnc);
+	struct inode *inode = btnc->host;
 	__u64 oldkey = ctxt->oldkey, newkey = ctxt->newkey;
 	int err;
 
fs/nilfs2/btnode.h +1 -0
index 0f88dbc9bcb3..05ab64d354dc 100644
@@ -30,6 +30,7 @@ struct nilfs_btnode_chkey_ctxt {
 	struct buffer_head *newbh;
 };
 
+void nilfs_init_btnc_inode(struct inode *btnc_inode);
 void nilfs_btnode_cache_clear(struct address_space *);
 struct buffer_head *nilfs_btnode_create_block(struct address_space *btnc,
 					      __u64 blocknr);
fs/nilfs2/btree.c +19 -8
index f42ab57201e7..77efd69213a3 100644
@@ -58,7 +58,8 @@ static void nilfs_btree_free_path(struct nilfs_btree_path *path)
 static int nilfs_btree_get_new_block(const struct nilfs_bmap *btree,
 				     __u64 ptr, struct buffer_head **bhp)
 {
-	struct address_space *btnc = &NILFS_BMAP_I(btree)->i_btnode_cache;
+	struct inode *btnc_inode = NILFS_BMAP_I(btree)->i_assoc_inode;
+	struct address_space *btnc = btnc_inode->i_mapping;
 	struct buffer_head *bh;
 
 	bh = nilfs_btnode_create_block(btnc, ptr);
@@ -470,7 +471,8 @@ static int __nilfs_btree_get_block(const struct nilfs_bmap *btree, __u64 ptr,
 				   struct buffer_head **bhp,
 				   const struct nilfs_btree_readahead_info *ra)
 {
-	struct address_space *btnc = &NILFS_BMAP_I(btree)->i_btnode_cache;
+	struct inode *btnc_inode = NILFS_BMAP_I(btree)->i_assoc_inode;
+	struct address_space *btnc = btnc_inode->i_mapping;
 	struct buffer_head *bh, *ra_bh;
 	sector_t submit_ptr = 0;
 	int ret;
@@ -1742,6 +1744,10 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *btree, __u64 key,
 		dat = nilfs_bmap_get_dat(btree);
 	}
 
+	ret = nilfs_attach_btree_node_cache(&NILFS_BMAP_I(btree)->vfs_inode);
+	if (ret < 0)
+		return ret;
+
 	ret = nilfs_bmap_prepare_alloc_ptr(btree, dreq, dat);
 	if (ret < 0)
 		return ret;
@@ -1914,7 +1920,7 @@ static int nilfs_btree_prepare_update_v(struct nilfs_bmap *btree,
 		path[level].bp_ctxt.newkey = path[level].bp_newreq.bpr_ptr;
 		path[level].bp_ctxt.bh = path[level].bp_bh;
 		ret = nilfs_btnode_prepare_change_key(
-			&NILFS_BMAP_I(btree)->i_btnode_cache,
+			NILFS_BMAP_I(btree)->i_assoc_inode->i_mapping,
 			&path[level].bp_ctxt);
 		if (ret < 0) {
 			nilfs_dat_abort_update(dat,
@@ -1940,7 +1946,7 @@ static void nilfs_btree_commit_update_v(struct nilfs_bmap *btree,
 
 	if (buffer_nilfs_node(path[level].bp_bh)) {
 		nilfs_btnode_commit_change_key(
-			&NILFS_BMAP_I(btree)->i_btnode_cache,
+			NILFS_BMAP_I(btree)->i_assoc_inode->i_mapping,
 			&path[level].bp_ctxt);
 		path[level].bp_bh = path[level].bp_ctxt.bh;
 	}
@@ -1959,7 +1965,7 @@ static void nilfs_btree_abort_update_v(struct nilfs_bmap *btree,
 			       &path[level].bp_newreq.bpr_req);
 	if (buffer_nilfs_node(path[level].bp_bh))
 		nilfs_btnode_abort_change_key(
-			&NILFS_BMAP_I(btree)->i_btnode_cache,
+			NILFS_BMAP_I(btree)->i_assoc_inode->i_mapping,
 			&path[level].bp_ctxt);
 }
 
@@ -2135,7 +2141,8 @@ static void nilfs_btree_add_dirty_buffer(struct nilfs_bmap *btree,
 static void nilfs_btree_lookup_dirty_buffers(struct nilfs_bmap *btree,
 					     struct list_head *listp)
 {
-	struct address_space *btcache = &NILFS_BMAP_I(btree)->i_btnode_cache;
+	struct inode *btnc_inode = NILFS_BMAP_I(btree)->i_assoc_inode;
+	struct address_space *btcache = btnc_inode->i_mapping;
 	struct list_head lists[NILFS_BTREE_LEVEL_MAX];
 	struct pagevec pvec;
 	struct buffer_head *bh, *head;
@@ -2189,12 +2196,12 @@ static int nilfs_btree_assign_p(struct nilfs_bmap *btree,
 		path[level].bp_ctxt.newkey = blocknr;
 		path[level].bp_ctxt.bh = *bh;
 		ret = nilfs_btnode_prepare_change_key(
-			&NILFS_BMAP_I(btree)->i_btnode_cache,
+			NILFS_BMAP_I(btree)->i_assoc_inode->i_mapping,
 			&path[level].bp_ctxt);
 		if (ret < 0)
 			return ret;
 		nilfs_btnode_commit_change_key(
-			&NILFS_BMAP_I(btree)->i_btnode_cache,
+			NILFS_BMAP_I(btree)->i_assoc_inode->i_mapping,
 			&path[level].bp_ctxt);
 		*bh = path[level].bp_ctxt.bh;
 	}
@@ -2399,6 +2406,10 @@ int nilfs_btree_init(struct nilfs_bmap *bmap)
 
 	if (nilfs_btree_root_broken(nilfs_btree_get_root(bmap), bmap->b_inode))
 		ret = -EIO;
+	else
+		ret = nilfs_attach_btree_node_cache(
+			&NILFS_BMAP_I(bmap)->vfs_inode);
+
 	return ret;
 }
 
fs/nilfs2/gcinode.c +4 -3
index 448320496856..aadea660c66c 100644
@@ -126,9 +126,10 @@ int nilfs_gccache_submit_read_data(struct inode *inode, sector_t blkoff,
 int nilfs_gccache_submit_read_node(struct inode *inode, sector_t pbn,
 				   __u64 vbn, struct buffer_head **out_bh)
 {
+	struct inode *btnc_inode = NILFS_I(inode)->i_assoc_inode;
 	int ret;
 
-	ret = nilfs_btnode_submit_block(&NILFS_I(inode)->i_btnode_cache,
+	ret = nilfs_btnode_submit_block(btnc_inode->i_mapping,
 					vbn ? : pbn, pbn, REQ_OP_READ, 0,
 					out_bh, &pbn);
 	if (ret == -EEXIST) /* internal code (cache hit) */
@@ -170,7 +171,7 @@ int nilfs_init_gcinode(struct inode *inode)
 	ii->i_flags = 0;
 	nilfs_bmap_init_gc(ii->i_bmap);
 
-	return 0;
+	return nilfs_attach_btree_node_cache(inode);
 }
 
 /**
@@ -185,7 +186,7 @@ void nilfs_remove_all_gcinodes(struct the_nilfs *nilfs)
 		ii = list_first_entry(head, struct nilfs_inode_info, i_dirty);
 		list_del_init(&ii->i_dirty);
 		truncate_inode_pages(&ii->vfs_inode.i_data, 0);
-		nilfs_btnode_cache_clear(&ii->i_btnode_cache);
+		nilfs_btnode_cache_clear(ii->i_assoc_inode->i_mapping);
 		iput(&ii->vfs_inode);
 	}
 }
fs/nilfs2/inode.c +90 -14

164 lines not on this page · raw message

fs/nilfs2/mdt.c +4 -3
index c0361ce45f62..d3f6cb9c32a0 100644
@@ -531,7 +531,7 @@ int nilfs_mdt_save_to_shadow_map(struct inode *inode)
 		goto out;
 
 	ret = nilfs_copy_dirty_pages(&shadow->frozen_btnodes,
-				     &ii->i_btnode_cache);
+				     ii->i_assoc_inode->i_mapping);
 	if (ret)
 		goto out;
 
@@ -622,8 +622,9 @@ void nilfs_mdt_restore_from_shadow_map(struct inode *inode)
 	nilfs_clear_dirty_pages(inode->i_mapping, true);
 	nilfs_copy_back_pages(inode->i_mapping, &shadow->frozen_data);
 
-	nilfs_clear_dirty_pages(&ii->i_btnode_cache, true);
-	nilfs_copy_back_pages(&ii->i_btnode_cache, &shadow->frozen_btnodes);
+	nilfs_clear_dirty_pages(ii->i_assoc_inode->i_mapping, true);
+	nilfs_copy_back_pages(ii->i_assoc_inode->i_mapping,
+			      &shadow->frozen_btnodes);
 
 	nilfs_bmap_restore(ii->i_bmap, &shadow->bmap_store);
 
fs/nilfs2/nilfs.h +5 -9
index f8450ee3fd06..635383b30d67 100644
@@ -28,7 +28,7 @@
  * @i_xattr: <TODO>
  * @i_dir_start_lookup: page index of last successful search
  * @i_cno: checkpoint number for GC inode
- * @i_btnode_cache: cached pages of b-tree nodes
+ * @i_assoc_inode: associated inode (B-tree node cache holder or back pointer)
  * @i_dirty: list for connecting dirty files
  * @xattr_sem: semaphore for extended attributes processing
  * @i_bh: buffer contains disk inode
@@ -43,7 +43,7 @@ struct nilfs_inode_info {
 	__u64 i_xattr;	/* sector_t ??? */
 	__u32 i_dir_start_lookup;
 	__u64 i_cno;		/* check point number for GC inode */
-	struct address_space i_btnode_cache;
+	struct inode *i_assoc_inode;
 	struct list_head i_dirty;	/* List for connecting dirty files */
 
 #ifdef CONFIG_NILFS_XATTR
@@ -75,13 +75,6 @@ NILFS_BMAP_I(const struct nilfs_bmap *bmap)
 	return container_of(bmap, struct nilfs_inode_info, i_bmap_data);
 }
 
-static inline struct inode *NILFS_BTNC_I(struct address_space *btnc)
-{
-	struct nilfs_inode_info *ii =
-		container_of(btnc, struct nilfs_inode_info, i_btnode_cache);
-	return &ii->vfs_inode;
-}
-
 /*
  * Dynamic state flags of NILFS on-memory inode (i_state)
  */
@@ -98,6 +91,7 @@ enum {
 	NILFS_I_INODE_SYNC,		/* dsync is not allowed for inode */
 	NILFS_I_BMAP,			/* has bmap and btnode_cache */
 	NILFS_I_GCINODE,		/* inode for GC, on memory only */
+	NILFS_I_BTNC,			/* inode for btree node cache */
 };
 
 /*
@@ -264,6 +258,8 @@ struct inode *nilfs_iget(struct super_block *sb, struct nilfs_root *root,
 			 unsigned long ino);
 extern struct inode *nilfs_iget_for_gc(struct super_block *sb,
 				       unsigned long ino, __u64 cno);
+int nilfs_attach_btree_node_cache(struct inode *inode);
+void nilfs_detach_btree_node_cache(struct inode *inode);
 extern void nilfs_update_inode(struct inode *, struct buffer_head *, int);
 extern void nilfs_truncate(struct inode *);
 extern void nilfs_evict_inode(struct inode *);
fs/nilfs2/page.c +3 -4
index 171fb5cd427f..d1a148f0cae3 100644
@@ -448,10 +448,9 @@ void nilfs_mapping_init(struct address_space *mapping, struct inode *inode)
 /*
  * NILFS2 needs clear_page_dirty() in the following two cases:
  *
- * 1) For B-tree node pages and data pages of the dat/gcdat, NILFS2 clears
- *    page dirty flags when it copies back pages from the shadow cache
- *    (gcdat->{i_mapping,i_btnode_cache}) to its original cache
- *    (dat->{i_mapping,i_btnode_cache}).
+ * 1) For B-tree node pages and data pages of DAT file, NILFS2 clears dirty
+ *    flag of pages when it copies back pages from shadow cache to the
+ *    original cache.
  *
  * 2) Some B-tree operations like insertion or deletion may dispose buffers
  *    in dirty state, and this needs to cancel the dirty state of their pages.
fs/nilfs2/segment.c +6 -3
index e3726aca28ed..8350c2eaee75 100644
@@ -738,15 +738,18 @@ static void nilfs_lookup_dirty_node_buffers(struct inode *inode,
 					    struct list_head *listp)
 {
 	struct nilfs_inode_info *ii = NILFS_I(inode);
-	struct address_space *mapping = &ii->i_btnode_cache;
+	struct inode *btnc_inode = ii->i_assoc_inode;
 	struct pagevec pvec;
 	struct buffer_head *bh, *head;
 	unsigned int i;
 	pgoff_t index = 0;
 
+	if (!btnc_inode)
+		return;
+
 	pagevec_init(&pvec);
 
-	while (pagevec_lookup_tag(&pvec, mapping, &index,
+	while (pagevec_lookup_tag(&pvec, btnc_inode->i_mapping, &index,
 					PAGECACHE_TAG_DIRTY)) {
 		for (i = 0; i < pagevec_count(&pvec); i++) {
 			bh = head = page_buffers(pvec.pages[i]);
@@ -2415,7 +2418,7 @@ nilfs_remove_written_gcinodes(struct the_nilfs *nilfs, struct list_head *head)
 			continue;
 		list_del_init(&ii->i_dirty);
 		truncate_inode_pages(&ii->vfs_inode.i_data, 0);
-		nilfs_btnode_cache_clear(&ii->i_btnode_cache);
+		nilfs_btnode_cache_clear(ii->i_assoc_inode->i_mapping);
 		iput(&ii->vfs_inode);
 	}
 }
fs/nilfs2/super.c +2 -3
index 4abd928b0bc8..b9d30e8c43b0 100644
@@ -157,7 +157,8 @@ struct inode *nilfs_alloc_inode(struct super_block *sb)
 	ii->i_bh = NULL;
 	ii->i_state = 0;
 	ii->i_cno = 0;
-	nilfs_mapping_init(&ii->i_btnode_cache, &ii->vfs_inode);
+	ii->i_assoc_inode = NULL;
+	ii->i_bmap = &ii->i_bmap_data;
 	return &ii->vfs_inode;
 }
 
@@ -1377,8 +1378,6 @@ static void nilfs_inode_init_once(void *obj)
 #ifdef CONFIG_NILFS_XATTR
 	init_rwsem(&ii->xattr_sem);
 #endif
-	address_space_init_once(&ii->i_btnode_cache);
-	ii->i_bmap = &ii->i_bmap_data;
 	inode_init_once(&ii->vfs_inode);
 }
 
Signature · 1 line
2.35.1
author maintainer
#24 lore raw

PATCH 25/25 nilfs2: fix lockdep warnings during disk space reclamation

In reply to Sasha Levin

From: Ryusuke Konishi <[email protected]>

[ Upstream commit 6e211930f79aa45d422009a5f2e5467d2369ffe5 ]

During disk space reclamation, nilfs2 still emits the following lockdep warning due to page/folio operations on shadowed page caches that nilfs2 uses to get a snapshot of DAT file in memory:

  WARNING: CPU: 0 PID: 2643 at include/linux/backing-dev.h:272 __folio_mark_dirty+0x645/0x670
  ...
  RIP: 0010:__folio_mark_dirty+0x645/0x670
  ...
  Call Trace:
    filemap_dirty_folio+0x74/0xd0
    __set_page_dirty_nobuffers+0x85/0xb0
    nilfs_copy_dirty_pages+0x288/0x510 [nilfs2]
    nilfs_mdt_save_to_shadow_map+0x50/0xe0 [nilfs2]
    nilfs_clean_segments+0xee/0x5d0 [nilfs2]
    nilfs_ioctl_clean_segments.isra.19+0xb08/0xf40 [nilfs2]
    nilfs_ioctl+0xc52/0xfb0 [nilfs2]
    __x64_sys_ioctl+0x11d/0x170

This fixes the remaining warning by using inode objects to hold those page caches.

---
 fs/nilfs2/dat.c   |  4 ++-
 fs/nilfs2/inode.c | 63 ++++++++++++++++++++++++++++++++++++++++++++---
 fs/nilfs2/mdt.c   | 38 +++++++++++++++++++---------
 fs/nilfs2/mdt.h   |  6 ++---
 fs/nilfs2/nilfs.h |  2 ++
 5 files changed, 92 insertions(+), 21 deletions(-)
fs/nilfs2/dat.c +3 -1
index 8bccdf1158fc..1a3d183027b9 100644
@@ -497,7 +497,9 @@ int nilfs_dat_read(struct super_block *sb, size_t entry_size,
 	di = NILFS_DAT_I(dat);
 	lockdep_set_class(&di->mi.mi_sem, &dat_lock_key);
 	nilfs_palloc_setup_cache(dat, &di->palloc_cache);
-	nilfs_mdt_setup_shadow_map(dat, &di->shadow);
+	err = nilfs_mdt_setup_shadow_map(dat, &di->shadow);
+	if (err)
+		goto failed;
 
 	err = nilfs_read_inode_common(dat, raw_inode);
 	if (err)
fs/nilfs2/inode.c +59 -4

123 lines not on this page · raw message

fs/nilfs2/mdt.c +26 -12
index d3f6cb9c32a0..e80ef2c0a785 100644
@@ -469,9 +469,18 @@ int nilfs_mdt_init(struct inode *inode, gfp_t gfp_mask, size_t objsz)
 void nilfs_mdt_clear(struct inode *inode)
 {
 	struct nilfs_mdt_info *mdi = NILFS_MDT(inode);
+	struct nilfs_shadow_map *shadow = mdi->mi_shadow;
 
 	if (mdi->mi_palloc_cache)
 		nilfs_palloc_destroy_cache(inode);
+
+	if (shadow) {
+		struct inode *s_inode = shadow->inode;
+
+		shadow->inode = NULL;
+		iput(s_inode);
+		mdi->mi_shadow = NULL;
+	}
 }
 
 /**
@@ -505,12 +514,15 @@ int nilfs_mdt_setup_shadow_map(struct inode *inode,
 			       struct nilfs_shadow_map *shadow)
 {
 	struct nilfs_mdt_info *mi = NILFS_MDT(inode);
+	struct inode *s_inode;
 
 	INIT_LIST_HEAD(&shadow->frozen_buffers);
-	address_space_init_once(&shadow->frozen_data);
-	nilfs_mapping_init(&shadow->frozen_data, inode);
-	address_space_init_once(&shadow->frozen_btnodes);
-	nilfs_mapping_init(&shadow->frozen_btnodes, inode);
+
+	s_inode = nilfs_iget_for_shadow(inode);
+	if (IS_ERR(s_inode))
+		return PTR_ERR(s_inode);
+
+	shadow->inode = s_inode;
 	mi->mi_shadow = shadow;
 	return 0;
 }
@@ -524,13 +536,14 @@ int nilfs_mdt_save_to_shadow_map(struct inode *inode)
 	struct nilfs_mdt_info *mi = NILFS_MDT(inode);
 	struct nilfs_inode_info *ii = NILFS_I(inode);
 	struct nilfs_shadow_map *shadow = mi->mi_shadow;
+	struct inode *s_inode = shadow->inode;
 	int ret;
 
-	ret = nilfs_copy_dirty_pages(&shadow->frozen_data, inode->i_mapping);
+	ret = nilfs_copy_dirty_pages(s_inode->i_mapping, inode->i_mapping);
 	if (ret)
 		goto out;
 
-	ret = nilfs_copy_dirty_pages(&shadow->frozen_btnodes,
+	ret = nilfs_copy_dirty_pages(NILFS_I(s_inode)->i_assoc_inode->i_mapping,
 				     ii->i_assoc_inode->i_mapping);
 	if (ret)
 		goto out;
@@ -547,7 +560,7 @@ int nilfs_mdt_freeze_buffer(struct inode *inode, struct buffer_head *bh)
 	struct page *page;
 	int blkbits = inode->i_blkbits;
 
-	page = grab_cache_page(&shadow->frozen_data, bh->b_page->index);
+	page = grab_cache_page(shadow->inode->i_mapping, bh->b_page->index);
 	if (!page)
 		return -ENOMEM;
 
@@ -579,7 +592,7 @@ nilfs_mdt_get_frozen_buffer(struct inode *inode, struct buffer_head *bh)
 	struct page *page;
 	int n;
 
-	page = find_lock_page(&shadow->frozen_data, bh->b_page->index);
+	page = find_lock_page(shadow->inode->i_mapping, bh->b_page->index);
 	if (page) {
 		if (page_has_buffers(page)) {
 			n = bh_offset(bh) >> inode->i_blkbits;
@@ -620,11 +633,11 @@ void nilfs_mdt_restore_from_shadow_map(struct inode *inode)
 		nilfs_palloc_clear_cache(inode);
 
 	nilfs_clear_dirty_pages(inode->i_mapping, true);
-	nilfs_copy_back_pages(inode->i_mapping, &shadow->frozen_data);
+	nilfs_copy_back_pages(inode->i_mapping, shadow->inode->i_mapping);
 
 	nilfs_clear_dirty_pages(ii->i_assoc_inode->i_mapping, true);
 	nilfs_copy_back_pages(ii->i_assoc_inode->i_mapping,
-			      &shadow->frozen_btnodes);
+			      NILFS_I(shadow->inode)->i_assoc_inode->i_mapping);
 
 	nilfs_bmap_restore(ii->i_bmap, &shadow->bmap_store);
 
@@ -639,10 +652,11 @@ void nilfs_mdt_clear_shadow_map(struct inode *inode)
 {
 	struct nilfs_mdt_info *mi = NILFS_MDT(inode);
 	struct nilfs_shadow_map *shadow = mi->mi_shadow;
+	struct inode *shadow_btnc_inode = NILFS_I(shadow->inode)->i_assoc_inode;
 
 	down_write(&mi->mi_sem);
 	nilfs_release_frozen_buffers(shadow);
-	truncate_inode_pages(&shadow->frozen_data, 0);
-	truncate_inode_pages(&shadow->frozen_btnodes, 0);
+	truncate_inode_pages(shadow->inode->i_mapping, 0);
+	truncate_inode_pages(shadow_btnc_inode->i_mapping, 0);
 	up_write(&mi->mi_sem);
 }
fs/nilfs2/mdt.h +2 -4
index e77aea4bb921..9d8ac0d27c16 100644
@@ -18,14 +18,12 @@
 /**
  * struct nilfs_shadow_map - shadow mapping of meta data file
  * @bmap_store: shadow copy of bmap state
- * @frozen_data: shadowed dirty data pages
- * @frozen_btnodes: shadowed dirty b-tree nodes' pages
+ * @inode: holder of page caches used in shadow mapping
  * @frozen_buffers: list of frozen buffers
  */
 struct nilfs_shadow_map {
 	struct nilfs_bmap_store bmap_store;
-	struct address_space frozen_data;
-	struct address_space frozen_btnodes;
+	struct inode *inode;
 	struct list_head frozen_buffers;
 };
 
fs/nilfs2/nilfs.h +2 -0
index 635383b30d67..9ca165bc97d2 100644
@@ -92,6 +92,7 @@ enum {
 	NILFS_I_BMAP,			/* has bmap and btnode_cache */
 	NILFS_I_GCINODE,		/* inode for GC, on memory only */
 	NILFS_I_BTNC,			/* inode for btree node cache */
+	NILFS_I_SHADOW,			/* inode for shadowed page cache */
 };
 
 /*
@@ -260,6 +261,7 @@ extern struct inode *nilfs_iget_for_gc(struct super_block *sb,
 				       unsigned long ino, __u64 cno);
 int nilfs_attach_btree_node_cache(struct inode *inode);
 void nilfs_detach_btree_node_cache(struct inode *inode);
+struct inode *nilfs_iget_for_shadow(struct inode *inode);
 extern void nilfs_update_inode(struct inode *, struct buffer_head *, int);
 extern void nilfs_truncate(struct inode *);
 extern void nilfs_evict_inode(struct inode *);
Signature · 1 line
2.35.1