git: b0c14eae1060 - main - re(4): add hw.re.aspm_disable loader tunable

The branch main has been updated by adrian:

URL: https://cgit.FreeBSD.org/src/commit/?id=b0c14eae10609f7d688b31d112de2a641d9cc516

commit b0c14eae10609f7d688b31d112de2a641d9cc516
Author:     Jérémie Jourdin <jeremie@jourdin.org>
AuthorDate: 2026-07-19 18:49:12 +0000
Commit:     Adrian Chadd <adrian@FreeBSD.org>
CommitDate: 2026-07-19 18:49:14 +0000

    re(4): add hw.re.aspm_disable loader tunable
    
    re(4) has unconditionally disabled ASPM L0s/L1 and CLKREQ at attach
    for years; on laptops this costs 200mW+ (requested by adrian@ in the PR).
    
    Make it a tunable following the existing hw.re.* pattern:
    
    * default 1 keeps today's behavior;
    * 0 preserves the firmware-configured ASPM state at attach and
      skips the watchdog re-assert from the previous revision.
    
    Documented in re.4.
    
    * Verified on RTL8168H (XID 0x541): with hw.re.aspm_disable=0,
      attach no longer logs "ASPM disabled" and pciconf -lcb shows
      the firmware Link Control state preserved -- including Clock PM,
      which the unconditional code previously cleared.
    
    * Default (1) is behaviorally identical to the current driver.
    
    * Note the tunable also stops the driver clearing CLKREQ, a small power
      win even where firmware leaves L0s/L1 off.
    
    Reviewed by:    adrian
    Differential Revision:  https://reviews.freebsd.org/D58280
    PR: kern/166724
---
 share/man/man4/re.4 | 11 ++++++++++-
 sys/dev/re/if_re.c  | 17 +++++++++++++----
 2 files changed, 23 insertions(+), 5 deletions(-)

diff --git a/share/man/man4/re.4 b/share/man/man4/re.4
index 0020af64b330..6d190c944b34 100644
--- a/share/man/man4/re.4
+++ b/share/man/man4/re.4
@@ -28,7 +28,7 @@
 .\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
 .\" THE POSSIBILITY OF SUCH DAMAGE.
 .\"
-.Dd June 9, 2026
+.Dd July 13, 2026
 .Dt RE 4
 .Os
 .Sh NAME
@@ -182,6 +182,15 @@ Tunables can be set at the
 prompt before booting the kernel or stored in
 .Xr loader.conf 5 .
 .Bl -tag -width "xxxxxx"
+.It Va hw.re.aspm_disable
+This tunable controls whether the driver disables PCI Express ASPM
+(the L0s and L1 link states and CLKREQ) on the device.
+ASPM is a known trigger of transmit stalls and watchdog timeouts on
+many supported controllers, so it is disabled by default.
+Set to 0 to keep the ASPM configuration programmed by the firmware,
+trading reliability under sustained transmit load for lower link
+power consumption.
+The default value is 1.
 .It Va hw.re.intr_filter
 This tunable makes driver use interrupt filter handler on
 controllers that support MSI/MSI-X capability.
diff --git a/sys/dev/re/if_re.c b/sys/dev/re/if_re.c
index 5a3e8e68d26f..5a1e8ad554cb 100644
--- a/sys/dev/re/if_re.c
+++ b/sys/dev/re/if_re.c
@@ -167,6 +167,8 @@ static int msix_disable = 0;
 TUNABLE_INT("hw.re.msix_disable", &msix_disable);
 static int prefer_iomap = 0;
 TUNABLE_INT("hw.re.prefer_iomap", &prefer_iomap);
+static int aspm_disable = 1;
+TUNABLE_INT("hw.re.aspm_disable", &aspm_disable);
 
 #define RE_CSUM_FEATURES    (CSUM_IP | CSUM_TCP | CSUM_UDP)
 
@@ -1369,8 +1371,14 @@ re_attach(device_t dev)
 		CSR_WRITE_1(sc, RL_EECMD, RL_EEMODE_OFF);
 	}
 
-	/* Disable ASPM L0S/L1 and CLKREQ. */
-	if (sc->rl_expcap != 0) {
+	/*
+	 * Disable ASPM L0S/L1 and CLKREQ.  ASPM is implicated in Tx
+	 * stalls and watchdog timeouts on many chip revisions, so it
+	 * is turned off by default; set the hw.re.aspm_disable tunable
+	 * to 0 to keep the firmware-configured ASPM state and trade
+	 * reliability under sustained load for link power management.
+	 */
+	if (aspm_disable != 0 && sc->rl_expcap != 0) {
 		cap = pci_read_config(dev, sc->rl_expcap +
 		    PCIER_LINK_CAP, 2);
 		if ((cap & PCIEM_LINK_CAP_ASPM) != 0) {
@@ -3658,11 +3666,12 @@ re_watchdog(struct rl_softc *sc)
 	}
 
 	/*
-	 * ASPM is a common cause of these Tx timeouts.  Make sure L0s/L1
+	 * ASPM is a common cause of these Tx timeouts.  Unless the
+	 * administrator has opted to keep ASPM enabled, make sure L0s/L1
 	 * and CLKREQ are still disabled on the link before reinitializing;
 	 * firmware or a power transition may have re-armed them.
 	 */
-	if (sc->rl_expcap != 0) {
+	if (aspm_disable != 0 && sc->rl_expcap != 0) {
 		uint16_t ctl;
 
 		ctl = pci_read_config(sc->rl_dev,