[01:20:23] Jianfeng: Page alignment makes sense when an IO-MMU is used. The device should only have access to the buffer. If the device has access to other data then that can affect reliability and security. [01:22:32] Jianfeng: O_DIRECT memory buffers only needs logical block size alignment though, not page alignment. [02:15:32] fun fact: PS4 firmware developers didn't know that vfio operates on entire pages. They were mapping a very small byte buffer allocated on the stack [02:15:44] that's how PS4 got hacked and people installed linux on it [02:19:37] providing page aligned buffers might be good from performance perspective. eg in case of nvme request might be split by driver if subsequent addresses are miss aligned. [02:36:11] jimharris: i pushed updated nvme memzone patches. I'd vote for merging this series as is, because it's solving real issues [02:36:55] we could worry about introducing a nicer contig mem allocation API later [02:38:06] with I/OAT out, there are just two places where physically contig memory is required after all - NVMe and virtio-pci [03:04:17] *** Joins: y4m4 (490fb8f3@gateway/web/cgi-irc/kiwiirc.com/ip.73.15.184.243) [03:04:33] Quick question here.. [03:05:05] I have been writing a multiple based AIO block device support, I am having a strange problem and not able to decipher what might be causing this problem [03:05:18] this issue is most probably related to my changes. [03:05:42] What i see is a segfault [03:05:43] ``` [03:05:52] ``` [03:06:11] ```vbdev_gpt.c: 297:spdk_gpt_bdev_complete: *ERROR*: Gpt: bdev=aio0 io error status=0Program received signal SIGSEGV, Segmentation fault.spdk_put_io_channel (ch=0xa0) at io_channel.c:505505 if (ch->ref == 0) {Missing separate debuginfos, use: debuginfo-install libibverbs-41mlnx1-OFED.4.3.0.1.8.43101.x86_64 librdmacm-41mlnx1-OFED.4.2.0.1 [03:06:12] .3.43101.x86_64(gdb) bt#0 spdk_put_io_channel (ch=0xa0) at io_channel.c:505#1 0x000000000042783a in spdk_gpt_bdev_complete (bdev_io=, status=, arg=0xfc5540) at vbdev_gpt.c:292#2 0x0000000000428ab5 in bdev_aio_poll (arg=0xfc5700) at bdev_aio.c:348#3 0x0000000000440c8a in _spdk_reactor_run (arg=0x6f5400) at reactor.c [03:06:12] :455#4 0x000000000044112d in spdk_reactors_start () at reactor.c:624#5 0x000000000044021d in spdk_app_start (opts=opts@entry=0x7fffffffdf90, start_fn=start_fn@entry=0x406490 , arg1=arg1@entry=0x0, arg2=arg2@entry=0x0) at app.c:485#6 0x000000000040557c in main (argc=1, argv=0x7fffffffe0e8) at nvmf_main.c:79(gdb) ``` [03:06:48] vbdev_gpt.c: 297:spdk_gpt_bdev_complete: *ERROR*: Gpt: bdev=aio0 io error status=0Program received signal SIGSEGV, Segmentation fault.spdk_put_io_channel (ch=0xa0) at io_channel.c:505505 if (ch->ref == 0) {Missing separate debuginfos, use: debuginfo-install libibverbs-41mlnx1-OFED.4.3.0.1.8.43101.x86_64 librdmacm-41mlnx1-OFED.4.2.0.1.3. [03:06:48] 43101.x86_64(gdb) bt#0 spdk_put_io_channel (ch=0xa0) at io_channel.c:505#1 0x000000000042783a in spdk_gpt_bdev_complete (bdev_io=, status=, arg=0xfc5540) at vbdev_gpt.c:292#2 0x0000000000428ab5 in bdev_aio_poll (arg=0xfc5700) at bdev_aio.c:348#3 0x0000000000440c8a in _spdk_reactor_run (arg=0x6f5400) at reactor.c:45 [03:06:49] 5#4 0x000000000044112d in spdk_reactors_start () at reactor.c:624#5 0x000000000044021d in spdk_app_start (opts=opts@entry=0x7fffffffdf90, start_fn=start_fn@entry=0x406490 , arg1=arg1@entry=0x0, arg2=arg2@entry=0x0) at app.c:485#6 0x000000000040557c in main (argc=1, argv=0x7fffffffe0e8) at nvmf_main.c:79(gdb) ``` [03:07:03] Sorry for double posting, my network seem to be acting up [03:07:13] Thanks for all the help [03:37:15] *** Quits: dlw (~Thunderbi@114.255.44.143) (Ping timeout: 256 seconds) [03:45:17] y4m4: do you unregister any bdev before that happens? [04:42:59] *** Joins: tzawadzki (tomzawadzk@nat/intel/x-uzlwprpzymlmsbjs) [04:43:00] *** Quits: tomzawadzki (tomzawadzk@nat/intel/x-sftfderxohkyfljz) (Remote host closed the connection) [05:44:03] *** Joins: lyan (~lyan@2605:a000:160e:2124:4a4d:7eff:fef2:eea3) [05:44:28] *** lyan is now known as Guest50666 [06:55:35] *** Quits: tzawadzki (tomzawadzk@nat/intel/x-uzlwprpzymlmsbjs) (Ping timeout: 240 seconds) [07:09:35] *** Joins: philipp-sk (~Philipp@ktnron0916w-lp130-02-76-66-162-159.dsl.bell.ca) [07:49:24] hmm, I haven't run UT w/coverage since we moved dirs around a little. Can't get coverage info off master for some reason, tail end of my output is here - anyone else having issues? https://gist.github.com/peluse/98feaf68de1284fefa6a539063f0cba1 [07:50:04] ut_coverage/ut_cov_unit.info exists but is 0 len... [08:07:43] *** Joins: tzawadzki (~tomzawadz@134.134.139.74) [08:22:34] *** Quits: tzawadzki (~tomzawadz@134.134.139.74) (Ping timeout: 264 seconds) [08:29:13] also, if I comment out the '$LCOV -q -r' lines in the unittest.sh that remove coverage we don't care about I do get html files generated, somehow I'm not picking up coverage for /lib [08:30:25] darsto: can we get this patch in first? https://review.gerrithub.io/#/c/spdk/spdk/+/416840/ it's not controversial at all [08:30:42] ok [08:30:47] i was just looking at it [08:31:25] i just posted a couple of nits on the patch before it - functionally it looks fine but want bwalker to weigh in before we submit it [08:32:14] this ioat one has a compilation failure https://review.gerrithub.io/#/c/spdk/spdk/+/416994/ [08:35:23] i know, there was a little bug in the original series and I just don't want to resubmit all patches at once [08:35:35] ok [08:35:40] otherwise klateck will bite [08:35:54] bite what? [08:35:55] :) [08:36:00] bite me [08:36:04] ouch [08:38:55] darsto: i was thinking more about the rte_vhost2 stuff last night after our meeting [08:40:09] when there's a new SET_MEM_TABLE, if the mem table entries match, does rte_vhost keep the old descriptors? or close the old ones and use the new ones? [08:40:52] i don't think our copy does that [08:41:00] but upstream rte_vhost does [08:41:16] upstream keeps the old descriptors? [08:41:41] sorry, upstream rte_vhost checks if the mappings changed - if they didn't, it just closes the new descriptors [08:42:02] it was introduced in rte_vhost after we forked it [08:42:11] ok [08:42:43] so upstream then works better now with zero-copy in that regard [08:44:22] i think it would be helpful if we had a specific list of each vhost message, and how the current upstream implementation is broken for zero-copy or storage use cases - with specifics [08:44:41] actually, upstream is wrong [08:44:58] file descriptors could technically change [08:45:07] so what rte_vhost does is a hack [08:45:21] could technically change how? [08:46:47] figured out the unittest issue... before unittest.sh was in the repo root. It has to be called from there even now after its been moved to /test/unit. So, to check coverage locally run it with "./test/unit/unittest.sh" and then your report will be off the root in "ut_coverage" [08:47:33] even if the memory has the same guest_user_addr and guest_phys_addr as before, it could be completely remapped now, right? [08:48:05] i wouldn't think so - on the qemu side it's still using the same file descriptor [08:48:58] sure, nobody would change mappings like that [08:49:05] just thinking loud [08:49:24] sure - keep doing that :) [08:50:24] but that patch you forwarded last night talks about the other vhost messages: https://patches.dpdk.org/patch/35663/ [08:51:33] why does virtio/vhost-net not see these problems with SET_VRING_CALL/KICK? [08:53:46] i wonder if the file descriptors for the eventfds change on the qemu side? [08:55:03] eventfds do change [08:55:09] i'm not sure why actually [08:55:28] change on our vhost side? or qemu side? [08:55:37] QEMU sends us SET_VRING_CALL at very beginning of initialization [08:56:09] right - and then again in virtqueue_start [08:56:13] as if it was setting up irqfds during vm start [08:56:14] peluse: this should make unittest.sh work from any directory: https://review.gerrithub.io/#/c/spdk/spdk/+/418217 [08:56:48] probably an issue on their side that's being worked-around [08:56:54] yeah it was for sure [08:56:57] drv tahnks [08:57:12] upstream rte_vhost has that per-vq read-write lock and Maxime offered to add some public APIs for us to use it [08:58:12] the discussion in this patch focused on SET_MEM_TABLE, but that's not the only issue with a public per-vq lock [08:58:32] we don't use internal DPDK data structures - so we don't care about their lock [08:58:42] per-vq locks is terrible [08:58:58] we need a notification that vq changed [09:01:37] drv, FYI I tested that too before I +1'd it :) [09:01:52] once its merged I'll send a reminder out on the dist list about local coverage and how to do it [09:02:05] cool, definitely want to make it as easy as possbile to generate coverage [09:05:31] now, what was I doing? :) [09:13:05] *** Quits: y4m4 (490fb8f3@gateway/web/cgi-irc/kiwiirc.com/ip.73.15.184.243) (Ping timeout: 240 seconds) [09:38:56] jimharris: Can you revisit the -1 on this one? [09:38:57] https://review.gerrithub.io/#/c/spdk/spdk/+/386166/ [09:39:06] we've gone round and round and I think we may be back to where we started [09:51:06] *** Joins: jkkariu (jkkariu@nat/intel/x-bspjqcsivomkhmzx) [09:56:27] jimharris: When creating the bdev layer's buffer pools, we still tell DPDK to create per-cpu caches [09:56:53] actually nevermind - I was thinking we cached buffers in a TAILQ, but we only do that for bdev_io [10:01:06] *** Joins: y4m4 (49def5c3@gateway/web/cgi-irc/kiwiirc.com/ip.73.222.245.195) [10:23:40] Thanks for the pointers I was able to get it resolved darsto [10:27:38] y4m4: you're welcome. was it a problem with your module, or with the existing vbdev_gpt? [10:27:58] from the backtrace you posted it seemed like we have a bug in vbdev_gpt [10:28:06] It was a problem in my extent_size [10:28:14] v/s some offset calculation [10:28:34] Now i am seeing hangs and replies with this error [10:28:38] rdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to s [10:28:39] end response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: [10:28:39] *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:re [10:28:40] quest_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response c [10:28:40] apsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsulerdma.c: 556:request_transfer_out: *ERROR*: Unable to send response capsule [10:28:44] Will have to dig more [10:36:07] *** Joins: nik_b (~niklasbeg@25.165.6.51.dyn.plus.net) [11:05:09] hi bwalker, all the comments in #416878 and #416879 have been addressed, would appreciate the ball kicked from your side ;) [11:07:33] I'll try to get through it today - thanks! [11:32:52] *** jimharris sets mode: +o bwalker [11:36:15] bwalker: https://review.gerrithub.io/#/c/spdk/spdk/+/418176/ needs another +2 then we can submit a set of 4 iSCSI patches from Shuhei [11:37:34] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [11:37:35] (spdk/master) bdev: submit queued IO after disabling QoS (Jim Harris) [11:37:35] Diff URL: https://github.com/spdk/spdk/compare/436c0c189bca...140eaaa08441 [11:37:35] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [11:38:09] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [11:38:10] (spdk/master) test: remove spdk_cunit library (Jim Harris) [11:38:10] Diff URL: https://github.com/spdk/spdk/compare/140eaaa08441...ca1de9eb7593 [11:38:10] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [11:40:32] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [11:40:33] (spdk/master) nvme/pcie: do not assume tracker array memory is physically contiguous (Dariusz Stojaczyk) [11:40:33] Diff URL: https://github.com/spdk/spdk/compare/ca1de9eb7593...d36ce206ec77 [11:40:33] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [11:45:42] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [11:45:43] (spdk/master) bdev: pack internal fields of spdk_bdev_io. (Seth Howell) [11:45:44] Diff URL: https://github.com/spdk/spdk/compare/d36ce206ec77...025997a0b23b [11:45:44] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [11:46:19] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [11:46:20] (spdk/master) test: make unittest.sh work from any directory (Daniel Verkamp) [11:46:20] Diff URL: https://github.com/spdk/spdk/compare/025997a0b23b...f1acfee36090 [11:46:20] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [12:06:49] *** Joins: johnmeneghini (~johnmeneg@pool-100-0-53-181.bstnma.fios.verizon.net) [12:11:28] I'm looking for some help with my vagrant and vm_setup changes. [12:12:10] I'm currently stuck because I can't get qemu/capstone to compile [12:12:30] capstone git [12:12:30] GEN x86_64-softmmu/config-devices.mak.tmp [12:12:30] GEN config-host.h [12:12:30] make[1]: *** No rule to make target '/home/vagrant/qemu/spdk-2.12-pre/capstone/libcapstone.a'. Stop. [12:12:30] make: *** [Makefile:430: subdir-capstone] Error 2 [12:12:30] make: *** Waiting for unfinished jobs.... [12:12:31] GEN x86_64-softmmu/config-devices.mak [12:12:46] Has anyone seen this problem before? [12:20:00] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [12:20:01] (spdk/master) virtio: move queue allocation to backend code (Dariusz Stojaczyk) [12:20:01] Diff URL: https://github.com/spdk/spdk/compare/f1acfee36090...3a15b49b05b3 [12:20:01] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [12:27:37] johnmemeghini, yes I have. It happened to me when I had tsocks installed, but had not written out my tsocks configuration file. [12:45:24] it looks like the qemu submodules are missing possibly? [12:45:40] so if you are behind a restrictive proxy, it is probably tsocks-related, as sethhowe mentioned [12:46:31] No, there's no proxy needed. I think Seth found the problem in that I have no tsocks configuration setup. [12:47:32] hmm, I see; if you have tsocks installed but not configured, maybe that causes problems [12:51:42] Seth and I have a plan to fix this. I'll update my patch shortly. Karol also sent me some email and asked about a QEMU problem. I'll look at that problem too. [13:08:34] drv: are you ok with https://review.gerrithub.io/#/c/spdk/spdk/+/417237/? i saw you commented but no vote [13:09:01] I just made a couple more comments and added my +2 to it [13:09:22] we should also probably wrap the long lines, but it's OK for now [13:12:04] *** Quits: nik_b (~niklasbeg@25.165.6.51.dyn.plus.net) (Quit: nik_b) [13:15:30] looking at the destruct_lvol_bdev patch https://review.gerrithub.io/#/c/spdk/spdk/+/407402/37/doc/lvol.md [13:16:09] the first three files are basically explaining the differences between delete_bdev and destruct_lvol_bdev - is there any reason we can't just get rid of delete_bdev entirely? [13:16:24] i.e. once all of the bdev modules have their own equivalent [13:16:51] yeah, if anything, it should be renamed to make it clear that it's for hot remove testing only [13:17:05] although I'm not sure we even need it for that [13:17:07] do we need it for hot remove testing? [13:17:20] probably not [13:17:42] and I agree, best to just not mention delete_bdev at all in that doc section [13:19:52] *** Joins: tomzawadzki (~tomzawadz@192.55.54.40) [13:37:13] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [13:37:14] (spdk/master) doc/jsonrpc: Add RPCs for logical volume block devices (Shuhei Matsumoto) [13:37:14] Diff URL: https://github.com/spdk/spdk/compare/1196deb504b4...d1aa8f622d57 [13:37:14] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [13:37:57] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [13:37:58] (spdk/master) test: fix scan-build errors in bdev unittests. (Seth Howell) [13:37:58] Diff URL: https://github.com/spdk/spdk/compare/3a15b49b05b3...1196deb504b4 [13:37:58] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [13:39:31] would love to get this one in: https://review.gerrithub.io/#/c/spdk/spdk/+/413913/ [13:48:19] wait, can examine_config be asynchronous? [13:48:50] lgtm, but probably needs bwalker's input [13:52:41] *** Joins: nik_b (~niklasbeg@25.165.6.51.dyn.plus.net) [13:52:43] examine_config currently requires the user to synchronously call spdk_bdev_module_examine_done or it throws an error [13:52:59] but if it must synchronously call that, why call that at all? returning should be plenty [13:54:01] i debated that - i'm ok with changing it, i just thought it might be weird for a bdev module to have to call spdk_bdev_module_examine_done in some cases but not others [13:54:36] I'd be ok if either examine_config was actually asynchronous and required the user to call it, or it was synchronous and just returned [13:54:59] it's weird to have to call a function within the examine callback to indicate that you're done when it's enforced to be synchronous [13:55:07] if examine_disk is implemented synchronously, it still has to call it [13:55:15] sure, but it can be asynchronous [13:55:24] if examine_config could be asynchronous, I wouldn't have an issue with it either [13:55:37] but he has code in spdk_bdev_start that's forcing it to be synchronous [13:56:30] I'm even ok if we're doing this with the thought that some time in the future we'd allow examine_config to become asynchronous [13:56:31] if we allow it to be asynchronous, it makes this patch way more complicated [13:56:32] to avoid an API change [13:57:24] the observation that examine_config is always synchronous (at least currently) makes this patch way simpler than I originally thought it woudl be [13:58:22] I guess I'm ok with this patch because it does catch mistakes, but it's a little unclear what to do in reponse to those callbacks [13:58:43] the documentation never actually mentions spdk_bdev_module_examine_done, but it didn't before either [13:59:09] the bdev_module.md docs need to be updated anyway - they are out of date before this patch [13:59:37] *** Quits: tomzawadzki (~tomzawadz@192.55.54.40) (Ping timeout: 268 seconds) [14:15:28] *** Joins: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) [14:15:29] (spdk/master) bdev: split examine into two parts (Piotr Pelplinski) [14:15:29] Diff URL: https://github.com/spdk/spdk/compare/d1aa8f622d57...9d258c75afa1 [14:15:29] *** Parts: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) () [14:20:15] *** Quits: philipp-sk (~Philipp@ktnron0916w-lp130-02-76-66-162-159.dsl.bell.ca) (Ping timeout: 245 seconds) [14:21:43] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [14:21:44] (spdk/master) iscsi: Remove resdiual count variables about unsupported bidirectional command (Shuhei Matsumoto) [14:21:44] Diff URL: https://github.com/spdk/spdk/compare/9d258c75afa1...0cd2f86977f5 [14:21:44] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [14:26:20] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [14:26:21] (spdk/master) scsi: only set SPDK_SCSI_STATUS_GOOD on bdev callback (Tomasz Zawadzki) [14:26:21] Diff URL: https://github.com/spdk/spdk/compare/0cd2f86977f5...bcfd6d0fb47b [14:26:21] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [14:37:16] bwalker: can you look at sethhowe's nvmf patch set? https://review.gerrithub.io/#/c/spdk/spdk/+/417345/ [14:43:59] bwalker: also please take a look at the mem map fixes if you have a chance: https://review.gerrithub.io/#/c/spdk/spdk/+/418105/ [14:45:26] ok, done on some, comments on others [14:46:11] *** Quits: Guest50666 (~lyan@2605:a000:160e:2124:4a4d:7eff:fef2:eea3) (Quit: Leaving) [15:00:54] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [15:00:55] (spdk/master) virtio/user: do not set vring physical address (Dariusz Stojaczyk) [15:00:55] Diff URL: https://github.com/spdk/spdk/compare/bcfd6d0fb47b...095f355bf3de [15:00:55] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [15:09:43] *** Quits: nik_b (~niklasbeg@25.165.6.51.dyn.plus.net) (Quit: nik_b) [15:18:18] anyone else had a problem with gerrit/gerrithub, where you're typing a comment and the browser randomly jumps to another part of the file? [15:18:37] *** Joins: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) [15:18:38] (spdk/master) test/unit: fix several scan-build errors. (Seth Howell) [15:18:38] Diff URL: https://github.com/spdk/spdk/compare/095f355bf3de...88e26908fa31 [15:18:38] *** Parts: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) () [15:18:41] it seems to especially happen with large patches to a single file - i.e. the pvol patch [15:23:40] yep [15:23:43] if the patch is big it does that [15:24:19] I think if you have something selected it does that [15:39:14] *** Joins: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) [15:39:15] (spdk/master) env_dpdk: fix mem_map translation for sizes >2MB (Daniel Verkamp) [15:39:15] Diff URL: https://github.com/spdk/spdk/compare/88e26908fa31...7d24e2a4cd23 [15:39:15] *** Parts: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) () [15:50:26] 30 patches merged today [15:50:40] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [15:50:41] (spdk/master) doc/jsonrpc: fix response of context_switch_monitor (Daniel Verkamp) [15:50:41] Diff URL: https://github.com/spdk/spdk/compare/ad51b333e887...b2d3a95b0e40 [15:50:41] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [15:52:08] *** Joins: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) [15:52:09] (spdk/master) test: remove leftover spdk_cunit declaration (Daniel Verkamp) [15:52:09] Diff URL: https://github.com/spdk/spdk/compare/b2d3a95b0e40...49e24df51c87 [15:52:09] *** Parts: travis-ci (~travis-ci@ec2-54-227-1-238.compute-1.amazonaws.com) () [15:52:25] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [15:52:26] (spdk/master) bdev: reorder items in spdk_bdev. (Seth Howell) [15:52:26] Diff URL: https://github.com/spdk/spdk/compare/49e24df51c87...274980704e8b [15:52:26] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [15:53:27] *** Joins: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) [15:53:28] (spdk/master) test/json: Add rbd bdev test (Pawel Kaminski) [15:53:29] Diff URL: https://github.com/spdk/spdk/compare/274980704e8b...57bce7953f58 [15:53:29] *** Parts: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) () [15:54:32] *** Joins: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) [15:54:33] (spdk/master) nvmf: FC-NVMe spec. header file (John Barnard) [15:54:33] Diff URL: https://github.com/spdk/spdk/compare/57bce7953f58...82c64e2a4167 [15:54:33] *** Parts: travis-ci (~travis-ci@ec2-54-81-190-44.compute-1.amazonaws.com) () [15:54:57] *** Joins: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) [15:54:58] (spdk/master) nvmf: give qpair_disconnect an asynchronous api. (Seth Howell) [15:54:59] Diff URL: https://github.com/spdk/spdk/compare/82c64e2a4167...09e3f4e3db2c [15:54:59] *** Parts: travis-ci (~travis-ci@ec2-54-163-84-24.compute-1.amazonaws.com) () [17:48:14] Jim, Seth and Paul, please take a look at : https://review.gerrithub.io/#/c/spdk/spdk/+/418236/ [17:49:03] I'd like to get these patches merged. This baby is overdue. [18:59:46] *** Quits: y4m4 (49def5c3@gateway/web/cgi-irc/kiwiirc.com/ip.73.222.245.195) (Ping timeout: 264 seconds) [22:44:42] *** Quits: johnmeneghini (~johnmeneg@pool-100-0-53-181.bstnma.fios.verizon.net) (Quit: Leaving.) [23:53:29] *** Joins: dlw (~Thunderbi@222.131.154.61)