linux-kernel - [irqchip: irq/irqchip-next] irqchip/gic-v4: Wait for GICR_VPENDBASER.Dirty to clear before descheduling

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Mon, 21 Mar 2022 08:59:40 -0000
From:   "irqchip-bot for Marc Zyngier" <tip-bot2@...utronix.de>
To:     linux-kernel@...r.kernel.org
Cc:     Jingyi Wang <wangjingyi11@...wei.com>,
        Nianyao Tang <tangnianyao@...wei.com>,
        Marc Zyngier <maz@...nel.org>, tglx@...utronix.de
Subject: [irqchip: irq/irqchip-next] irqchip/gic-v4: Wait for
 GICR_VPENDBASER.Dirty to clear before descheduling

The following commit has been merged into the irq/irqchip-next branch of irqchip:

Commit-ID:     686121ebe6fce266f16978e98ecda35a72c39dbf
Gitweb:        https://git.kernel.org/pub/scm/linux/kernel/git/maz/arm-platforms/686121ebe6fce266f16978e98ecda35a72c39dbf
Author:        Marc Zyngier <maz@...nel.org>
AuthorDate:    Thu, 17 Mar 2022 09:49:02 
Committer:     Marc Zyngier <maz@...nel.org>
CommitterDate: Mon, 21 Mar 2022 08:53:05 

irqchip/gic-v4: Wait for GICR_VPENDBASER.Dirty to clear before descheduling

The way KVM drives GICv4.{0,1} is as follows:
- vcpu_load() makes the VPE resident, instructing the RD to start
  scanning for interrupts
- just before entering the guest, we check that the RD has finished
  scanning and that we can start running the vcpu
- on preemption, we deschedule the VPE by making it invalid on
  the RD

However, we are preemptible between the first two steps. If it so
happens *and* that the RD was still scanning, we nonetheless write
to the GICR_VPENDBASER register while Dirty is set, and bad things
happen (we're in UNPRED land).

This affects both the 4.0 and 4.1 implementations.

Make sure Dirty is cleared before performing the deschedule,
meaning that its_clear_vpend_valid() becomes a sort of full VPE
residency barrier.

Reported-by: Jingyi Wang <wangjingyi11@...wei.com>
Tested-by: Nianyao Tang <tangnianyao@...wei.com>
Signed-off-by: Marc Zyngier <maz@...nel.org>
Fixes: 57e3cebd022f ("KVM: arm64: Delay the polling of the GICR_VPENDBASER.Dirty
bit")
Link: https://lore.kernel.org/r/4aae10ba-b39a-5f84-754b-69c2eb0a2c03@huawei.com
---
 drivers/irqchip/irq-gic-v3-its.c | 28 +++++++++++++++++++---------
 1 file changed, 19 insertions(+), 9 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 9e93ff2..c9b1df9 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -3011,18 +3011,12 @@ static int __init allocate_lpi_tables(void)
 	return 0;
 }
 
-static u64 its_clear_vpend_valid(void __iomem *vlpi_base, u64 clr, u64 set)
+static u64 read_vpend_dirty_clear(void __iomem *vlpi_base)
 {
 	u32 count = 1000000;	/* 1s! */
 	bool clean;
 	u64 val;
 
-	val = gicr_read_vpendbaser(vlpi_base + GICR_VPENDBASER);
-	val &= ~GICR_VPENDBASER_Valid;
-	val &= ~clr;
-	val |= set;
-	gicr_write_vpendbaser(val, vlpi_base + GICR_VPENDBASER);
-
 	do {
 		val = gicr_read_vpendbaser(vlpi_base + GICR_VPENDBASER);
 		clean = !(val & GICR_VPENDBASER_Dirty);
@@ -3033,10 +3027,26 @@ static u64 its_clear_vpend_valid(void __iomem *vlpi_base, u64 clr, u64 set)
 		}
 	} while (!clean && count);
 
-	if (unlikely(val & GICR_VPENDBASER_Dirty)) {
+	if (unlikely(!clean))
 		pr_err_ratelimited("ITS virtual pending table not cleaning\n");
+
+	return val;
+}
+
+static u64 its_clear_vpend_valid(void __iomem *vlpi_base, u64 clr, u64 set)
+{
+	u64 val;
+
+	/* Make sure we wait until the RD is done with the initial scan */
+	val = read_vpend_dirty_clear(vlpi_base);
+	val &= ~GICR_VPENDBASER_Valid;
+	val &= ~clr;
+	val |= set;
+	gicr_write_vpendbaser(val, vlpi_base + GICR_VPENDBASER);
+
+	val = read_vpend_dirty_clear(vlpi_base);
+	if (unlikely(val & GICR_VPENDBASER_Dirty))
 		val |= GICR_VPENDBASER_PendingLast;
-	}
 
 	return val;
 }