linux-kernel - Re: MSI interrupt for xhci still lost on 5.6-rc6 after cpu hotplug

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <87r1xjp3gn.fsf@nanos.tec.linutronix.de>
Date:   Mon, 23 Mar 2020 15:10:00 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Mathias Nyman <mathias.nyman@...ux.intel.com>, x86@...nel.org
Cc:     linux-pci <linux-pci@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Evan Green <evgreen@...omium.org>,
        "Ghorai\, Sukumar" <sukumar.ghorai@...el.com>,
        "Amara\, Madhusudanarao" <madhusudanarao.amara@...el.com>,
        "Nandamuri\, Srikanth" <srikanth.nandamuri@...el.com>
Subject: Re: MSI interrupt for xhci still lost on 5.6-rc6 after cpu hotplug

Mathias,

Mathias Nyman <mathias.nyman@...ux.intel.com> writes:
> Full function trace with patch is huge, can be found compressed at
>
> This time xhci interrupts stopped after
> migration/3-24    [003] d..1    48.530271: msi_set_affinity: twostep update msi, irq 122, vector 33 -> 34, apicid: 6 -> 4

thanks for providing the data. I think I decoded the issue. Can you
please test the patch below?

Thanks,

        tglx
----
diff --git a/arch/x86/kernel/apic/msi.c b/arch/x86/kernel/apic/msi.c
index 159bd0cb8548..49aec92fa35b 100644
--- a/arch/x86/kernel/apic/msi.c
+++ b/arch/x86/kernel/apic/msi.c
@@ -63,12 +63,13 @@ msi_set_affinity(struct irq_data *irqd, const struct cpumask *mask, bool force)
 {
 	struct irq_cfg old_cfg, *cfg = irqd_cfg(irqd);
 	struct irq_data *parent = irqd->parent_data;
-	unsigned int cpu;
+	unsigned int cpu, old_vector;
 	int ret;
 
 	/* Save the current configuration */
 	cpu = cpumask_first(irq_data_get_effective_affinity_mask(irqd));
 	old_cfg = *cfg;
+	old_vector = old_cfg.vector;
 
 	/* Allocate a new target vector */
 	ret = parent->chip->irq_set_affinity(parent, mask, force);
@@ -90,10 +91,10 @@ msi_set_affinity(struct irq_data *irqd, const struct cpumask *mask, bool force)
 	 */
 	if (!irqd_msi_nomask_quirk(irqd) ||
 	    cfg->vector == old_cfg.vector ||
-	    old_cfg.vector == MANAGED_IRQ_SHUTDOWN_VECTOR ||
+	    old_vector == MANAGED_IRQ_SHUTDOWN_VECTOR ||
 	    cfg->dest_apicid == old_cfg.dest_apicid) {
 		irq_msi_update_msg(irqd, cfg);
-		return ret;
+		goto check_old;
 	}
 
 	/*
@@ -102,7 +103,7 @@ msi_set_affinity(struct irq_data *irqd, const struct cpumask *mask, bool force)
 	 */
 	if (WARN_ON_ONCE(cpu != smp_processor_id())) {
 		irq_msi_update_msg(irqd, cfg);
-		return ret;
+		goto check_old;
 	}
 
 	/*
@@ -163,6 +164,11 @@ msi_set_affinity(struct irq_data *irqd, const struct cpumask *mask, bool force)
 	if (lapic_vector_set_in_irr(cfg->vector))
 		irq_data_get_irq_chip(irqd)->irq_retrigger(irqd);
 
+check_old:
+	if (!in_irq() && old_vector != MANAGED_IRQ_SHUTDOWN_VECTOR &&
+	    irqd_msi_nomask_quirk(irqd) && lapic_vector_set_in_irr(old_vector))
+		irq_data_get_irq_chip(irqd)->irq_retrigger(irqd);
+
 	return ret;
 }