nfp: add async reconfiguration mechanism

Some callers of nfp_net_reconfig() are in atomic context so
we used to busy wait for commands to complete.  In worst case
scenario that means locking up a core for up to 5 seconds
when a command times out.  Lets add a timer-based mechanism
of asynchronously checking whether reconfiguration completed
successfully for atomic callers to use.  Non-atomic callers
can now just sleep.

The approach taken is quite simple because (1) synchronous
reconfigurations always happen under RTNL (or before device
is registered); (2) we can coalesce pending reconfigs.
There is no need for request queues, timer which eventually
takes a look at reconfiguration result to report errors is
good enough.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 10 insertions, 2 deletions

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index 3d53fcf323eb..e744acc18ef4 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -59,8 +59,8 @@
 			netdev_warn((nn)->netdev, fmt, ## args);	\
 	} while (0)
 
-/* Max time to wait for NFP to respond on updates (in ms) */
-#define NFP_NET_POLL_TIMEOUT	5000
+/* Max time to wait for NFP to respond on updates (in seconds) */
+#define NFP_NET_POLL_TIMEOUT	5
 
 /* Bar allocation */
 #define NFP_NET_CRTL_BAR	0
@@ -447,6 +447,10 @@ static inline bool nfp_net_fw_ver_eq(struct nfp_net_fw_version *fw_ver,
  * @shared_name:        Name for shared interrupt
  * @me_freq_mhz:        ME clock_freq (MHz)
  * @reconfig_lock:	Protects HW reconfiguration request regs/machinery
+ * @reconfig_posted:	Pending reconfig bits coming from async sources
+ * @reconfig_timer_active:  Timer for reading reconfiguration results is pending
+ * @reconfig_sync_present:  Some thread is performing synchronous reconfig
+ * @reconfig_timer:	Timer for async reading of reconfig results
  * @link_up:            Is the link up?
  * @link_status_lock:	Protects @link_up and ensures atomicity with BAR reading
  * @rx_coalesce_usecs:      RX interrupt moderation usecs delay parameter
@@ -531,6 +535,10 @@ struct nfp_net {
 	spinlock_t link_status_lock;
 
 	spinlock_t reconfig_lock;
+	u32 reconfig_posted;
+	bool reconfig_timer_active;
+	bool reconfig_sync_present;
+	struct timer_list reconfig_timer;
 
 	u32 rx_coalesce_usecs;
 	u32 rx_coalesce_max_frames;