netdev - [PATCH net-next v2 02/11] ice: retry acquiring hardware semaphore during cross-timestamp request

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: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20230918212814.435688-3-anthony.l.nguyen@intel.com>
Date: Mon, 18 Sep 2023 14:28:05 -0700
From: Tony Nguyen <anthony.l.nguyen@...el.com>
To: davem@...emloft.net,
	kuba@...nel.org,
	pabeni@...hat.com,
	edumazet@...gle.com,
	netdev@...r.kernel.org
Cc: Karol Kolacinski <karol.kolacinski@...el.com>,
	anthony.l.nguyen@...el.com,
	richardcochran@...il.com,
	Jacob Keller <jacob.e.keller@...el.com>,
	Pucha Himasekhar Reddy <himasekharx.reddy.pucha@...el.com>
Subject: [PATCH net-next v2 02/11] ice: retry acquiring hardware semaphore during cross-timestamp request

From: Karol Kolacinski <karol.kolacinski@...el.com>

The hardware for performing a cross-timestamp on E822 uses a hardware
semaphore which we must acquire before initiating the cross-timestamp
operation.

The current implementation only attempts to acquire the semaphore once, and
assumes that it will succeed. If the semaphore is busy for any reason, the
cross-timestamp operation fails with -EFAULT.

Instead of immediately failing, try the acquire the lock a few times with a
small sleep between attempts. This ensures that most requests will go
through without issue.

Additionally, return -EBUSY instead of -EFAULT if the operation can't
continue due to the semaphore being busy.

Signed-off-by: Karol Kolacinski <karol.kolacinski@...el.com>
Signed-off-by: Jacob Keller <jacob.e.keller@...el.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@...el.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@...el.com>
---
 drivers/net/ethernet/intel/ice/ice_ptp.c | 22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.c b/drivers/net/ethernet/intel/ice/ice_ptp.c
index 81d96a40d5a7..e75bb6e7d680 100644
--- a/drivers/net/ethernet/intel/ice/ice_ptp.c
+++ b/drivers/net/ethernet/intel/ice/ice_ptp.c
@@ -1976,11 +1976,21 @@ ice_ptp_get_syncdevicetime(ktime_t *device,
 	u32 hh_lock, hh_art_ctl;
 	int i;
 
-	/* Get the HW lock */
-	hh_lock = rd32(hw, PFHH_SEM + (PFTSYN_SEM_BYTES * hw->pf_id));
+#define MAX_HH_HW_LOCK_TRIES	5
+#define MAX_HH_CTL_LOCK_TRIES	100
+
+	for (i = 0; i < MAX_HH_HW_LOCK_TRIES; i++) {
+		/* Get the HW lock */
+		hh_lock = rd32(hw, PFHH_SEM + (PFTSYN_SEM_BYTES * hw->pf_id));
+		if (hh_lock & PFHH_SEM_BUSY_M) {
+			usleep_range(10000, 15000);
+			continue;
+		}
+		break;
+	}
 	if (hh_lock & PFHH_SEM_BUSY_M) {
 		dev_err(ice_pf_to_dev(pf), "PTP failed to get hh lock\n");
-		return -EFAULT;
+		return -EBUSY;
 	}
 
 	/* Start the ART and device clock sync sequence */
@@ -1988,9 +1998,7 @@ ice_ptp_get_syncdevicetime(ktime_t *device,
 	hh_art_ctl = hh_art_ctl | GLHH_ART_CTL_ACTIVE_M;
 	wr32(hw, GLHH_ART_CTL, hh_art_ctl);
 
-#define MAX_HH_LOCK_TRIES 100
-
-	for (i = 0; i < MAX_HH_LOCK_TRIES; i++) {
+	for (i = 0; i < MAX_HH_CTL_LOCK_TRIES; i++) {
 		/* Wait for sync to complete */
 		hh_art_ctl = rd32(hw, GLHH_ART_CTL);
 		if (hh_art_ctl & GLHH_ART_CTL_ACTIVE_M) {
@@ -2019,7 +2027,7 @@ ice_ptp_get_syncdevicetime(ktime_t *device,
 	hh_lock = hh_lock & ~PFHH_SEM_BUSY_M;
 	wr32(hw, PFHH_SEM + (PFTSYN_SEM_BYTES * hw->pf_id), hh_lock);
 
-	if (i == MAX_HH_LOCK_TRIES)
+	if (i == MAX_HH_CTL_LOCK_TRIES)
 		return -ETIMEDOUT;
 
 	return 0;
-- 
2.38.1