Commit 1cf897fc authored by Dexuan Cui's avatar Dexuan Cui Committed by Greg Kroah-Hartman

Drivers: hv: vmbus: finally fix hv_need_to_signal_on_read()

commit 433e19cf upstream.

Commit a389fcfd ("Drivers: hv: vmbus: Fix signaling logic in
hv_need_to_signal_on_read()")
added the proper mb(), but removed the test "prev_write_sz < pending_sz"
when making the signal decision.

As a result, the guest can signal the host unnecessarily,
and then the host can throttle the guest because the host
thinks the guest is buggy or malicious; finally the user
running stress test can perceive intermittent freeze of
the guest.

This patch brings back the test, and properly handles the
in-place consumption APIs used by NetVSC (see get_next_pkt_raw(),
put_pkt_raw() and commit_rd_index()).

Fixes: a389fcfd ("Drivers: hv: vmbus: Fix signaling logic in
hv_need_to_signal_on_read()")
Signed-off-by: default avatarDexuan Cui <decui@microsoft.com>
Reported-by: default avatarRolf Neugebauer <rolf.neugebauer@docker.com>
Tested-by: default avatarRolf Neugebauer <rolf.neugebauer@docker.com>
Cc: "K. Y. Srinivasan" <kys@microsoft.com>
Cc: Haiyang Zhang <haiyangz@microsoft.com>
Cc: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: default avatarK. Y. Srinivasan <kys@microsoft.com>
Cc: Rolf Neugebauer <rolf.neugebauer@docker.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
parent 964dfbe3
......@@ -383,6 +383,7 @@ int hv_ringbuffer_read(struct vmbus_channel *channel,
return ret;
}
init_cached_read_index(channel);
next_read_location = hv_get_next_read_location(inring_info);
next_read_location = hv_copyfrom_ringbuffer(inring_info, &desc,
sizeof(desc),
......
......@@ -1288,6 +1288,9 @@ void netvsc_channel_cb(void *context)
ndev = hv_get_drvdata(device);
buffer = get_per_channel_state(channel);
/* commit_rd_index() -> hv_signal_on_read() needs this. */
init_cached_read_index(channel);
do {
desc = get_next_pkt_raw(channel);
if (desc != NULL) {
......@@ -1340,6 +1343,9 @@ void netvsc_channel_cb(void *context)
bufferlen = bytes_recvd;
}
init_cached_read_index(channel);
} while (1);
if (bufferlen > NETVSC_PACKET_SIZE)
......
......@@ -128,6 +128,7 @@ struct hv_ring_buffer_info {
u32 ring_data_startoffset;
u32 priv_write_index;
u32 priv_read_index;
u32 cached_read_index;
};
/*
......@@ -180,6 +181,19 @@ static inline u32 hv_get_bytes_to_write(struct hv_ring_buffer_info *rbi)
return write;
}
static inline u32 hv_get_cached_bytes_to_write(
const struct hv_ring_buffer_info *rbi)
{
u32 read_loc, write_loc, dsize, write;
dsize = rbi->ring_datasize;
read_loc = rbi->cached_read_index;
write_loc = rbi->ring_buffer->write_index;
write = write_loc >= read_loc ? dsize - (write_loc - read_loc) :
read_loc - write_loc;
return write;
}
/*
* VMBUS version is 32 bit entity broken up into
* two 16 bit quantities: major_number. minor_number.
......@@ -1482,7 +1496,7 @@ hv_get_ring_buffer(struct hv_ring_buffer_info *ring_info)
static inline void hv_signal_on_read(struct vmbus_channel *channel)
{
u32 cur_write_sz;
u32 cur_write_sz, cached_write_sz;
u32 pending_sz;
struct hv_ring_buffer_info *rbi = &channel->inbound;
......@@ -1506,12 +1520,24 @@ static inline void hv_signal_on_read(struct vmbus_channel *channel)
cur_write_sz = hv_get_bytes_to_write(rbi);
if (cur_write_sz >= pending_sz)
if (cur_write_sz < pending_sz)
return;
cached_write_sz = hv_get_cached_bytes_to_write(rbi);
if (cached_write_sz < pending_sz)
vmbus_setevent(channel);
return;
}
static inline void
init_cached_read_index(struct vmbus_channel *channel)
{
struct hv_ring_buffer_info *rbi = &channel->inbound;
rbi->cached_read_index = rbi->ring_buffer->read_index;
}
/*
* An API to support in-place processing of incoming VMBUS packets.
*/
......@@ -1573,6 +1599,8 @@ static inline void put_pkt_raw(struct vmbus_channel *channel,
* This call commits the read index and potentially signals the host.
* Here is the pattern for using the "in-place" consumption APIs:
*
* init_cached_read_index();
*
* while (get_next_pkt_raw() {
* process the packet "in-place";
* put_pkt_raw();
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment