linux-kernel - [PATCH 4.9 38/44] irqchip/gic-v3-its: Align PCI Multi-MSI allocation on their size

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: <20190129113143.037142661@linuxfoundation.org>
Date:   Tue, 29 Jan 2019 12:36:33 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        Marc Zyngier <marc.zyngier@....com>
Subject: [PATCH 4.9 38/44] irqchip/gic-v3-its: Align PCI Multi-MSI allocation on their size

4.9-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Marc Zyngier <marc.zyngier@....com>

commit 8208d1708b88b412ca97f50a6d951242c88cbbac upstream.

The way we allocate events works fine in most cases, except
when multiple PCI devices share an ITS-visible DevID, and that
one of them is trying to use MultiMSI allocation.

In that case, our allocation is not guaranteed to be zero-based
anymore, and we have to make sure we allocate it on a boundary
that is compatible with the PCI Multi-MSI constraints.

Fix this by allocating the full region upfront instead of iterating
over the number of MSIs. MSI-X are always allocated one by one,
so this shouldn't change anything on that front.

Fixes: b48ac83d6bbc2 ("irqchip: GICv3: ITS: MSI support")
Cc: stable@...r.kernel.org
Reported-by: Ard Biesheuvel <ard.biesheuvel@...aro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@...aro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@....com>
[ardb: rebased onto v4.9.153, should apply cleanly onto v4.4.y as well]
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@...aro.org>

---
 drivers/irqchip/irq-gic-v3-its.c |   25 +++++++++++++------------
 1 file changed, 13 insertions(+), 12 deletions(-)

--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -1372,13 +1372,14 @@ static void its_free_device(struct its_d
 	kfree(its_dev);
 }
 
-static int its_alloc_device_irq(struct its_device *dev, irq_hw_number_t *hwirq)
+static int its_alloc_device_irq(struct its_device *dev, int nvecs, irq_hw_number_t *hwirq)
 {
 	int idx;
 
-	idx = find_first_zero_bit(dev->event_map.lpi_map,
-				  dev->event_map.nr_lpis);
-	if (idx == dev->event_map.nr_lpis)
+	idx = bitmap_find_free_region(dev->event_map.lpi_map,
+				      dev->event_map.nr_lpis,
+				      get_count_order(nvecs));
+	if (idx < 0)
 		return -ENOSPC;
 
 	*hwirq = dev->event_map.lpi_base + idx;
@@ -1464,20 +1465,20 @@ static int its_irq_domain_alloc(struct i
 	int err;
 	int i;
 
-	for (i = 0; i < nr_irqs; i++) {
-		err = its_alloc_device_irq(its_dev, &hwirq);
-		if (err)
-			return err;
+	err = its_alloc_device_irq(its_dev, nr_irqs, &hwirq);
+	if (err)
+		return err;
 
-		err = its_irq_gic_domain_alloc(domain, virq + i, hwirq);
+	for (i = 0; i < nr_irqs; i++) {
+		err = its_irq_gic_domain_alloc(domain, virq + i, hwirq + i);
 		if (err)
 			return err;
 
 		irq_domain_set_hwirq_and_chip(domain, virq + i,
-					      hwirq, &its_irq_chip, its_dev);
+					      hwirq + i, &its_irq_chip, its_dev);
 		pr_debug("ID:%d pID:%d vID:%d\n",
-			 (int)(hwirq - its_dev->event_map.lpi_base),
-			 (int) hwirq, virq + i);
+			 (int)(hwirq + i - its_dev->event_map.lpi_base),
+			 (int)(hwirq + i), virq + i);
 	}
 
 	return 0;