From: Marc Zyngier <maz@kernel.org>
To: linux-kernel@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>,
Jason Cooper <jason@lakedaemon.net>,
lorenzo.pieralisi@arm.com, Andrew.Murray@arm.com,
yuzenghui@huawei.com, Heyi Guo <guoheyi@huawei.com>
Subject: [PATCH 09/11] irqchip/gic-v3-its: Synchronise INT/CLEAR commands targetting a VLPI using VSYNC
Date: Tue, 5 Nov 2019 16:22:56 +0000 [thread overview]
Message-ID: <20191105162258.22214-10-maz@kernel.org> (raw)
In-Reply-To: <20191105162258.22214-1-maz@kernel.org>
We have so far always injected/cleared VLPIs using either
INT+SYNC or CLEAR+SYNC sequences, but that's pretty wrong
for two reasons:
- SYNC only synchronises physical LPIs
- The collection ID that for the associated LPI doesn't match
the redistributor the vPE is associated with
Instead, send an {INT,CLEAR}+VSYNC for forwarded LPIs, ensuring
that the ITS synchronises against the virtual pending table.
Reported-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
---
drivers/irqchip/irq-gic-v3-its.c | 79 ++++++++++++++++++++++++++++++--
1 file changed, 75 insertions(+), 4 deletions(-)
diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index d4b9f12e2e53..dbb994f52e42 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -718,6 +718,42 @@ static struct its_vpe *its_build_vinv_cmd(struct its_node *its,
return valid_vpe(its, map->vpe);
}
+static struct its_vpe *its_build_vint_cmd(struct its_node *its,
+ struct its_cmd_block *cmd,
+ struct its_cmd_desc *desc)
+{
+ struct its_vlpi_map *map;
+
+ map = dev_event_to_vlpi_map(desc->its_int_cmd.dev,
+ desc->its_int_cmd.event_id);
+
+ its_encode_cmd(cmd, GITS_CMD_INT);
+ its_encode_devid(cmd, desc->its_int_cmd.dev->device_id);
+ its_encode_event_id(cmd, desc->its_int_cmd.event_id);
+
+ its_fixup_cmd(cmd);
+
+ return valid_vpe(its, map->vpe);
+}
+
+static struct its_vpe *its_build_vclear_cmd(struct its_node *its,
+ struct its_cmd_block *cmd,
+ struct its_cmd_desc *desc)
+{
+ struct its_vlpi_map *map;
+
+ map = dev_event_to_vlpi_map(desc->its_clear_cmd.dev,
+ desc->its_clear_cmd.event_id);
+
+ its_encode_cmd(cmd, GITS_CMD_CLEAR);
+ its_encode_devid(cmd, desc->its_clear_cmd.dev->device_id);
+ its_encode_event_id(cmd, desc->its_clear_cmd.event_id);
+
+ its_fixup_cmd(cmd);
+
+ return valid_vpe(its, map->vpe);
+}
+
static u64 its_cmd_ptr_to_offset(struct its_node *its,
struct its_cmd_block *ptr)
{
@@ -1098,6 +1134,34 @@ static void its_send_vinv(struct its_device *dev, u32 event_id)
its_send_single_vcommand(dev->its, its_build_vinv_cmd, &desc);
}
+static void its_send_vint(struct its_device *dev, u32 event_id)
+{
+ struct its_cmd_desc desc;
+
+ /*
+ * There is no real VINT command. This is just a normal INT,
+ * with a VSYNC instead of a SYNC.
+ */
+ desc.its_int_cmd.dev = dev;
+ desc.its_int_cmd.event_id = event_id;
+
+ its_send_single_vcommand(dev->its, its_build_vint_cmd, &desc);
+}
+
+static void its_send_vclear(struct its_device *dev, u32 event_id)
+{
+ struct its_cmd_desc desc;
+
+ /*
+ * There is no real VCLEAR command. This is just a normal CLEAR,
+ * with a VSYNC instead of a SYNC.
+ */
+ desc.its_clear_cmd.dev = dev;
+ desc.its_clear_cmd.event_id = event_id;
+
+ its_send_single_vcommand(dev->its, its_build_vclear_cmd, &desc);
+}
+
/*
* irqchip functions - assumes MSI, mostly.
*/
@@ -1291,10 +1355,17 @@ static int its_irq_set_irqchip_state(struct irq_data *d,
if (which != IRQCHIP_STATE_PENDING)
return -EINVAL;
- if (state)
- its_send_int(its_dev, event);
- else
- its_send_clear(its_dev, event);
+ if (irqd_is_forwarded_to_vcpu(d)) {
+ if (state)
+ its_send_vint(its_dev, event);
+ else
+ its_send_vclear(its_dev, event);
+ } else {
+ if (state)
+ its_send_int(its_dev, event);
+ else
+ its_send_clear(its_dev, event);
+ }
return 0;
}
--
2.20.1
next prev parent reply other threads:[~2019-11-05 16:23 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-05 16:22 [PATCH 00/11] irqchip/gic-v3-its: Cleanup and fixes for Linux 5.5 Marc Zyngier
2019-11-05 16:22 ` [PATCH 01/11] irqchip/gic-v3-its: Free collection mapping on device teardown Marc Zyngier
2019-11-08 13:00 ` Zenghui Yu
2019-11-08 15:25 ` Marc Zyngier
2019-11-05 16:22 ` [PATCH 02/11] irqchip/gic-v3-its: Factor out wait_for_syncr primitive Marc Zyngier
2019-11-05 16:22 ` [PATCH 03/11] irqchip/gic-v3-its: Allow LPI invalidation via the DirectLPI interface Marc Zyngier
2019-11-05 16:22 ` [PATCH 04/11] irqchip/gic-v3-its: Make is_v4 use a TYPER copy Marc Zyngier
2019-11-05 16:22 ` [PATCH 05/11] irqchip/gic-v3-its: Kill its->ite_size and use TYPER copy instead Marc Zyngier
2019-11-05 16:22 ` [PATCH 06/11] irqchip/gic-v3-its: Kill its->device_ids " Marc Zyngier
2019-11-05 16:22 ` [PATCH 07/11] irqchip/gic-v3-its: Add its_vlpi_map helpers Marc Zyngier
2019-11-05 16:22 ` [PATCH 08/11] irqchip/gic-v3-its: Synchronise INV command targetting a VLPI using VSYNC Marc Zyngier
2019-11-05 16:22 ` Marc Zyngier [this message]
2019-11-05 16:22 ` [PATCH 10/11] irqchip/gic-v3-its: Lock VLPI map array before translating it Marc Zyngier
2019-11-05 16:22 ` [PATCH 11/11] irqchip/gic-v3-its: Make vlpi_lock a spinlock Marc Zyngier
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191105162258.22214-10-maz@kernel.org \
--to=maz@kernel.org \
--cc=Andrew.Murray@arm.com \
--cc=guoheyi@huawei.com \
--cc=jason@lakedaemon.net \
--cc=linux-kernel@vger.kernel.org \
--cc=lorenzo.pieralisi@arm.com \
--cc=tglx@linutronix.de \
--cc=yuzenghui@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).