CVE-2025-71117

*** Pendiente de traducción *** In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> block: Remove queue freezing from several sysfs store callbacks<br /> <br /> Freezing the request queue from inside sysfs store callbacks may cause a<br /> deadlock in combination with the dm-multipath driver and the<br /> queue_if_no_path option. Additionally, freezing the request queue slows<br /> down system boot on systems where sysfs attributes are set synchronously.<br /> <br /> Fix this by removing the blk_mq_freeze_queue() / blk_mq_unfreeze_queue()<br /> calls from the store callbacks that do not strictly need these callbacks.<br /> Add the __data_racy annotation to request_queue.rq_timeout to suppress<br /> KCSAN data race reports about the rq_timeout reads.<br /> <br /> This patch may cause a small delay in applying the new settings.<br /> <br /> For all the attributes affected by this patch, I/O will complete<br /> correctly whether the old or the new value of the attribute is used.<br /> <br /> This patch affects the following sysfs attributes:<br /> * io_poll_delay<br /> * io_timeout<br /> * nomerges<br /> * read_ahead_kb<br /> * rq_affinity<br /> <br /> Here is an example of a deadlock triggered by running test srp/002<br /> if this patch is not applied:<br /> <br /> task:multipathd<br /> Call Trace:<br /> <br /> __schedule+0x8c1/0x1bf0<br /> schedule+0xdd/0x270<br /> schedule_preempt_disabled+0x1c/0x30<br /> __mutex_lock+0xb89/0x1650<br /> mutex_lock_nested+0x1f/0x30<br /> dm_table_set_restrictions+0x823/0xdf0<br /> __bind+0x166/0x590<br /> dm_swap_table+0x2a7/0x490<br /> do_resume+0x1b1/0x610<br /> dev_suspend+0x55/0x1a0<br /> ctl_ioctl+0x3a5/0x7e0<br /> dm_ctl_ioctl+0x12/0x20<br /> __x64_sys_ioctl+0x127/0x1a0<br /> x64_sys_call+0xe2b/0x17d0<br /> do_syscall_64+0x96/0x3a0<br /> entry_SYSCALL_64_after_hwframe+0x4b/0x53<br /> <br /> task:(udev-worker)<br /> Call Trace:<br /> <br /> __schedule+0x8c1/0x1bf0<br /> schedule+0xdd/0x270<br /> blk_mq_freeze_queue_wait+0xf2/0x140<br /> blk_mq_freeze_queue_nomemsave+0x23/0x30<br /> queue_ra_store+0x14e/0x290<br /> queue_attr_store+0x23e/0x2c0<br /> sysfs_kf_write+0xde/0x140<br /> kernfs_fop_write_iter+0x3b2/0x630<br /> vfs_write+0x4fd/0x1390<br /> ksys_write+0xfd/0x230<br /> __x64_sys_write+0x76/0xc0<br /> x64_sys_call+0x276/0x17d0<br /> do_syscall_64+0x96/0x3a0<br /> entry_SYSCALL_64_after_hwframe+0x4b/0x53<br />