Date: Sun, 19 Jul 2026 18:50:52 +0000 From: Adrian Chadd <adrian@FreeBSD.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-main@FreeBSD.org Cc: =?utf-8?Q?J=C3=A9r=C3=A9mie?= Jourdin <jeremie@jourdin.org> Subject: git: b0c14eae1060 - main - re(4): add hw.re.aspm_disable loader tunable Message-ID: <6a5d1c8c.3f48b.65ba8348@gitrepo.freebsd.org>
index | next in thread | raw e-mail
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,home | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6a5d1c8c.3f48b.65ba8348>