ghsa-3r3f-7rrr-c2h9
Vulnerability from github
In the Linux kernel, the following vulnerability has been resolved:
PCI/PM: Drain runtime-idle callbacks before driver removal
A race condition between the .runtime_idle() callback and the .remove() callback in the rtsx_pcr PCI driver leads to a kernel crash due to an unhandled page fault [1].
The problem is that rtsx_pci_runtime_idle() is not expected to be running after pm_runtime_get_sync() has been called, but the latter doesn't really guarantee that. It only guarantees that the suspend and resume callbacks will not be running when it returns.
However, if a .runtime_idle() callback is already running when pm_runtime_get_sync() is called, the latter will notice that the runtime PM status of the device is RPM_ACTIVE and it will return right away without waiting for the former to complete. In fact, it cannot wait for .runtime_idle() to complete because it may be called from that callback (it arguably does not make much sense to do that, but it is not strictly prohibited).
Thus in general, whoever is providing a .runtime_idle() callback needs to protect it from running in parallel with whatever code runs after pm_runtime_get_sync(). [Note that .runtime_idle() will not start after pm_runtime_get_sync() has returned, but it may continue running then if it has started earlier.]
One way to address that race condition is to call pm_runtime_barrier() after pm_runtime_get_sync() (not before it, because a nonzero value of the runtime PM usage counter is necessary to prevent runtime PM callbacks from being invoked) to wait for the .runtime_idle() callback to complete should it be running at that point. A suitable place for doing that is in pci_device_remove() which calls pm_runtime_get_sync() before removing the driver, so it may as well call pm_runtime_barrier() subsequently, which will prevent the race in question from occurring, not just in the rtsx_pcr driver, but in any PCI drivers providing .runtime_idle() callbacks.
{
"affected": [],
"aliases": [
"CVE-2024-35809"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2024-05-17T14:15:14Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\nPCI/PM: Drain runtime-idle callbacks before driver removal\n\nA race condition between the .runtime_idle() callback and the .remove()\ncallback in the rtsx_pcr PCI driver leads to a kernel crash due to an\nunhandled page fault [1].\n\nThe problem is that rtsx_pci_runtime_idle() is not expected to be running\nafter pm_runtime_get_sync() has been called, but the latter doesn\u0027t really\nguarantee that. It only guarantees that the suspend and resume callbacks\nwill not be running when it returns.\n\nHowever, if a .runtime_idle() callback is already running when\npm_runtime_get_sync() is called, the latter will notice that the runtime PM\nstatus of the device is RPM_ACTIVE and it will return right away without\nwaiting for the former to complete. In fact, it cannot wait for\n.runtime_idle() to complete because it may be called from that callback (it\narguably does not make much sense to do that, but it is not strictly\nprohibited).\n\nThus in general, whoever is providing a .runtime_idle() callback needs\nto protect it from running in parallel with whatever code runs after\npm_runtime_get_sync(). [Note that .runtime_idle() will not start after\npm_runtime_get_sync() has returned, but it may continue running then if it\nhas started earlier.]\n\nOne way to address that race condition is to call pm_runtime_barrier()\nafter pm_runtime_get_sync() (not before it, because a nonzero value of the\nruntime PM usage counter is necessary to prevent runtime PM callbacks from\nbeing invoked) to wait for the .runtime_idle() callback to complete should\nit be running at that point. A suitable place for doing that is in\npci_device_remove() which calls pm_runtime_get_sync() before removing the\ndriver, so it may as well call pm_runtime_barrier() subsequently, which\nwill prevent the race in question from occurring, not just in the rtsx_pcr\ndriver, but in any PCI drivers providing .runtime_idle() callbacks.",
"id": "GHSA-3r3f-7rrr-c2h9",
"modified": "2024-06-27T15:30:39Z",
"published": "2024-05-17T15:31:09Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2024-35809"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/47d8aafcfe313511a98f165a54d0adceb34e54b1"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/6347348c6aba52dda0b33296684cbb627bdc6970"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/7cc94dd36e48879e76ae7a8daea4ff322b7d9674"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/900b81caf00c89417172afe0e7e49ac4eb110f4b"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/9a87375bb586515c0af63d5dcdcd58ec4acf20a6"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/9d5286d4e7f68beab450deddbb6a32edd5ecf4bf"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/bbe068b24409ef740657215605284fc7cdddd491"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/d534198311c345e4b062c4b88bb609efb8bd91d5"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/d86ad8c3e152349454b82f37007ff6ba45f26989"
},
{
"type": "WEB",
"url": "https://lists.debian.org/debian-lts-announce/2024/06/msg00017.html"
},
{
"type": "WEB",
"url": "https://lists.debian.org/debian-lts-announce/2024/06/msg00020.html"
}
],
"schema_version": "1.4.0",
"severity": []
}