Page MenuHomeFreeBSD
Feed Advanced Search

Sep 26 2017

sepherosa_gmail.com added inline comments to D12430: tcp: Don't "negotiate" MSS..
Sep 26 2017, 3:18 AM
sepherosa_gmail.com updated the diff for D12430: tcp: Don't "negotiate" MSS..

tuexen's inputs.

Sep 26 2017, 2:48 AM
sepherosa_gmail.com added inline comments to D12430: tcp: Don't "negotiate" MSS..
Sep 26 2017, 2:26 AM
sepherosa_gmail.com added a comment to D12430: tcp: Don't "negotiate" MSS..

Shouldn't we also apply something like

 	/* Map our computed MSS into the 3-bit index. */
-	mss = min(tcp_mssopt(&sc->sc_inc), max(sc->sc_peer_mss, V_tcp_minmss));
+	mss = sc->sc_peer_mss;
 	for (i = nitems(tcp_sc_msstab) - 1; tcp_sc_msstab[i] > mss && i > 0;
 	     i--)
 		;

since we will use the computed index to restore the sc->sc_peer_mss later.

Sep 26 2017, 2:25 AM

Sep 23 2017

sepherosa_gmail.com added a comment to D12453: mbuf: Remove UDP_IPV4_EX, which is never defined..

This seems generally sensible.
I'm not quite sure about keeping the "Non-standard RSS hash types" comment, as M_HASHTYPE_OPAQUE is basically intended to be a catch-all for any other model, and is not specifically intended to be RSS (hence not having RSS in its name). So I'd be tempted to drop that comment.

Sep 23 2017, 6:25 AM
sepherosa_gmail.com added a comment to D12450: ixl: Fix mbuf hash type settings..

Did microsoft update the referenced link? I don't see the removal notice in it.

Sep 23 2017, 6:24 AM

Sep 22 2017

sepherosa_gmail.com created D12455: rss: Remove never defined UDP_IPV4_EX.
Sep 22 2017, 6:30 AM
sepherosa_gmail.com created D12454: ixgbe: Remove never defined UDP_IPV4_EX.
Sep 22 2017, 5:42 AM
sepherosa_gmail.com created D12453: mbuf: Remove UDP_IPV4_EX, which is never defined..
Sep 22 2017, 5:36 AM
sepherosa_gmail.com created D12450: ixl: Fix mbuf hash type settings..
Sep 22 2017, 3:49 AM

Sep 21 2017

sepherosa_gmail.com created D12430: tcp: Don't "negotiate" MSS..
Sep 21 2017, 8:11 AM
sepherosa_gmail.com created D12429: hyperv/hn: Fix UDP checksum offload issue in Azure.
Sep 21 2017, 7:25 AM

Sep 20 2017

sepherosa_gmail.com added reviewers for D12417: hyperv/hn: Set tcp header offset for CSUM/LSO offloading.: decui_microsoft.com, honzhan_microsoft.com, howard0su_gmail.com, royger.
Sep 20 2017, 5:16 AM
sepherosa_gmail.com created D12417: hyperv/hn: Set tcp header offset for CSUM/LSO offloading..
Sep 20 2017, 5:16 AM

Sep 14 2017

sepherosa_gmail.com created D12365: hyperv/hn: Incease max supported MTU.
Sep 14 2017, 3:58 AM

Sep 13 2017

sepherosa_gmail.com created D12352: hyperv/hn: Fix MTU setting.
Sep 13 2017, 7:25 AM

Sep 12 2017

sepherosa_gmail.com created D12333: hyperv/hn: Apply VF's RSS setting.
Sep 12 2017, 8:26 AM

Sep 7 2017

sepherosa_gmail.com updated the diff for D12176: mlx4_en: Implement SIOCGIFRSS{KEY,HASH}.

Address Hans' comment

Sep 7 2017, 3:05 AM
sepherosa_gmail.com added inline comments to D12176: mlx4_en: Implement SIOCGIFRSS{KEY,HASH}.
Sep 7 2017, 2:29 AM

Sep 5 2017

sepherosa_gmail.com created D12229: mlx4_en: Setup mbuf hash type properly.
Sep 5 2017, 8:40 AM

Sep 1 2017

sepherosa_gmail.com added inline comments to D12176: mlx4_en: Implement SIOCGIFRSS{KEY,HASH}.
Sep 1 2017, 12:54 AM
sepherosa_gmail.com added inline comments to D12175: hyperv/hn: Implement SIOCGIFRSS{KEY,HASH}..
Sep 1 2017, 12:53 AM

Aug 31 2017

sepherosa_gmail.com created D12177: hyperv/hn: Log RSS capabilities mask..
Aug 31 2017, 7:28 AM
sepherosa_gmail.com added inline comments to D12175: hyperv/hn: Implement SIOCGIFRSS{KEY,HASH}..
Aug 31 2017, 5:51 AM
sepherosa_gmail.com updated the summary of D12174: if: Add ioctls to get RSS key and hash type/function..
Aug 31 2017, 5:33 AM
sepherosa_gmail.com added reviewers for D12175: hyperv/hn: Implement SIOCGIFRSS{KEY,HASH}.: adrian, rwatson, glebius, gallatin.
Aug 31 2017, 5:32 AM
sepherosa_gmail.com created D12176: mlx4_en: Implement SIOCGIFRSS{KEY,HASH}.
Aug 31 2017, 5:31 AM
sepherosa_gmail.com created D12175: hyperv/hn: Implement SIOCGIFRSS{KEY,HASH}..
Aug 31 2017, 5:28 AM
sepherosa_gmail.com created D12174: if: Add ioctls to get RSS key and hash type/function..
Aug 31 2017, 5:26 AM

Aug 11 2017

sepherosa_gmail.com created D11982: hyperv: Update copyright for the files changed in 2017.
Aug 11 2017, 8:30 AM
sepherosa_gmail.com created D11981: hyperv/hn: Re-set datapath after synthetic parts reattached..
Aug 11 2017, 7:40 AM
sepherosa_gmail.com created D11979: hyperv/hn: Minor cleanup.
Aug 11 2017, 6:13 AM
sepherosa_gmail.com created D11978: hyperv/hn: Fix/enhance receiving path when VF is activated..
Aug 11 2017, 5:40 AM

Aug 10 2017

sepherosa_gmail.com created D11948: hyperv/hn: Update VF's ibytes properly under transparent VF mode..
Aug 10 2017, 5:39 AM

Aug 1 2017

sepherosa_gmail.com created D11803: hyperv/hn: Implement transparent mode network VF..
Aug 1 2017, 6:31 AM

Jul 25 2017

sepherosa_gmail.com updated the diff for D11635: hyperv: Add VF bringup scripts and devd rules..

Add comment about the network VF and hn(4) work together

Jul 25 2017, 1:59 AM
sepherosa_gmail.com added a comment to D11635: hyperv: Add VF bringup scripts and devd rules..

Sorry for jumping out of the blue, but I don't understand what this is all about. You seem to use lagg together with VFs (I assume PCI Virtual Functions), yet there's no commit message at all explaining what's going on, I think you should add a proper commit message describing how this is supposed to work.

Jul 25 2017, 1:04 AM

Jul 24 2017

sepherosa_gmail.com created D11710: hyperv/hn: Add comment about ether_ifattach event subscription..
Jul 24 2017, 3:49 AM
sepherosa_gmail.com created D11708: hyperv/hn: Renaming and minor cleanup.
Jul 24 2017, 2:27 AM
sepherosa_gmail.com updated the diff for D11635: hyperv: Add VF bringup scripts and devd rules..
Jul 24 2017, 1:49 AM

Jul 21 2017

sepherosa_gmail.com added a comment to D11003: Load balance sockets with new SO_REUSEPORT_LB option.
In D11003#242017, @kevin.bowling_kev009.com wrote:

@sepherosa_gmail.com does this look good to you now?

Jul 21 2017, 5:47 AM

Jul 19 2017

sepherosa_gmail.com created D11651: hyperv/hn: Ignore LINK_SPEED_CHANGE status..
Jul 19 2017, 1:31 AM
sepherosa_gmail.com created D11650: rndis: Add LINK_SPEED_CHANGE status.
Jul 19 2017, 1:30 AM

Jul 18 2017

sepherosa_gmail.com updated the diff for D11618: hyperv/hn: Export VF list and VF-HN mapping.

Consistent temp variable naming

Jul 18 2017, 8:24 AM
sepherosa_gmail.com updated the diff for D11618: hyperv/hn: Export VF list and VF-HN mapping.

Fix eventhandler deregister on detach path

Jul 18 2017, 8:16 AM
sepherosa_gmail.com created D11635: hyperv: Add VF bringup scripts and devd rules..
Jul 18 2017, 7:04 AM
sepherosa_gmail.com accepted D11634: fix the installation failure of FreeBSD on Hyper-V GEN2.

Heh.

Jul 18 2017, 3:05 AM

Jul 17 2017

sepherosa_gmail.com created D11618: hyperv/hn: Export VF list and VF-HN mapping.
Jul 17 2017, 6:46 AM
sepherosa_gmail.com created D11617: ethernet: Add ethernet interface attached event and devctl notification..
Jul 17 2017, 6:40 AM

Jun 30 2017

sepherosa_gmail.com accepted D11415: hv_kbdc.c: remove unused include.
Jun 30 2017, 1:10 AM

Jun 21 2017

sepherosa_gmail.com added a comment to D11003: Load balance sockets with new SO_REUSEPORT_LB option.

And your implementation actually suffers the same issue as Linux: if any of the listen socket is closed, the sockets on the closed listen socket's accept queue are dropped; but well, that's may be acceptable, since I don't think Linux even fixed that.

Would implementing https://github.com/DragonFlyBSD/DragonFlyBSD/commit/02ad2f0b874fb0a45eb69750219f79f5e8982272 solve this?

Jun 21 2017, 7:38 AM

Jun 20 2017

sepherosa_gmail.com accepted D11271: Only dumping the error message under verbose mode.
Jun 20 2017, 9:04 AM
sepherosa_gmail.com added inline comments to D11271: Only dumping the error message under verbose mode.
Jun 20 2017, 3:05 AM
sepherosa_gmail.com accepted D11271: Only dumping the error message under verbose mode.
Jun 20 2017, 1:27 AM

Jun 16 2017

sepherosa_gmail.com added a comment to D11003: Load balance sockets with new SO_REUSEPORT_LB option.

I think you took a pretty old version of the code. I fixed the hash stability issue in a later version; you probably want to check that. And your implementation actually suffers the same issue as Linux: if any of the listen socket is closed, the sockets on the closed listen socket's accept queue are dropped; but well, that's may be acceptable, since I don't think Linux even fixed that.

Jun 16 2017, 10:38 PM

May 3 2017

sepherosa_gmail.com added a comment to D10564: pcicfg: Fix direct calls of pci_cfg{read,write} on system w/o PCI host bridge..
In D10564#219024, @jhb wrote:

I'd kind of prefer we fix the drivers, though I think this patch is fine for now. I have an open item to refactor even the ACPI code to not use pci_cfgreg* directly but instead punt that up to the parent of acpi0 (nexus0) to support arm64 (and to support systems with multiple domains).

May 3 2017, 1:32 AM

May 2 2017

sepherosa_gmail.com updated the diff for D10565: vmbus: Reorganize vmbus device tree.

minor style fixes; remove unnecessary comment.

May 2 2017, 6:32 AM
sepherosa_gmail.com created D10565: vmbus: Reorganize vmbus device tree.
May 2 2017, 6:27 AM
sepherosa_gmail.com created D10564: pcicfg: Fix direct calls of pci_cfg{read,write} on system w/o PCI host bridge..
May 2 2017, 1:26 AM

Apr 20 2017

sepherosa_gmail.com abandoned D8519: hyperv/hn: If the mbuf does not have flowid, fallback to m_ether_hash.

It nullifies Hyper-V hot channel effect, thus actually hurts performance. Drop it.

Apr 20 2017, 6:05 AM

Apr 7 2017

sepherosa_gmail.com added a comment to D10272: Improve handling of syncookies.

Here's an idea from gnn that I agree with:
The fundamental issue here is that the syncookie is good enough to pass validation, but there is sufficient information in the syncache entry to reject it. The patch proposed here is a workaround for deleting the syncache entry.
I think the original problem is that we deleted the syncache entry. Why not set a flag or change the state of the syncache entry to prevent future packets from being accepted for that connection, and let the syncache expiry mechanism deal with deleting the syncache entry eventually.

Apr 7 2017, 1:08 AM

Apr 1 2017

sepherosa_gmail.com created D10230: hyperv/hn: Fixat RNDIS rxfilter after the successful RNDIS initialization.
Apr 1 2017, 3:42 AM
sepherosa_gmail.com accepted D10212: fix the issue of using wrong SRB status.
Apr 1 2017, 1:02 AM

Mar 30 2017

sepherosa_gmail.com accepted D10196: implement initial Hyperv keyboard synthetic driver.

Looks pretty good to me. Some minor code style suggestion.

Mar 30 2017, 7:54 AM

Mar 6 2017

sepherosa_gmail.com accepted D9904: fix UEFI boot on physical machines.
Mar 6 2017, 9:19 AM

Feb 27 2017

sepherosa_gmail.com accepted D9686: fix UEFI VM's bootup on Hyper-V (i.e. Hyper-V Generation-2 VM).

We plan to commit this near the end of this week, if no objection comes. It detects possible errors, and obviously will not make the situation worse than before.

Feb 27 2017, 2:46 AM

Feb 22 2017

sepherosa_gmail.com retitled D9714: hyperv/hn: Misaligned chimney sending buffers should not be used from to hyperv/hn: Misaligned chimney sending buffers should not be used.
Feb 22 2017, 3:08 AM
sepherosa_gmail.com retitled D9713: hyperv/hn: Make sure that RNDIS packet message is at least 4B aligned. from to hyperv/hn: Make sure that RNDIS packet message is at least 4B aligned..
Feb 22 2017, 2:48 AM
sepherosa_gmail.com retitled D9712: hyperv/hn: Simplify RNDIS packet total length calculation. from to hyperv/hn: Simplify RNDIS packet total length calculation..
Feb 22 2017, 2:22 AM

Feb 21 2017

sepherosa_gmail.com retitled D9699: hyperv/hn: Simplify RNDIS packet data offset calculation. from to hyperv/hn: Simplify RNDIS packet data offset calculation..
Feb 21 2017, 8:23 AM

Jan 27 2017

sepherosa_gmail.com accepted D9345: ifnet: move the new ifnet_event EVENTHANDLER_DECLARE to net/if_var.h.
Jan 27 2017, 2:43 AM

Jan 19 2017

sepherosa_gmail.com added inline comments to D9242: lib/libc/x86/sys/__vdso_gettc.c: Conditionalize hyperv support based on MK_HYPERV.
Jan 19 2017, 8:27 AM
sepherosa_gmail.com accepted D9242: lib/libc/x86/sys/__vdso_gettc.c: Conditionalize hyperv support based on MK_HYPERV.

OK, I see. Please fix the commit message when committing then. BTW, in addition to include/dev/hyperv, there are usr.sbin/hyperv and several manpages that you probably want to manage w/ MK_HYPERV.

Jan 19 2017, 8:26 AM
sepherosa_gmail.com added a comment to D9242: lib/libc/x86/sys/__vdso_gettc.c: Conditionalize hyperv support based on MK_HYPERV.

I don't see why it is needed. If hyperv is not in kernel or the ref rdtsc is not the current timecounter, this path will never be hit. Or do you have other concerns?

Jan 19 2017, 8:13 AM

Jan 18 2017

sepherosa_gmail.com added a comment to D8963: ifnet: introduce event handlers for ifup/ifdown events.

If no objection comes, we plan to commit this earlier next week.

Jan 18 2017, 7:02 AM

Jan 9 2017

sepherosa_gmail.com accepted D9102: hyperv/hn: add devctl_notify for VF_UP/DOWN events.
Jan 9 2017, 2:39 PM
sepherosa_gmail.com accepted D9101: hyperv/hn: add a sysctl name for the VF interface.
Jan 9 2017, 2:35 PM
sepherosa_gmail.com accepted D8964: hyperv/hn: add the support for VF drivers.
Jan 9 2017, 9:13 AM

Jan 6 2017

sepherosa_gmail.com retitled D9058: alc: Add Killer E2500 support. from to alc: Add Killer E2500 support..
Jan 6 2017, 6:53 AM
sepherosa_gmail.com abandoned D8904: inet: Fix routing issue by calling if_up().
Jan 6 2017, 6:28 AM
sepherosa_gmail.com retitled D9057: hyperv: Add method to read 64bit Hyper-V specific time value. from to hyperv: Add method to read 64bit Hyper-V specific time value..
Jan 6 2017, 5:35 AM

Jan 5 2017

sepherosa_gmail.com added a comment to D8904: inet: Fix routing issue by calling if_up().

I think the change is a step in the right direction. Certainly, "ifconfig xxN down" followed by an implicit UP should not cause any change to the routing table. Does anyone know why the "down" is removing the route? That seems wrong to me.

Jan 5 2017, 8:16 AM

Jan 4 2017

sepherosa_gmail.com added a comment to D9038: Always null-terminate ccb_pathinq.(sim_vid|hba_vid|dev_name) (part 2).

storvsc parts looks good to me.

Jan 4 2017, 3:09 AM

Dec 30 2016

sepherosa_gmail.com accepted D8964: hyperv/hn: add the support for VF drivers.
Dec 30 2016, 6:28 AM
sepherosa_gmail.com added a comment to D8964: hyperv/hn: add the support for VF drivers.

See the comment, rest looks good.

Dec 30 2016, 5:47 AM
sepherosa_gmail.com added a comment to D8905: if: Defer the if_up until the ifnet.if_ioctl is called..
In D8905#185507, @jhb wrote:

I would suggest rewording the message a bit to something like:

Defer if_up() until after the interface's if_ioctl method is called.
This ensures the interface is initialized by the interface driver before
it can be used by the rest of the system.
Dec 30 2016, 1:09 AM
sepherosa_gmail.com updated D8905: if: Defer the if_up until the ifnet.if_ioctl is called..
Dec 30 2016, 1:09 AM

Dec 29 2016

sepherosa_gmail.com added inline comments to D8964: hyperv/hn: add the support for VF drivers.
Dec 29 2016, 8:50 AM
sepherosa_gmail.com added inline comments to D8964: hyperv/hn: add the support for VF drivers.
Dec 29 2016, 8:50 AM
sepherosa_gmail.com added reviewers for D8963: ifnet: introduce event handlers for ifup/ifdown events: glebius, gnn, rwatson.
Dec 29 2016, 8:17 AM
sepherosa_gmail.com accepted D8962: hyperv/hn: remove the MTU and IFF_DRV_RUNNING checking in hn_rxpkt().

ship it

Dec 29 2016, 7:51 AM

Dec 28 2016

sepherosa_gmail.com retitled D8928: hyperv/hn: Factor out function to set rxfilter. from to hyperv/hn: Factor out function to set rxfilter..
Dec 28 2016, 1:26 AM
sepherosa_gmail.com accepted D8909: hyperv/hn: remember the channel pointer in struct hn_rx_ring.

ship it.

Dec 28 2016, 1:25 AM

Dec 26 2016

sepherosa_gmail.com added inline comments to D8909: hyperv/hn: remember the channel pointer in struct hn_rx_ring.
Dec 26 2016, 1:10 PM
sepherosa_gmail.com retitled D8908: hyperv/hn: Function renaming; no functional changes. from to hyperv/hn: Function renaming; no functional changes..
Dec 26 2016, 9:13 AM
sepherosa_gmail.com retitled D8907: hyperv/hn: Consolidate hn_{suspend,resume} from to hyperv/hn: Consolidate hn_{suspend,resume}.
Dec 26 2016, 9:02 AM
sepherosa_gmail.com retitled D8906: hyperv/vmbus: Nuke unnecessary critical sections. from to hyperv/vmbus: Nuke unnecessary critical sections..
Dec 26 2016, 7:59 AM
sepherosa_gmail.com added a comment to D8904: inet: Fix routing issue by calling if_up().
In D8904#184430, @hrs wrote:

The cause is that the prefix route was removed by in_scrubprefix() in the PRC_IFDOWN handler and never reinstalled upon PRC_IFUP because the reinstallation is done only for ifa passed to SIOCAIFADDR. Just calling if_up(ifp) looks too heavy to me because it causes extra pr_ctlinput() calls of each protocol, not only inetdomain, in order to recover the routes. What do you think about adding reinstalltion of the prefix routes in the protocol specific PRC_IFUP handler instead?

Dec 26 2016, 7:52 AM
sepherosa_gmail.com retitled D8905: if: Defer the if_up until the ifnet.if_ioctl is called. from to if: Defer the if_up until the ifnet.if_ioctl is called..
Dec 26 2016, 5:14 AM
sepherosa_gmail.com retitled D8904: inet: Fix routing issue by calling if_up() from to inet: Fix routing issue by calling if_up().
Dec 26 2016, 3:28 AM