linux-kernel - [PATCH v2] PCI: vmd: Fix spinlock usage on config access for RT kernel

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20241218115951.83062-1-ryotkkr98@gmail.com>
Date: Wed, 18 Dec 2024 20:59:51 +0900
From: Ryo Takakura <ryotkkr98@...il.com>
To: lgoncalv@...hat.com,
	bhelgaas@...gle.com,
	jonathan.derrick@...ux.dev,
	kw@...ux.com,
	lpieralisi@...nel.org,
	manivannan.sadhasivam@...aro.org,
	nirmal.patel@...ux.intel.com,
	robh@...nel.org,
	bigeasy@...utronix.de,
	rostedt@...dmis.org
Cc: linux-kernel@...r.kernel.org,
	linux-pci@...r.kernel.org,
	linux-rt-devel@...ts.linux.dev,
	Ryo Takakura <ryotkkr98@...il.com>
Subject: [PATCH v2] PCI: vmd: Fix spinlock usage on config access for RT kernel

PCI config access is locked with pci_lock which serializes
pci_user/bus_write_config*() and pci_user/bus_read_config*().
The subsequently invoked vmd_pci_write() and vmd_pci_read() are also
serialized as they are only invoked by them respectively.

Remove cfg_lock which is taken by vmd_pci_write() and vmd_pci_read()
for their serialization as its already serialized by pci_lock.

This fixes the spinlock_t(cfg_lock) usage within
raw_spinlock_t(pci_lock) on RT kernels where spinlock_t becomes
sleepable.

Reported as following:
[   18.756807] BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48
[   18.756810] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 1617, name: nodedev-init
[   18.756810] preempt_count: 1, expected: 0
[   18.756811] RCU nest depth: 0, expected: 0
[   18.756811] INFO: lockdep is turned off.
[   18.756812] irq event stamp: 0
[   18.756812] hardirqs last  enabled at (0): [<0000000000000000>] 0x0
[   18.756815] hardirqs last disabled at (0): [<ffffffff864f1fe2>] copy_process+0xa62/0x2d90
[   18.756819] softirqs last  enabled at (0): [<ffffffff864f1fe2>] copy_process+0xa62/0x2d90
[   18.756820] softirqs last disabled at (0): [<0000000000000000>] 0x0
[   18.756822] CPU: 3 UID: 0 PID: 1617 Comm: nodedev-init Tainted: G        W          6.12.1 #11
[   18.756823] Tainted: [W]=WARN
[   18.756824] Hardware name: Dell Inc. Vostro 3710/0K1D6X, BIOS 1.14.0 06/09/2023
[   18.756825] Call Trace:
[   18.756826]  <TASK>
[   18.756827]  dump_stack_lvl+0x9b/0xf0
[   18.756830]  dump_stack+0x10/0x20
[   18.756831]  __might_resched+0x158/0x230
[   18.756833]  rt_spin_lock+0x4e/0x130
[   18.756837]  ? vmd_pci_read+0x8d/0x100 [vmd]
[   18.756839]  vmd_pci_read+0x8d/0x100 [vmd]
[   18.756840]  pci_user_read_config_byte+0x6f/0xe0
[   18.756843]  pci_read_config+0xfe/0x290
[   18.756845]  sysfs_kf_bin_read+0x68/0x90
[   18.756847]  kernfs_fop_read_iter+0xd7/0x200
[   18.756850]  vfs_read+0x26d/0x360
[   18.756853]  ksys_read+0x70/0xf0
[   18.756855]  __x64_sys_read+0x1a/0x20
[   18.756857]  x64_sys_call+0x1715/0x20d0
[   18.756859]  do_syscall_64+0x8f/0x170
[   18.756861]  ? syscall_exit_to_user_mode+0xcd/0x2c0
[   18.756863]  ? do_syscall_64+0x9b/0x170
[   18.756865]  entry_SYSCALL_64_after_hwframe+0x76/0x7e

Signed-off-by: Ryo Takakura <ryotkkr98@...il.com>
---

Thanks Luis for feedback!

Changes since v1:
https://lore.kernel.org/lkml/20241215141321.383144-1-ryotkkr98@gmail.com/T/

- Remove cfg_lock instead of converting it to raw spinlock as suggested
  by Sebastian[0].
  
[0] https://lore.kernel.org/linux-rt-users/20230620154434.MosrzRUh@linutronix.de/

---
 drivers/pci/controller/vmd.c | 8 --------
 1 file changed, 8 deletions(-)

diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c
index 9d9596947..2be898424 100644
--- a/drivers/pci/controller/vmd.c
+++ b/drivers/pci/controller/vmd.c
@@ -125,7 +125,6 @@ struct vmd_irq_list {
 struct vmd_dev {
 	struct pci_dev		*dev;
 
-	spinlock_t		cfg_lock;
 	void __iomem		*cfgbar;
 
 	int msix_count;
@@ -385,13 +384,11 @@ static int vmd_pci_read(struct pci_bus *bus, unsigned int devfn, int reg,
 {
 	struct vmd_dev *vmd = vmd_from_bus(bus);
 	void __iomem *addr = vmd_cfg_addr(vmd, bus, devfn, reg, len);
-	unsigned long flags;
 	int ret = 0;
 
 	if (!addr)
 		return -EFAULT;
 
-	spin_lock_irqsave(&vmd->cfg_lock, flags);
 	switch (len) {
 	case 1:
 		*value = readb(addr);
@@ -406,7 +403,6 @@ static int vmd_pci_read(struct pci_bus *bus, unsigned int devfn, int reg,
 		ret = -EINVAL;
 		break;
 	}
-	spin_unlock_irqrestore(&vmd->cfg_lock, flags);
 	return ret;
 }
 
@@ -420,13 +416,11 @@ static int vmd_pci_write(struct pci_bus *bus, unsigned int devfn, int reg,
 {
 	struct vmd_dev *vmd = vmd_from_bus(bus);
 	void __iomem *addr = vmd_cfg_addr(vmd, bus, devfn, reg, len);
-	unsigned long flags;
 	int ret = 0;
 
 	if (!addr)
 		return -EFAULT;
 
-	spin_lock_irqsave(&vmd->cfg_lock, flags);
 	switch (len) {
 	case 1:
 		writeb(value, addr);
@@ -444,7 +438,6 @@ static int vmd_pci_write(struct pci_bus *bus, unsigned int devfn, int reg,
 		ret = -EINVAL;
 		break;
 	}
-	spin_unlock_irqrestore(&vmd->cfg_lock, flags);
 	return ret;
 }
 
@@ -1009,7 +1002,6 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	if (features & VMD_FEAT_OFFSET_FIRST_VECTOR)
 		vmd->first_vec = 1;
 
-	spin_lock_init(&vmd->cfg_lock);
 	pci_set_drvdata(dev, vmd);
 	err = vmd_enable_domain(vmd, features);
 	if (err)
-- 
2.34.1