Date: Mon, 25 Feb 2019 15:17:52 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 235944] jedec_dimm(4) does not attach to KFA2 (aka Galax) Hall of Fame DDR4 sticks Message-ID: <bug-235944-227-UBLALeElEU@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-235944-227@https.bugs.freebsd.org/bugzilla/> References: <bug-235944-227@https.bugs.freebsd.org/bugzilla/>
index | next in thread | previous in thread | raw e-mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=235944 --- Comment #17 from Ravi Pokala <rpokala@panasas.com> --- (In reply to Andriy Gapon from comment #15) > also, I think that it is wrong to fail the attach on a failure to access TSOD. > Many DDR3 DIMMs simply do not have it (not sure about DDR4). > I think that all "server" RDIMMS/LRDIMMs and consumer ECC-capable UDIMMS that I've seen had it. But all non-ECC UDIMMs didn't. > So, a missing TSOD should not be a fatal error. Agreed, which is why it isn't: ================================================================ /* The MSBit of the TSOD-presence byte reports whether or not the TSOD * is in fact present. If it is, read manufacturer and device info from * it to confirm that it's a valid TSOD device. It's an error if any of * those bytes are unreadable; it's not an error if the device is simply * not known to us (tsod_match == NULL). * While DDR3 and DDR4 don't explicitly require a TSOD, essentially all * DDR3 and DDR4 DIMMs include one. */ rc = smbus_readb(sc->smbus, sc->spd_addr, tsod_present_offset, &byte); ... if (byte & 0x80) { tsod_present = true; ... } else { tsod_match = NULL; tsod_present = false; } ... /* Create the temperature sysctl IFF the TSOD is present and valid */ if (tsod_present && (tsod_match != NULL)) { ================================================================ -- You are receiving this mail because: You are the assignee for the bug.help
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-235944-227-UBLALeElEU>