Date: Tue, 07 Jan 2020 12:50:48 +0000 From: bugzilla-noreply@freebsd.org To: net@FreeBSD.org Subject: [Bug 243126] Assertion fl->ifl_cidx == cidx failed at /usr/src/sys/net/iflib.c:2531 Message-ID: <bug-243126-7501-xVoIS4fTuq@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-243126-7501@https.bugs.freebsd.org/bugzilla/> References: <bug-243126-7501@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=243126 Andriy Gapon <avg@FreeBSD.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|New |Open --- Comment #3 from Andriy Gapon <avg@FreeBSD.org> --- I wonder if irf_idx can "jump ahead" of ifl_cidx because of a skipped zero-length packet? Hmm, it seems like it could be the case. Given the current value of iri_cidx/ifr_cq_cidx, I examined three descriptors before that index: (kgdb) p $19.vxcr_u.rxcd[476] $21 = {rxd_idx = 142, pad1 = 0, eop = 1, sop = 1, qid = 0, rss_type = 2, no_csum = 0, pad2 = 0, rss_hash = 600473664, len = 1514, error = 0, vlan = 0, vtag = 0, csum = 0, csum_ok = 1, udp = 0, tcp = 1, ipcsum_ok = 1, ipv6 = 0, ipv4 = 1, fragment = 0, fcs = 0, type = 3, gen = 1} (kgdb) p $19.vxcr_u.rxcd[475] $22 = {rxd_idx = 141, pad1 = 0, eop = 1, sop = 1, qid = 0, rss_type = 0, no_csum = 0, pad2 = 0, rss_hash = 0, len = 0, error = 0, vlan = 0, vtag = 0, csum = 0, csum_ok = 0, udp = 0, tcp = 0, ipcsum_ok = 0, ipv6 = 0, ipv4 = 0, fragment = 0, fcs = 0, type = 3, gen = 1} (kgdb) p $19.vxcr_u.rxcd[474] $23 = {rxd_idx = 140, pad1 = 0, eop = 1, sop = 1, qid = 0, rss_type = 2, no_csum = 0, pad2 = 0, rss_hash = 600473664, len = 66, error = 0, vlan = 0, vtag = 0, csum = 0, csum_ok = 1, udp = 0, tcp = 1, ipcsum_ok = 1, ipv6 = 0, ipv4 = 1, fragment = 0, fcs = 0, type = 3, gen = 1} The descriptor at 476 with rxd_idx = 142 seems like the current packet. And the previous descriptor at 475 with rxd_idx = 141 is a zero-length packet: eop = 1, sop = 1, len = 0. The packet before it is a normal packet again: eop = 1, sop = 1, len = 66. Now, how to fix the problem? I see two ways: - a driver can notify iflib of a zero length packet (via a new callback), so that iflib can skip the corresponding entry in the appropriate free list - rxd_frag_to_sd() can skip through fl entries until ifl_cidx becomes equal to irf_idx -- You are receiving this mail because: You are the assignee for the bug.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-243126-7501-xVoIS4fTuq>