From: Saurabh Sengar <ssengar@linux.microsoft.com>
To: bhelgaas@google.com, linux-pci@vger.kernel.org,
linux-kernel@vger.kernel.org
Cc: alexander.stein@ew.tq-group.com, ssengar@linux.microsoft.com,
decui@microsoft.com
Subject: [PATCH] PCI/sysfs: Fix race in pci sysfs creation
Date: Fri, 8 Dec 2023 19:46:16 -0800 [thread overview]
Message-ID: <1702093576-30405-1-git-send-email-ssengar@linux.microsoft.com> (raw)
Currently there is a race in calling pci_create_resource_files function
from two different therads, first therad is triggered by pci_sysfs_init
from the late initcall where as the second thread is initiated by
pci_bus_add_devices from the respective PCI drivers probe.
The synchronization between these threads relies on the sysfs_initialized
flag. However, in pci_sysfs_init, sysfs_initialized is set right before
calling pci_create_resource_files which is wrong as it can create race
condition with pci_bus_add_devices threads. Fix this by setting
sysfs_initialized flag at the end of pci_sysfs_init and direecly call the
pci_create_resource_files function from it.
There can be an additional case where driver probe is so delayed that
pci_bus_add_devices is called after the sysfs is created by pci_sysfs_init.
In such cases, attempting to access already existing sysfs resources is
unnecessary. Fix this by adding a check for sysfs attributes and return
if they are already allocated.
In both cases, the consequence will be the removal of sysfs resources that
were appropriately allocated by pci_sysfs_init following the warning below.
[ 3.376688] sysfs: cannot create duplicate filename '/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A03:00/device:07/VMBUS:01/47505500-0001-0000-3130-444531454238/pci0001:00/0001:00:00.0/resource0'
[ 3.385103] CPU: 3 PID: 9 Comm: kworker/u8:0 Not tainted 5.15.0-1046-azure #53~20.04.1-Ubuntu
[ 3.389585] Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090008 12/07/2018
[ 3.394663] Workqueue: events_unbound async_run_entry_fn
[ 3.397687] Call Trace:
[ 3.399312] <TASK>
[ 3.400780] dump_stack_lvl+0x38/0x4d
[ 3.402998] dump_stack+0x10/0x16
[ 3.406050] sysfs_warn_dup.cold+0x17/0x2b
[ 3.408476] sysfs_add_file_mode_ns+0x17b/0x190
[ 3.411072] sysfs_create_bin_file+0x64/0x90
[ 3.413514] pci_create_attr+0xc7/0x260
[ 3.415827] pci_create_resource_files+0x6f/0x150
[ 3.418455] pci_create_sysfs_dev_files+0x18/0x30
[ 3.421136] pci_bus_add_device+0x30/0x70
[ 3.423512] pci_bus_add_devices+0x31/0x70
[ 3.425958] hv_pci_probe+0x4ce/0x640
[ 3.428106] vmbus_probe+0x67/0x90
[ 3.430121] really_probe.part.0+0xcb/0x380
[ 3.432516] really_probe+0x40/0x80
[ 3.434581] __driver_probe_device+0xe8/0x140
[ 3.437119] driver_probe_device+0x23/0xb0
[ 3.439504] __driver_attach_async_helper+0x31/0x90
[ 3.442296] async_run_entry_fn+0x33/0x120
[ 3.444666] process_one_work+0x225/0x3d0
[ 3.447043] worker_thread+0x4d/0x3e0
[ 3.449233] ? process_one_work+0x3d0/0x3d0
[ 3.451632] kthread+0x12a/0x150
[ 3.453583] ? set_kthread_struct+0x50/0x50
[ 3.456103] ret_from_fork+0x22/0x30
Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com>
---
There has been earlier attempts to fix this problem, below are the patches
for reference of these attempts.
1. https://lore.kernel.org/linux-pci/20230316103036.1837869-1-alexander.stein@ew.tq-group.com/T/#u
2. https://lwn.net/ml/linux-kernel/20230316091540.494366-1-alexander.stein@ew.tq-group.com/
Bug details: https://bugzilla.kernel.org/show_bug.cgi?id=215515
drivers/pci/pci-sysfs.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
index f2909ae93f2f..a31f6f2cf309 100644
--- a/drivers/pci/pci-sysfs.c
+++ b/drivers/pci/pci-sysfs.c
@@ -1230,6 +1230,10 @@ static int pci_create_resource_files(struct pci_dev *pdev)
if (!pci_resource_len(pdev, i))
continue;
+ /* Check if resource already allocated and proceed no further */
+ if (pdev->res_attr[i] || pdev->res_attr_wc[i])
+ return 0;
+
retval = pci_create_attr(pdev, i, 0);
/* for prefetchable resources, create a WC mappable file */
if (!retval && arch_can_pci_mmap_wc() &&
@@ -1411,9 +1415,8 @@ static int __init pci_sysfs_init(void)
struct pci_bus *pbus = NULL;
int retval;
- sysfs_initialized = 1;
for_each_pci_dev(pdev) {
- retval = pci_create_sysfs_dev_files(pdev);
+ retval = pci_create_resource_files(pdev);
if (retval) {
pci_dev_put(pdev);
return retval;
@@ -1423,6 +1426,8 @@ static int __init pci_sysfs_init(void)
while ((pbus = pci_find_next_bus(pbus)))
pci_create_legacy_files(pbus);
+ sysfs_initialized = 1;
+
return 0;
}
late_initcall(pci_sysfs_init);
--
2.25.1
next reply other threads:[~2023-12-09 3:46 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-09 3:46 Saurabh Sengar [this message]
2023-12-12 7:19 ` [PATCH] PCI/sysfs: Fix race in pci sysfs creation Alexander Stein
2023-12-12 8:21 ` Saurabh Singh Sengar
2023-12-12 8:28 ` Saurabh Singh Sengar
2024-01-04 5:38 ` Saurabh Singh Sengar
2024-01-20 6:41 ` Saurabh Singh Sengar
2024-02-02 7:17 ` Saurabh Singh Sengar
2024-02-06 22:07 ` Bjorn Helgaas
2024-02-07 16:30 ` Saurabh Singh Sengar
2024-02-27 17:14 ` Saurabh Singh Sengar
2024-02-28 15:22 ` Bjorn Helgaas
2024-02-28 17:22 ` Krzysztof Wilczyński
2024-02-28 18:16 ` Saurabh Singh Sengar
2024-04-15 18:15 ` [EXTERNAL] " Saurabh Singh Sengar
2024-03-02 8:57 ` Lukas Wunner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1702093576-30405-1-git-send-email-ssengar@linux.microsoft.com \
--to=ssengar@linux.microsoft.com \
--cc=alexander.stein@ew.tq-group.com \
--cc=bhelgaas@google.com \
--cc=decui@microsoft.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).