CVE-2023-52587

En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: IB/ipoib: corrige el bloqueo de la lista mcast Al liberar `priv->lock` mientras se itera `priv->multicast_list` en `ipoib_mcast_join_task()`, se abre una ventana para `ipoib_mcast_dev_flush( )` para eliminar los elementos mientras se encuentra en medio de la iteración. Si se elimina mcast mientras se elimina el bloqueo, el bucle for gira para siempre, lo que resulta en un bloqueo total (como se informó en el kernel RHEL 4.18.0-372.75.1.el8_6): Tarea A (kworker/u72:2 a continuación) | Tarea B (kworker/u72:0 a continuación) -----------------------------------+---- ------------------------------- ipoib_mcast_join_task(trabajo) | ipoib_ib_dev_flush_light(trabajo) spin_lock_irq(&priv->lock) | __ipoib_ib_dev_flush(priv, ...) list_for_each_entry(mcast, | ipoib_mcast_dev_flush(dev = priv->dev) &priv->multicast_list, lista) | ipoib_mcast_join(dev, mcast) | spin_unlock_irq(&priv->bloquear) | | spin_lock_irqsave(&priv->bloqueo, banderas) | list_for_each_entry_safe(mcast, tmcast, | &priv->multicast_list, lista) | list_del(&mcast->lista); | list_add_tail(&mcast->lista, &remove_list) | spin_unlock_irqrestore(&priv->bloquear, banderas) spin_lock_irq(&priv->bloquear) | | ipoib_mcast_remove_list(&remove_list) (Aquí, `mcast` ya no está en | list_for_each_entry_safe(mcast, tmcast, `priv->multicast_list` y seguimos | remove_list, list) girando en `remove_list` de | >>> wait_for_completion(&mcast ->hecho) el otro hilo que está bloqueado | y la lista sigue siendo válida | en su pila.) Solucione este problema manteniendo el bloqueo mantenido y cambiando a GFP_ATOMIC para evitar eventuales suspensiones. Desafortunadamente, no pudimos reproducir el bloqueo y confirmar esta solución, pero según la revisión del código, creo que esta solución debería abordar dichos bloqueos. crash> bc 31 PID: 747 TAREA: ff1c6a1a007e8000 CPU: 31 COMANDO: "kworker/u72:2" -- [excepción RIP: ipoib_mcast_join_task+0x1b1] RIP: ffffffffc0944ac1 RSP: ff646f199a8c7e00 RFLAGS: 000000 02 RAX: 0000000000000000 RBX: ff1c6a1a04dc82f8 RCX: 00000000000000000 trabajo (&priv->mcast_task{,.work}) RDX: ff1c6a192d60ac68 RSI: 0000000000000286 RDI: ff1c6a1a04dc8000 &mcast->list RBP: ff646f199a8c7e90 R8: ff1c699980019420 R9: ff 1c6a1920c9a000 R10: ff646f199a8c7e00 R11: ff1c6a191a7d9800 R12: ff1c6a192d60ac00 mcast R13: ff1c6a1d82200000 R14: ff1c6a1a04dc8000 R15: ff1c6a1a04dc82d8 dev priv (&priv->lock) &priv->multicast_list (también conocido como head) ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018 --- --- #5 [ff646f199a8c7e00] ipoib_mcast_join_task+0x1b1 en ffffff ffc0944ac1 [ ib_ipoib] #6 [ff646f199a8c7e98] Process_one_work+0x1a7 en ffffffff9bf10967 crash> rx ff646f199a8c7e68 ff646f199a8c7e68: ff1c6a1a04dc82f8 <<< work = &priv->mcast_task.work crash> lista -hO ipoib_dev_priv.multicast_list ff1c6a1a04dc8000 (vacío) falla> ipoib_dev_priv.mcast_task.work. func,mcast_mutex.owner.counter ff1c6a1a04dc8000 mcast_task.work.func = 0xffffffffc0944910 , mcast_mutex.owner.counter = 0xff1c69998efec000 crash> b 8 PID: 8 TAREA: ff1c69998efec000 CPU: 33 COMANDO: "kworker/u72:0" -- #3 [ff646f1980153d50] wait_for_completion+0x96 en ffffffff9c7d7646 #4 [ff646f1980153d90] ipoib_mcast_remove_list+0x56 en ffffffffc0944dc6 [ib_ipoib] #5 [ff646f1980153de8] ipoib_ mcast_dev_flush+0x1a7 en ffffffffc09455a7 [ib_ipoib] #6 [ff646f1980153e58] __ipoib_ib_dev_flush+0x1a4 en ffffffffc09431a4 [ib_ipoib] # 7 [ff ---truncado---

In the Linux kernel, the following vulnerability has been resolved: IB/ipoib: Fix mcast list locking Releasing the `priv->lock` while iterating the `priv->multicast_list` in `ipoib_mcast_join_task()` opens a window for `ipoib_mcast_dev_flush()` to remove the items while in the middle of iteration. If the mcast is removed while the lock was dropped, the for loop spins forever resulting in a hard lockup (as was reported on RHEL 4.18.0-372.75.1.el8_6 kernel): Task A (kworker/u72:2 below) | Task B (kworker/u72:0 below) -----------------------------------+----------------------------------- ipoib_mcast_join_task(work) | ipoib_ib_dev_flush_light(work) spin_lock_irq(&priv->lock) | __ipoib_ib_dev_flush(priv, ...) list_for_each_entry(mcast, | ipoib_mcast_dev_flush(dev = priv->dev) &priv->multicast_list, list) | ipoib_mcast_join(dev, mcast) | spin_unlock_irq(&priv->lock) | | spin_lock_irqsave(&priv->lock, flags) | list_for_each_entry_safe(mcast, tmcast, | &priv->multicast_list, list) | list_del(&mcast->list); | list_add_tail(&mcast->list, &remove_list) | spin_unlock_irqrestore(&priv->lock, flags) spin_lock_irq(&priv->lock) | | ipoib_mcast_remove_list(&remove_list) (Here, `mcast` is no longer on the | list_for_each_entry_safe(mcast, tmcast, `priv->multicast_list` and we keep | remove_list, list) spinning on the `remove_list` of | >>> wait_for_completion(&mcast->done) the other thread which is blocked | and the list is still valid on | it's stack.) Fix this by keeping the lock held and changing to GFP_ATOMIC to prevent eventual sleeps. Unfortunately we could not reproduce the lockup and confirm this fix but based on the code review I think this fix should address such lockups. crash> bc 31 PID: 747 TASK: ff1c6a1a007e8000 CPU: 31 COMMAND: "kworker/u72:2" -- [exception RIP: ipoib_mcast_join_task+0x1b1] RIP: ffffffffc0944ac1 RSP: ff646f199a8c7e00 RFLAGS: 00000002 RAX: 0000000000000000 RBX: ff1c6a1a04dc82f8 RCX: 0000000000000000 work (&priv->mcast_task{,.work}) RDX: ff1c6a192d60ac68 RSI: 0000000000000286 RDI: ff1c6a1a04dc8000 &mcast->list RBP: ff646f199a8c7e90 R8: ff1c699980019420 R9: ff1c6a1920c9a000 R10: ff646f199a8c7e00 R11: ff1c6a191a7d9800 R12: ff1c6a192d60ac00 mcast R13: ff1c6a1d82200000 R14: ff1c6a1a04dc8000 R15: ff1c6a1a04dc82d8 dev priv (&priv->lock) &priv->multicast_list (aka head) ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018 --- <NMI exception stack> --- #5 [ff646f199a8c7e00] ipoib_mcast_join_task+0x1b1 at ffffffffc0944ac1 [ib_ipoib] #6 [ff646f199a8c7e98] process_one_work+0x1a7 at ffffffff9bf10967 crash> rx ff646f199a8c7e68 ff646f199a8c7e68: ff1c6a1a04dc82f8 <<< work = &priv->mcast_task.work crash> list -hO ipoib_dev_priv.multicast_list ff1c6a1a04dc8000 (empty) crash> ipoib_dev_priv.mcast_task.work.func,mcast_mutex.owner.counter ff1c6a1a04dc8000 mcast_task.work.func = 0xffffffffc0944910 <ipoib_mcast_join_task>, mcast_mutex.owner.counter = 0xff1c69998efec000 crash> b 8 PID: 8 TASK: ff1c69998efec000 CPU: 33 COMMAND: "kworker/u72:0" -- #3 [ff646f1980153d50] wait_for_completion+0x96 at ffffffff9c7d7646 #4 [ff646f1980153d90] ipoib_mcast_remove_list+0x56 at ffffffffc0944dc6 [ib_ipoib] #5 [ff646f1980153de8] ipoib_mcast_dev_flush+0x1a7 at ffffffffc09455a7 [ib_ipoib] #6 [ff646f1980153e58] __ipoib_ib_dev_flush+0x1a4 at ffffffffc09431a4 [ib_ipoib] #7 [ff ---truncated---

Alon Zahavi discovered that the NVMe-oF/TCP subsystem in the Linux kernel did not properly validate H2C PDU data, leading to a null pointer dereference vulnerability. A remote attacker could use this to cause a denial of service. Sander Wiebing, Alvise de Faveri Tron, Herbert Bos, and Cristiano Giuffrida discovered that the Linux kernel mitigations for the initial Branch History Injection vulnerability were insufficient for Intel processors. A local attacker could potentially use this to expose sensitive information.