"value":"In the Linux kernel, the following vulnerability has been resolved:\n\nice: fix 'scheduling while atomic' on aux critical err interrupt\n\nThere's a kernel BUG splat on processing aux critical error\ninterrupts in ice_misc_intr():\n\n[ 2100.917085] BUG: scheduling while atomic: swapper/15/0/0x00010000\n...\n[ 2101.060770] Call Trace:\n[ 2101.063229] <IRQ>\n[ 2101.065252] dump_stack+0x41/0x60\n[ 2101.068587] __schedule_bug.cold.100+0x4c/0x58\n[ 2101.073060] __schedule+0x6a4/0x830\n[ 2101.076570] schedule+0x35/0xa0\n[ 2101.079727] schedule_preempt_disabled+0xa/0x10\n[ 2101.084284] __mutex_lock.isra.7+0x310/0x420\n[ 2101.088580] ? ice_misc_intr+0x201/0x2e0 [ice]\n[ 2101.093078] ice_send_event_to_aux+0x25/0x70 [ice]\n[ 2101.097921] ice_misc_intr+0x220/0x2e0 [ice]\n[ 2101.102232] __handle_irq_event_percpu+0x40/0x180\n[ 2101.106965] handle_irq_event_percpu+0x30/0x80\n[ 2101.111434] handle_irq_event+0x36/0x53\n[ 2101.115292] handle_edge_irq+0x82/0x190\n[ 2101.119148] handle_irq+0x1c/0x30\n[ 2101.122480] do_IRQ+0x49/0xd0\n[ 2101.125465] common_interrupt+0xf/0xf\n[ 2101.129146] </IRQ>\n...\n\nAs Andrew correctly mentioned previously[0], the following call\nladder happens:\n\nice_misc_intr() <- hardirq\n ice_send_event_to_aux()\n device_lock()\n mutex_lock()\n might_sleep()\n might_resched() <- oops\n\nAdd a new PF state bit which indicates that an aux critical error\noccurred and serve it in ice_service_task() in process context.\nThe new ice_pf::oicr_err_reg is read-write in both hardirq and\nprocess contexts, but only 3 bits of non-critical data probably\naren't worth explicit synchronizing (and they're even in the same\nbyte [31:24]).\n\n[0] https://lore.kernel.org/all/YeSRUVmrdmlUXHDn@lunn.ch"
"value":"En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: ice: fix 'scheduling while atomic' on aux critical err interrupci\u00f3n Hay un ERROR del kernel en el procesamiento de interrupciones de error cr\u00edtico auxiliar en ice_misc_intr(): [ 2100.917085] ERROR: scheduling while atomic: swapper/15/0/0x00010000 ... [ 2101.060770] Seguimiento de llamadas: [ 2101.063229] [ 2101.065252] dump_stack+0x41/0x60 [ 2101.068587] __schedule_bug.cold.100+0x4c/0x58 [ 2101.073060] __schedule+0x6a4/0x830 [ [2101.076570] schedule+0x35/0xa0 [2101.079727] schedule_preempt_disabled+0xa/0x10 [2101.084284] __mutex_lock.isra.7+0x310/0x420 [2101.088580] ? ice_misc_intr+0x201/0x2e0 [hielo] [ 2101.093078] ice_send_event_to_aux+0x25/0x70 [hielo] [ 2101.097921] ice_misc_intr+0x220/0x2e0 [hielo] [ 2101.102232] __handle_irq_event_percpu+0x40/0x180 [ 2101.106965] handle_irq_event_percpu+0x30/0x80 [ 2101.111434] handle_irq_event+0x36/0x53 [ 2101.115292] handle_edge_irq+0x82/0x190 [ 2101.119148] handle_irq+0x1c/0x30 [ 2101.122480] do_IRQ+0x49/0xd0 [ 2101.125465] common_interrupt+0xf/0xf [ 2101.129146] ... Como Andrew mencion\u00f3 correctamente anteriormente[0], ocurre la siguiente escalera de llamadas: ice_misc_intr() <- hardirq ice_send_event_to_aux() device_lock() mutex_lock() might_sleep() might_resched() <- oops Agregue un nuevo bit de estado de PF que indique que ocurri\u00f3 un error cr\u00edtico auxiliar y sirvalo en ice_service_task() en el contexto del proceso. El nuevo ice_pf::oicr_err_reg es de lectura y escritura tanto en contextos de hardirq como de proceso, pero solo 3 bits de datos no cr\u00edticos probablemente no merezcan una sincronizaci\u00f3n expl\u00edcita (e incluso est\u00e1n en el mismo byte [31:24]). [0] https://lore.kernel.org/all/YeSRUVmrdmlUXHDn@lunn.ch"