mirror of
https://github.com/fkie-cad/nvd-json-data-feeds.git
synced 2025-05-29 01:31:20 +00:00
33 lines
2.8 KiB
JSON
33 lines
2.8 KiB
JSON
{
|
|
"id": "CVE-2023-52767",
|
|
"sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
|
|
"published": "2024-05-21T16:15:15.917",
|
|
"lastModified": "2024-05-21T16:53:56.550",
|
|
"vulnStatus": "Awaiting Analysis",
|
|
"cveTags": [],
|
|
"descriptions": [
|
|
{
|
|
"lang": "en",
|
|
"value": "In the Linux kernel, the following vulnerability has been resolved:\n\ntls: fix NULL deref on tls_sw_splice_eof() with empty record\n\nsyzkaller discovered that if tls_sw_splice_eof() is executed as part of\nsendfile() when the plaintext/ciphertext sk_msg are empty, the send path\ngets confused because the empty ciphertext buffer does not have enough\nspace for the encryption overhead. This causes tls_push_record() to go on\nthe `split = true` path (which is only supposed to be used when interacting\nwith an attached BPF program), and then get further confused and hit the\ntls_merge_open_record() path, which then assumes that there must be at\nleast one populated buffer element, leading to a NULL deref.\n\nIt is possible to have empty plaintext/ciphertext buffers if we previously\nbailed from tls_sw_sendmsg_locked() via the tls_trim_both_msgs() path.\ntls_sw_push_pending_record() already handles this case correctly; let's do\nthe same check in tls_sw_splice_eof()."
|
|
},
|
|
{
|
|
"lang": "es",
|
|
"value": "En el kernel de Linux, se resolvi\u00f3 la siguiente vulnerabilidad: tls: corrige NULL deref en tls_sw_splice_eof() con registro vac\u00edo syzkaller descubri\u00f3 que si tls_sw_splice_eof() se ejecuta como parte de sendfile() cuando el texto plano/texto cifrado sk_msg est\u00e1 vac\u00edo, el env\u00edo La ruta se confunde porque el b\u00fafer de texto cifrado vac\u00edo no tiene suficiente espacio para la sobrecarga de cifrado. Esto hace que tls_push_record() vaya a la ruta `split = true` (que se supone que solo debe usarse al interactuar con un programa BPF adjunto), y luego se confunda a\u00fan m\u00e1s y acceda a la ruta tls_merge_open_record(), que luego supone que hay debe haber al menos un elemento de b\u00fafer poblado, lo que lleva a una deref NULL. Es posible tener buffers de texto plano/texto cifrado vac\u00edos si previamente salimos de tls_sw_sendmsg_locked() a trav\u00e9s de la ruta tls_trim_both_msgs(). tls_sw_push_pending_record() ya maneja este caso correctamente; hagamos la misma verificaci\u00f3n en tls_sw_splice_eof()."
|
|
}
|
|
],
|
|
"metrics": {},
|
|
"references": [
|
|
{
|
|
"url": "https://git.kernel.org/stable/c/2214e2bb5489145aba944874d0ee1652a0a63dc8",
|
|
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
|
|
},
|
|
{
|
|
"url": "https://git.kernel.org/stable/c/53f2cb491b500897a619ff6abd72f565933760f0",
|
|
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
|
|
},
|
|
{
|
|
"url": "https://git.kernel.org/stable/c/944900fe2736c07288efe2d9394db4d3ca23f2c9",
|
|
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
|
|
}
|
|
]
|
|
} |