"value":"In the Linux kernel, the following vulnerability has been resolved:\n\nmd: Don't register sync_thread for reshape directly\n\nCurrently, if reshape is interrupted, then reassemble the array will\nregister sync_thread directly from pers->run(), in this case\n'MD_RECOVERY_RUNNING' is set directly, however, there is no guarantee\nthat md_do_sync() will be executed, hence stop_sync_thread() will hang\nbecause 'MD_RECOVERY_RUNNING' can't be cleared.\n\nLast patch make sure that md_do_sync() will set MD_RECOVERY_DONE,\nhowever, following hang can still be triggered by dm-raid test\nshell/lvconvert-raid-reshape.sh occasionally:\n\n[root@fedora ~]# cat /proc/1982/stack\n[<0>] stop_sync_thread+0x1ab/0x270 [md_mod]\n[<0>] md_frozen_sync_thread+0x5c/0xa0 [md_mod]\n[<0>] raid_presuspend+0x1e/0x70 [dm_raid]\n[<0>] dm_table_presuspend_targets+0x40/0xb0 [dm_mod]\n[<0>] __dm_destroy+0x2a5/0x310 [dm_mod]\n[<0>] dm_destroy+0x16/0x30 [dm_mod]\n[<0>] dev_remove+0x165/0x290 [dm_mod]\n[<0>] ctl_ioctl+0x4bb/0x7b0 [dm_mod]\n[<0>] dm_ctl_ioctl+0x11/0x20 [dm_mod]\n[<0>] vfs_ioctl+0x21/0x60\n[<0>] __x64_sys_ioctl+0xb9/0xe0\n[<0>] do_syscall_64+0xc6/0x230\n[<0>] entry_SYSCALL_64_after_hwframe+0x6c/0x74\n\nMeanwhile mddev->recovery is:\nMD_RECOVERY_RUNNING |\nMD_RECOVERY_INTR |\nMD_RECOVERY_RESHAPE |\nMD_RECOVERY_FROZEN\n\nFix this problem by remove the code to register sync_thread directly\nfrom raid10 and raid5. And let md_check_recovery() to register\nsync_thread."
"value":"En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: md: No registre sync_thread para remodelar directamente Actualmente, si se interrumpe el proceso de remodelaci\u00f3n, volver a ensamblar la matriz registrar\u00e1 sync_thread directamente desde pers->run(), en este caso 'MD_RECOVERY_RUNNING ' se configura directamente, sin embargo, no hay garant\u00eda de que md_do_sync() se ejecute, por lo tanto, stop_sync_thread() se bloquear\u00e1 porque 'MD_RECOVERY_RUNNING' no se puede borrar. En el \u00faltimo parche, aseg\u00farese de que md_do_sync() establezca MD_RECOVERY_DONE; sin embargo, dm-raid test shell/lvconvert-raid-reshape.sh ocasionalmente puede activar el siguiente bloqueo: [root@fedora ~]# cat /proc/1982/stack [<0>] stop_sync_thread+0x1ab/0x270 [md_mod] [<0>] md_frozen_sync_thread+0x5c/0xa0 [md_mod] [<0>] raid_presuspend+0x1e/0x70 [dm_raid] [<0>] dm_table_presuspend_targets+0x40/0xb0 [ dm_mod] [<0>] __dm_destroy+0x2a5/0x310 [dm_mod] [<0>] dm_destroy+0x16/0x30 [dm_mod] [<0>] dev_remove+0x165/0x290 [dm_mod] [<0>] ctl_ioctl+0x4bb/ 0x7b0 [dm_mod] [<0>] dm_ctl_ioctl+0x11/0x20 [dm_mod] [<0>] vfs_ioctl+0x21/0x60 [<0>] __x64_sys_ioctl+0xb9/0xe0 [<0>] do_syscall_64+0xc6/0x230 [<0 >] Entry_SYSCALL_64_after_hwframe+0x6c/0x74 Mientras tanto mddev->recovery es: MD_RECOVERY_RUNNING | MD_RECOVERY_INTR | MD_RECOVERY_RESHAPE | MD_RECOVERY_FROZEN Solucione este problema eliminando el c\u00f3digo para registrar sync_thread directamente desde raid10 y raid5. Y deje que md_check_recovery() registre sync_thread."
