GHSA-FQRQ-QCW6-WMQQ
Vulnerability from github – Published: 2026-10-28 12:31 – Updated: 2026-10-28 12:31In the Linux kernel, the following vulnerability has been resolved:
accel/rocket: search every core slot when a core is removed
rocket_remove() decrements rdev->num_cores for each core it removes, while find_core_for_dev() searches slots 0 to num_cores - 1. Unbinding the cores in the order they were bound therefore loses the last one: by the time it is removed the search range has already shrunk past its slot, so find_core_for_dev() returns -1 and rocket_remove() gives up without doing anything.
num_cores never reaches zero, rocket_device_fini() never runs, and the file-scoped rdev keeps pointing at a device that is going away. Binding the cores again starts from that stale count, because rocket_probe() takes rdev->num_cores as the slot to fill. On an RK3588, which describes three cores, the second round lands on slots 1, 2 and 3 while rdev->cores was allocated with room for three:
rocket fdab0000.npu: Rockchip NPU core 1 version: 1179210309 rocket fdac0000.npu: Rockchip NPU core 2 version: 1179210309 rocket fdad0000.npu: Rockchip NPU core 3 version: 1179210309
The write to rdev->cores[3] is past the end of the array.
Nothing in tree reads the core array often enough to notice, so the overrun is silent today. It turned up while testing a devfreq series on top of this, where a worker walks every core a few times a second, and UBSAN caught the first bool it read out of the overrun entry:
UBSAN: invalid-load in drivers/accel/rocket/rocket_devfreq.c:47:10 load of value 5 is not a valid value for type '_Bool' Workqueue: devfreq_wq devfreq_monitor
Record how many slots were allocated and search all of them. Every core is then found on removal, num_cores reaches zero, the device is torn down and a later bind starts from a clean rdev.
This does not make unbinding a single core out of several work. probe still takes num_cores as the slot to fill, so rebinding one core while its siblings stay bound would write over a slot that is already in use, and rocket_open() still reaches for cores[0] whether or not anything is there. Both of those want more thought than a fix should carry.
Found by unbinding and rebinding all three cores on an Orange Pi 5 Plus. With this applied, 25 unbind/rebind rounds and 5 module unload/reload rounds run clean there: the cores land in slots 0, 1 and 2 every time, whichever order they are bound in, and the shared supply goes back to a single user after each round.
{
"affected": [],
"aliases": [
"CVE-2026-99184"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-10-28T11:17:39Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\naccel/rocket: search every core slot when a core is removed\n\nrocket_remove() decrements rdev-\u003enum_cores for each core it removes,\nwhile find_core_for_dev() searches slots 0 to num_cores - 1. Unbinding\nthe cores in the order they were bound therefore loses the last one: by\nthe time it is removed the search range has already shrunk past its\nslot, so find_core_for_dev() returns -1 and rocket_remove() gives up\nwithout doing anything.\n\nnum_cores never reaches zero, rocket_device_fini() never runs, and the\nfile-scoped rdev keeps pointing at a device that is going away. Binding\nthe cores again starts from that stale count, because rocket_probe()\ntakes rdev-\u003enum_cores as the slot to fill. On an RK3588, which describes\nthree cores, the second round lands on slots 1, 2 and 3 while rdev-\u003ecores\nwas allocated with room for three:\n\n rocket fdab0000.npu: Rockchip NPU core 1 version: 1179210309\n rocket fdac0000.npu: Rockchip NPU core 2 version: 1179210309\n rocket fdad0000.npu: Rockchip NPU core 3 version: 1179210309\n\nThe write to rdev-\u003ecores[3] is past the end of the array.\n\nNothing in tree reads the core array often enough to notice, so the\noverrun is silent today. It turned up while testing a devfreq series on\ntop of this, where a worker walks every core a few times a second, and\nUBSAN caught the first bool it read out of the overrun entry:\n\n UBSAN: invalid-load in drivers/accel/rocket/rocket_devfreq.c:47:10\n load of value 5 is not a valid value for type \u0027_Bool\u0027\n Workqueue: devfreq_wq devfreq_monitor\n\nRecord how many slots were allocated and search all of them. Every core\nis then found on removal, num_cores reaches zero, the device is torn down\nand a later bind starts from a clean rdev.\n\nThis does not make unbinding a single core out of several work. probe\nstill takes num_cores as the slot to fill, so rebinding one core while\nits siblings stay bound would write over a slot that is already in use,\nand rocket_open() still reaches for cores[0] whether or not anything is\nthere. Both of those want more thought than a fix should carry.\n\nFound by unbinding and rebinding all three cores on an Orange Pi 5 Plus.\nWith this applied, 25 unbind/rebind rounds and 5 module unload/reload\nrounds run clean there: the cores land in slots 0, 1 and 2 every time,\nwhichever order they are bound in, and the shared supply goes back to a\nsingle user after each round.",
"id": "GHSA-fqrq-qcw6-wmqq",
"modified": "2026-10-28T12:31:07Z",
"published": "2026-10-28T12:31:07Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-99184"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/134451cf95cb6aa04b725fbfa04e878d072756f5"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/5078c91f5c84bb383d0a10ed39856f98575e39cf"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/6eff1d69dae808a1518aee5a58d0853a8db3f1be"
}
],
"schema_version": "1.4.0",
"severity": []
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.