CVE-2022-50488

*** Pendiente de traducción *** In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> block, bfq: fix possible uaf for &amp;#39;bfqq-&gt;bic&amp;#39;<br /> <br /> Our test report a uaf for &amp;#39;bfqq-&gt;bic&amp;#39; in 5.10:<br /> <br /> ==================================================================<br /> BUG: KASAN: use-after-free in bfq_select_queue+0x378/0xa30<br /> <br /> CPU: 6 PID: 2318352 Comm: fsstress Kdump: loaded Not tainted 5.10.0-60.18.0.50.h602.kasan.eulerosv2r11.x86_64 #1<br /> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-20220320_160524-szxrtosci10000 04/01/2014<br /> Call Trace:<br /> bfq_select_queue+0x378/0xa30<br /> bfq_dispatch_request+0xe8/0x130<br /> blk_mq_do_dispatch_sched+0x62/0xb0<br /> __blk_mq_sched_dispatch_requests+0x215/0x2a0<br /> blk_mq_sched_dispatch_requests+0x8f/0xd0<br /> __blk_mq_run_hw_queue+0x98/0x180<br /> __blk_mq_delay_run_hw_queue+0x22b/0x240<br /> blk_mq_run_hw_queue+0xe3/0x190<br /> blk_mq_sched_insert_requests+0x107/0x200<br /> blk_mq_flush_plug_list+0x26e/0x3c0<br /> blk_finish_plug+0x63/0x90<br /> __iomap_dio_rw+0x7b5/0x910<br /> iomap_dio_rw+0x36/0x80<br /> ext4_dio_read_iter+0x146/0x190 [ext4]<br /> ext4_file_read_iter+0x1e2/0x230 [ext4]<br /> new_sync_read+0x29f/0x400<br /> vfs_read+0x24e/0x2d0<br /> ksys_read+0xd5/0x1b0<br /> do_syscall_64+0x33/0x40<br /> entry_SYSCALL_64_after_hwframe+0x61/0xc6<br /> <br /> Commit 3bc5e683c67d ("bfq: Split shared queues on move between cgroups")<br /> changes that move process to a new cgroup will allocate a new bfqq to<br /> use, however, the old bfqq and new bfqq can point to the same bic:<br /> <br /> 1) Initial state, two process with io in the same cgroup.<br /> <br /> Process 1 Process 2<br /> (BIC1) (BIC2)<br /> | Λ | Λ<br /> | | | |<br /> V | V |<br /> bfqq1 bfqq2<br /> <br /> 2) bfqq1 is merged to bfqq2.<br /> <br /> Process 1 Process 2<br /> (BIC1) (BIC2)<br /> | |<br /> \-------------\|<br /> V<br /> bfqq1 bfqq2(coop)<br /> <br /> 3) Process 1 exit, then issue new io(denoce IOA) from Process 2.<br /> <br /> (BIC2)<br /> | Λ<br /> | |<br /> V |<br /> bfqq2(coop)<br /> <br /> 4) Before IOA is completed, move Process 2 to another cgroup and issue io.<br /> <br /> Process 2<br /> (BIC2)<br /> Λ<br /> |\--------------\<br /> | V<br /> bfqq2 bfqq3<br /> <br /> Now that BIC2 points to bfqq3, while bfqq2 and bfqq3 both point to BIC2.<br /> If all the requests are completed, and Process 2 exit, BIC2 will be<br /> freed while there is no guarantee that bfqq2 will be freed before BIC2.<br /> <br /> Fix the problem by clearing bfqq-&gt;bic while bfqq is detached from bic.