What about the following simplification. If on older hosts you can avoid more code
and the additional barrier. Also cleanup comment wording somewhat.
>From de2d566d9092cbbc8e5974dea581617ef787ff69 Mon Sep 17 00:00:00 2001
From: Michael Kelley <mhkelley@xxxxxxxxxxx>
Date: Sat, 10 Feb 2018 20:48:49 +0000
Subject: [PATCH 1/9] Drivers: hv: vmbus: Fix ring buffer signaling
Fix bugs in signaling the Hyper-V host when freeing space in the
host->guest ring buffer:
1. The interrupt_mask must not be used to determine whether to signal
on the host->guest ring buffer
2. The ring buffer write_index must be read (via hv_get_bytes_to_write)
*after* pending_send_sz is read in order to avoid a race condition
3. Comparisons with pending_send_sz must treat the "equals" case as
not-enough-space
4. Don't signal if the pending_send_sz feature is not present. Older
versions of Hyper-V that don't implement this feature will poll.
Fixes: 03bad714a161 ("vmbus: more host signalling avoidance")
Signed-off-by: Michael Kelley <mhkelley@xxxxxxxxxxx>
---
drivers/hv/ring_buffer.c | 67 +++++++++++++++++++++++++++++-------------------
1 file changed, 40 insertions(+), 27 deletions(-)
diff --git a/drivers/hv/ring_buffer.c b/drivers/hv/ring_buffer.c
index 50e071444a5c..d27cbb66279c 100644
--- a/drivers/hv/ring_buffer.c
+++ b/drivers/hv/ring_buffer.c
@@ -417,13 +417,25 @@ __hv_pkt_iter_next(struct vmbus_channel *channel,
}
EXPORT_SYMBOL_GPL(__hv_pkt_iter_next);
+/* How many bytes were read in this iterator cycle */
+static u32 hv_pkt_iter_bytes_read(const struct hv_ring_buffer_info *rbi)
+{
+ if (rbi->ring_buffer->read_index < rbi->priv_read_index)
+ return rbi->priv_read_index - rbi->ring_buffer->read_index;
+ else
+ return rbi->ring_datasize -
+ rbi->ring_buffer->read_index + rbi->priv_read_index;
+}
+
/*
* Update host ring buffer after iterating over packets.
*/
void hv_pkt_iter_close(struct vmbus_channel *channel)
{
struct hv_ring_buffer_info *rbi = &channel->inbound;
- u32 orig_write_sz = hv_get_bytes_to_write(rbi);
+ u32 curr_write_sz;
+ u32 delta = hv_pkt_iter_bytes_read(rbi);
+ u32 pending_sz;
/*
* Make sure all reads are done before we update the read index since
@@ -433,39 +445,40 @@ void hv_pkt_iter_close(struct vmbus_channel *channel)
virt_rmb();
rbi->ring_buffer->read_index = rbi->priv_read_index;
+ /* Older versions do not require signalling */
+ if (!rbi->ring_buffer->feature_bits.feat_pending_send_sz)
+ return;
+
/*
- * Issue a full memory barrier before making the signaling decision.
- * Here is the reason for having this barrier:
- * If the reading of the pend_sz (in this function)
- * were to be reordered and read before we commit the new read
- * index (in the calling function) we could
- * have a problem. If the host were to set the pending_sz after we
- * have sampled pending_sz and go to sleep before we commit the
- * read index, we could miss sending the interrupt. Issue a full
- * memory barrier to address this.
+ * If the reading of the pend_sz were to be reordered and read
+ * before we commit the new read index then we would have a
+ * problem. If the host were to set the pending_sz after we
+ * have sampled pending_sz and go to sleep before we commit
+ * the read index, we could miss sending the interrupt.
*/
virt_mb();
- /* If host has disabled notifications then skip */
- if (rbi->ring_buffer->interrupt_mask)
- return;
+ pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz);
- if (rbi->ring_buffer->feature_bits.feat_pending_send_sz) {
- u32 pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz);
+ /*
+ * Ensure the read of write_index in hv_get_bytes_to_write()
+ * happens after the read of pending_send_sz.
+ */
+ virt_rmb();
+ curr_write_sz = hv_get_bytes_to_write(rbi);
- /*
- * If there was space before we began iteration,
- * then host was not blocked. Also handles case where
- * pending_sz is zero then host has nothing pending
- * and does not need to be signaled.
- */
- if (orig_write_sz > pending_sz)
- return;
+ /*
+ * If there was space before we began iteration,
+ * then host was not blocked. Also handles case where
+ * pending_sz is zero then host has nothing pending and does
+ * not need to be signaled.
+ */
+ if (curr_write_sz - delta > pending_sz)
+ return;
- /* If pending write will not fit, don't give false hope. */
- if (hv_get_bytes_to_write(rbi) < pending_sz)
- return;
- }
+ /* If pending write will not fit, don't give false hope. */
+ if (curr_write_sz <= pending_sz)
+ return;
vmbus_setevent(channel);
}
--
2.15.1
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
