Bug 248159
| Summary: | mpr(4): Unable to suspend to ACPI S3 state | ||
|---|---|---|---|
| Product: | Base System | Reporter: | Joshua Kinard <freebsd> |
| Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> |
| Status: | Open --- | ||
| Severity: | Affects Only Me | CC: | emaste, scottl |
| Priority: | --- | ||
| Version: | 12.1-RELEASE | ||
| Hardware: | amd64 | ||
| OS: | Any | ||
Presumably we can remove "on Ryzen platform" from the title - this issue is an mpr(4) limitation regardless of platform. (In reply to Joshua Kinard from comment #0) This code appears to have come in with the initial commit in 2014. I'm going to Cc: scottl but only because it seems he is the only one actively working on underlying issues like this in this driver. scottl, feel free to remove yourself from the Cc: if my guess here is wrong. |
It looks like the mpr(4) driver w/ an LSI/AVAGO/Broadcom SAS9300-8i won't allow an AMD Ryzen system to enter ACPI S3 state. I have on system running Ryzen on an ASUS PRIME B450M-A/CSM motherboard, and building a second system using a Gigabyte B450M DS3H motherboard. Both systems have a SAS9300-8i card in them, and I finally pinned down that removing the SAS card allowed 'acpiconf -s 3' to work and put the system to sleep. Issue is reproducible on 12.1-RELEASE-p7, 12.1-STABLE (r361761), and 13-CURRENT (r361779). The only (somewhat) usable error message is "acpi0: device_suspend failed", followed by the system running resume functions (mostly bringing the USB ports back online). Doing some looking around, it appears the culprit is....by design: /usr/src/sys/dev/mpr/mpr_pci.c:454 static int mpr_pci_suspend(device_t dev) { return (EINVAL); } Are there plans to implement mpr_pci_suspend so that this device can be put into S3 state? are here any SAS/SATA HBA drivers in FreeBSD that support suspend/resume?