IABSD.fr/src/sys/dev/pv
Branch :
Log
| Author | Commit | Date | CI | Message |
|---|---|---|---|---|
 |
4b1a56af | 2022-01-09 05:42:36 | spelling feedback and ok tb@ jmc@ ok ratchov@ | |
 |
ffaee248 | 2021-11-05 11:38:29 | Constify struct cfattach. | |
 |
04c68813 | 2021-10-24 09:16:53 | pretty & normalize the cfdriver decl | |
 |
69b036cd | 2021-08-31 15:52:10 | When running on Hyper-V, make use of its timecounter as delay func in case we're still using the i8254 for that. On Hyper-V Gen 2 VMs there is no i8254 we can trust, so we need some kind of fallback, especially if there is no TSC either. Discussed with the hackroom ok kettenis@ | |
|
4a58382d | 2021-07-26 11:06:36 | fix an mbuf leak with m_len 0 mbufs from niklas@ via mikeb@ | |
 |
1ab70bb7 | 2021-07-26 06:00:37 | Add mtx_enter/mtx_leave in kvp_pool_keys(). ok mikeb | |
|
e977d71d | 2021-06-11 12:47:15 | Drop received packets unless IFF_RUNNING is set. When hvn(4) attaches it sends commands and waits for replies to come back in. The receive pipe seems to contain both command completions and data packets. When waiting for command completion during hvn(4) attach, it's possible for packets to show up as well. We shouldn't call if_input() if hvn(4) is not set up, so drop them when we're not running. ok mikeb@ | |
|
4123b6a7 | 2021-05-16 15:10:19 | panic does not require a \n at the end. When one is provided, it looks wrong. | |
 |
8611d3cd | 2021-02-23 04:44:30 | timecounting: use C99-style initialization for all timecounter structs The timecounter struct is large and I think it may change in the future. Changing it later will be easier if we use C99-style initialization for all timecounter structs. It also makes reading the code a bit easier. For reasons I cannot explain, switching to C99-style initialization sometimes changes the hash of the resulting object file, even though the resulting struct should be the same. So there is a binary change here, but only sometimes. No behavior should change in either case. I can't compile-test this everywhere but I have been staring at the diff for days now and I'm relatively confident this will not break compilation. Fingers crossed. ok gnezdo@ | |
 |
2f79a14d | 2021-02-11 11:57:32 | Initialize var since it's used in a condition a little bit afterwards. CID 1501713 ok jmatthew@ | |
 |
319c37cd | 2021-01-15 06:14:41 | Report full IP address information to the host. Contrary to what I wrote here 10 years ago, xdr isn't scary at all. ok dlg@ | |
|
73a8717b | 2021-01-13 23:56:48 | Make vm_rpc_send_rpci_tx_buf() use the buffer passed in, not the buffer allocated as part of the softc, though it's always been the same buffer so far. ok dlg@ | |
|
b02e1729 | 2021-01-13 23:50:21 | put M_ZERO in the right place ok dlg@ | |
 |
471f2571 | 2020-12-12 11:48:52 | Rename the macro MCLGETI to MCLGETL and removes the dead parameter ifp. OK dlg@, bluhm@ No Opinion mpi@ Not against it claudio@ | |
|
b6a411b5 | 2020-12-04 18:05:26 | hvn(4): msleep(9) -> msleep_nsec(9) In hvn_alloc_cmd() we may spin waiting for a free rndis_cmd. We check the list once per tick and block with msleep(9) if there aren't any free objects. In practice though we don't need to poll for a free rndis_cmd because our sleep is protected by a mutex, so we can't miss a wakeup(9). That is, it's safe to use msleep_nsec(9) here and not set a timeout (INFSLP). Tested by Andre Stoebe <as@nul.space> (Hyper-V on Windows 10). "LGTM" mikeb@, ok mpi@ | |
|
b910befe | 2020-12-04 03:22:46 | hvn(4), hyperv(4): more tsleep(9) -> tsleep_nsec(9) conversions Replace all tsleep(..., 1) calls in these drivers with tsleep_nsec(9) calls. In every case we can use the nearby delay(9) intervals as the timeout. Tested by Andre Stoebe <as@nul.space> (Hyper-V on Windows 10). "LGTM" mikeb@, ok mpi@ | |
 |
3f3b0cb2 | 2020-10-15 17:47:24 | Ensure resid is correctly set in the INQUIRY case. | |
|
a83ec286 | 2020-10-15 13:22:12 | Replace simple one-line bcopy()/memcpy() of faked data into scsi_xfer with call to scsi_copy_internal_data(). Thus getting xs->resid properly set and adding the usual uio/size sanity checks. | |
|
176be72c | 2020-10-15 11:30:08 | Nuke commented out variable declaration. | |
|
664c6166 | 2020-09-22 19:32:51 | Since dlg@'s 2009 scsi midlayer refactoring the 'struct scsi_generic *cmd' member of 'struct scsi_xfer' has always been pointed at the 'struct scsi_generic cmdstore' member of the same instance. So nuke 'cmdstore' and remove the '*' from cmd. Take the address of cmd as required by the various casts. No intentional functional change. luna88k test by aoyama@, sparc64 test by jmatthew@ Identification of 2009's last *cmd use and ok jmatthew@ | |
|
65e0582e | 2020-09-05 14:21:52 | Rename SID_INQUIRY_HDR (a.k.a. 5) to more explicit SID_SCSI2_HDRLEN and use it to replace various uses of '5' when calculating the amount of data in the INQUIRY response. Matches up more naturally with SID_SCSI2_ALEN. Also use to fix SCSIDEBUG display of INQUIRY responses to show correct count of bytes received/available. | |
|
3ca8dabf | 2020-09-05 13:05:06 | Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY responses. This is what the SCSI specifications say is the correct value and already used in several cases. | |
 |
9e34d16a | 2020-09-04 13:10:16 | Replace TAILQ concatenation loop with TAILQ_CONCAT OK millert@, florian@ | |
|
bb4b71eb | 2020-09-03 12:41:28 | Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'. | |
|
25a9dbe6 | 2020-09-02 22:00:26 | When building emulated INQUIRY results use the SCSI_REV_* #define's to initialize the 'version' field. Not numbers. | |
|
b291b595 | 2020-09-02 21:16:29 | When building emulated INQUIRY results use the SCSI_REV_* #define's to initialize the 'version' field. Not numbers. | |
|
eccd596d | 2020-09-01 12:17:52 | Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to struct scsi_rw_10. ok gnezdo@ jmatthew@ (who also did sparc64 compile test) | |
|
594bc22f | 2020-08-30 10:36:33 | avoid a invalid pointer deref in hvn_stop() intr_barrier() now uses the argument so directly call sched_barrier(NULL) which is what intr_barrier() used to do until recently. problem reported by and fix tested by Andre Stoebe | |
 |
85fcfc35 | 2020-08-26 03:29:05 | Declare hw_{prod,serial,uuid,vendor,ver} in <sys/systm.h>. OK deraadt@, mpi@ | |
|
7380a3a4 | 2020-07-22 13:16:04 | Nuke unused struct scsi_link members of adapter softc's where the driver successfully compiles on one or more of amd64, i386, hppa. | |
|
e5eae15d | 2020-07-20 14:41:12 | Move remaining scsi bus initialization info from "prototype scsi link" fields to struct scsibus_attach_args. Nuke the struct scsi_link * (saa_sc_link) in scaibus_attach_args. Explicitly initialize each field in scsibus_attach_args variables. | |
|
ead808c4 | 2020-07-19 18:57:57 | Move the adapter related items (luns, adapter, adapter_target, adapter_buswidth, adapter_softc) from struct scsi_link to struct scsibus_attach_args. Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k). | |
|
0b29cb40 | 2020-07-16 21:18:29 | Access adapter softc via link->bus->sb_adapter_softc. In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth' via sb_adapter_buswidth. Removes last post-config uses of the copies of bus related information in scsi_link. | |
|
a0d66453 | 2020-07-11 20:15:52 | Garbage collect unused vioscsi_softc member "sc_scsibus". | |
|
e803e20a | 2020-07-11 13:34:05 | Expunge some Captain Obvious comments, tweak whitespace a bit, move a debug statement. All to make upcoming diff(s) smaller and easier to read. | |
|
c0d956de | 2020-07-10 19:43:09 | Shuffle a bit of code to stop using sc->sc_link to hold values for openings outside of the one used at config_found() time. Adapters should take care of their own data. ok dlg@ for a sligthly different vioblk.c version. | |
|
0cae21bd | 2020-07-10 13:26:36 | Change users of IFQ_SET_MAXLEN() and IFQ_IS_EMPTY() to use the "new" API. ok dlg@ tobhe@ | |
 |
d82e6535 | 2020-07-06 13:33:05 | Add support for timeconting in userland. This diff exposes parts of clock_gettime(2) and gettimeofday(2) to userland via libc eliberating processes from the need for a context switch everytime they want to count the passage of time. If a timecounter clock can be exposed to userland than it needs to set its tc_user member to a non-zero value. Tested with one or multiple counters per architecture. The timing data is shared through a pointer found in the new ELF auxiliary vector AUX_openbsd_timekeep containing timehands information that is frequently updated by the kernel. Timing differences between the last kernel update and the current time are adjusted in userland by the tc_get_timecount() function inside the MD usertc.c file. This permits a much more responsive environment, quite visible in browsers, office programs and gaming (apparently one is are able to fly in Minecraft now). Tested by robert@, sthen@, naddy@, kmos@, phessler@, and many others! OK from at least kettenis@, cheloha@, naddy@, sthen@ | |
|
3eec3140 | 2020-06-29 06:50:52 | avoid a NULL deref in xen_intr_barrier() intr_barrier() now uses the argument so directly call sched_barrier(NULL) which is what intr_barrier(NULL) used to do until recently. From Todd Carson on bugs@ ok dlg@ | |
|
382c0517 | 2020-06-27 17:28:58 | Nuke trailing whitespace. | |
|
a7997e14 | 2020-06-27 14:29:44 | No need to bzero()/memset() 'struct scsibus_attach_args' variables immediately before initializing the only field in the struct. | |
|
040fedc0 | 2020-06-27 07:20:57 | Replace TAILQ concatenation loop with TAILQ_CONCAT As a result remove unneeded variable OK kn@, millert@ | |
|
3209772d | 2020-06-24 22:03:40 | kernel: use gettime(9)/getuptime(9) in lieu of time_second(9)/time_uptime(9) time_second(9) and time_uptime(9) are widely used in the kernel to quickly get the system UTC or system uptime as a time_t. However, time_t is 64-bit everywhere, so it is not generally safe to use them on 32-bit platforms: you have a split-read problem if your hardware cannot perform atomic 64-bit reads. This patch replaces time_second(9) with gettime(9), a safer successor interface, throughout the kernel. Similarly, time_uptime(9) is replaced with getuptime(9). There is a performance cost on 32-bit platforms in exchange for eliminating the split-read problem: instead of two register reads you now have a lockless read loop to pull the values from the timehands. This is really not *too* bad in the grand scheme of things, but compared to what we were doing before it is several times slower. There is no performance cost on 64-bit (__LP64__) platforms. With input from visa@, dlg@, and tedu@. Several bugs squashed by visa@. ok kettenis@ | |
|
182e13b7 | 2020-06-24 18:59:30 | Use SDEV_NO_ADAPTER_TARGET instead of the value that is being aassigned to adapter_buswidth to indicate the adapter is not a target on the bus. ok dlg@ as part of a larger diff. | |
|
9667f96c | 2020-06-24 18:47:56 | Use SDEV_NO_ADAPTER_TARGET instead of the value of adapter_buswidth to indicate the adapter is not a target on the bus. ok dlg@ as part of a larger diff. | |
|
3c596f75 | 2020-06-24 01:35:29 | Nuke some pointless *_dev_probe() and *_dev_free() functions and set/trust the scsiconf.c probing limits. Same as was done to vioblk(4). ok dlg@ | |
 |
447a2df6 | 2020-06-22 02:31:32 | use ifiq_input and use it's return value to apply backpressure to rxrs. this is a step toward deprecating softclock based livelock detection. | |
|
16ac8c7f | 2020-06-21 16:46:34 | Nuke pointless vioblk_dev_probe() and vioblk_dev_free() functions and tweak adapter_target and adapter_buswidth values to provide desired semantic of providing only target 0/lun 0 device per vioblk(4) device. Tested by sf@ | |
|
c8f27247 | 2020-05-29 04:42:23 | dev/rndvar.h no longer has statistical interfaces (removed during various conversion steps). it only contains kernel prototypes for 4 interfaces, all of which legitimately belong in sys/systm.h, which are already included by all enqueue_randomness() users. | |
|
421bfb1f | 2020-02-16 03:23:05 | Force a guest info update when we get a Capabilities_Register command. ESXi doesn't carry the guest info over when a guest gets moved to a different host, and this appears to be the only hint the guest gets that this has happened. ok dlg@ | |
|
c8d4e68a | 2020-02-14 15:56:47 | IPS_MAXFER, NAX_XFER, and MAXBSIZE are all defined as (64 * 1024). Reduce confusion by using MAXPHYS (a.k.a. (64 * 1024)) as other drivers do. | |
 |
78588113 | 2020-02-13 15:39:02 | Resolve a few issues with interrupt handling Upon a failed task_add, the adjusted reference counter needs to be decremented. xen_intr_schedule should follow the same logic as xen_intr. Tested by Niklas Hallqvist. Ok mpi@ | |
|
4d7d9900 | 2020-02-12 14:08:56 | Bring a few stragglers into line with the idiom used by the other 40+ scsi drivers. i.e. eliminate the struct scsi_adapter member in the softc and rely on the pointer to a static scsi_adapter in the struct scsi_link member. | |
|
21ceeee0 | 2020-02-05 16:29:29 | Nuke unnecessary abstraction 'scsi_minphys()' which just calls 'minphys()'. Just use & check for NULL instead, since 'minphys()' is always called on the code path ([cd|sd|st]minphys) that calls physio(). | |
|
9f36bc94 | 2020-01-26 00:53:31 | Shuffle some names around to make reading the code less headache inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys' to reflect what it is supposed to do. Use consistent naming convention (<dev>_minphys) for the actual device functions. No functional change. | |
|
854a1b2d | 2020-01-24 05:14:51 | cleanup unused headers generated by config ok tedu@ krw@ deraadt@ | |
|
a867c811 | 2020-01-23 07:52:59 | Use a consistant idiom/format when declaring scsi_adapter structures in drivers. Terse one liners, NULLs instead of 0's, explicitly specify all members, etc. Nuke #ifdef notyet blocks related to the scsi_adapter in aic. No intentional functional change. ok tedu@ | |
|
c19253d5 | 2020-01-22 02:02:31 | xbf(4): tsleep(9) -> tsleep_nsec(9); ok mikeb@ | |
|
11cf39dc | 2020-01-11 21:30:00 | xenstore: *sleep(9) -> *sleep_nsec(9); ok mikeb@ | |
|
03604742 | 2019-12-31 10:05:32 | Convert infinite sleeps to {m,t}sleep_nsec(9). ok kn@ | |
|
44bcf650 | 2019-12-31 01:26:56 | include sys/sensors.h for sensor use | |
 |
aa3e15ec | 2019-12-19 10:15:21 | Add size for free() in vio(4). There is an existing allocsize variable tracking size of allocations, turns out we can pass it to free in the error path. OK florian@, mpi@ | |
 |
274ac25b | 2019-12-13 06:43:46 | pvclock(4): attach even if when PVCLOCK_FLAG_TSC_STABLE is unset Attaches pvclock with lower priority (500) in case of unstable tsc (PVCLOCK_FLAG_TSC_STABLE) instead of not attaching at all. In this state, we do make sure to return a monotonically increasing number. This mostly helps openbsd guests on openbsd vmm(4) where a pvclock with unstable tsc is still better than i8254. ok mlarkin@ | |
|
0296da5a | 2019-12-13 02:16:53 | Always release interrupt source when unmasking the interrupt xen_intr_unmask_release was not decrementing the reference counter on the interrupt source structure when bailing out early which led to the refcnt overflow. From niklas, ok mlarkin | |
 |
cc77704b | 2019-10-25 07:13:54 | Remove some space before tab and other small indentation errors. | |
|
3b758f9b | 2019-10-07 15:36:01 | Attach Hyper-V guest services to VMBus 4.0 Tested by Andre Stoebe and Remi Locherer. ok deraadt, tb | |
|
7b732f28 | 2019-09-01 15:03:32 | Adopt the SCSI versioning #define's from FreeBSD. Eliminate the now unneeded version_to_spc() mapping array, a duplicate #define and a couple of magic numbers. Toss in some comments for future generations of spelunkers. Makes it possible to check for specific SPC versions when new features or eliminated features require such a check. No intentional functional change. | |
|
0b90b277 | 2019-08-30 17:41:18 | Don't use misleadingly named SCSISPC() which returns SCSI version not SPC version. Use raw data in version field to check the SPC level until SCSISPC() is straightened out. Actually a bit clearer to read as the raw data is what is manipulated in adjacent code. | |
 |
35937439 | 2019-08-06 19:24:45 | The vio(4) interface did not recover from mbuf shortage. The logic to start the rx tick did not work since we use if_rxr_get(). An easy fix is to start the rx timeout every second unconditionally. OK dlg@ | |
 |
d8499287 | 2019-05-26 15:22:30 | Support virtio 1.0 for virtio_pci virtio 1.0 for virtio_mmio it not yet implemented, but 0.9 devices continue to work. | |
|
0091658a | 2019-05-26 15:20:04 | Rework virtio_negotiate_features() Add a sc_driver_features field that is automatically used by virtio_negotiate_features() and during reinit. Make virtio_negotiate_features() return an error code. Virtio 1.0 has a special status bit for feature negotiation that means that negotiation can fail. Make virtio_negotiate_features() return an error code instead of the features. Make virtio_reinit_start() automatically call virtio_negotiate_features(). Add a convenience function virtio_has_feature() to make checking bits easier. Add an error check in viomb for virtio_negotiate_features because it has some feature bits that may cause negotiation to fail. More error checking in the child drivers is still missing. ok mlarkin@ | |
|
b5ec98cc | 2019-05-13 15:40:34 | vmm: add host side pvclock Emulate kvm pvclock in vmm(4). Compatible with pvclock(4) in OpenBSD. Linux does not attach to this (yet). Fixes by reyk@ and tested extensively by reyk@, tb@ and phessler@ ok mlarkin@ phessler@ reyk@ | |
|
42004ccb | 2019-04-17 21:44:32 | Remove unused parameter from virtio_init_vq() | |
|
4cb18956 | 2019-04-17 21:43:46 | Initialize virtqueue before passing it to device Reported by Gary Zibrat | |
|
b3dbd5f2 | 2019-03-24 18:22:36 | virtio: Add a few feature bit defines and names ok mlarkin@ | |
|
750a9ae1 | 2019-03-24 18:21:12 | virtio: Prepare for 64 feature bits virtio 1.0 supports an arbitrary number of feature bits. However, so far no more than 64 are used (compared to 32 in virtio 0.9). Adjust data types to support 64 feature bits. Later, we may want to use bitmaps and setbit(), ... to support even more feature bits. ok mlarkin@ | |
|
ee262b2e | 2019-03-24 18:17:24 | virtio: adjust virtio_setup_queue prototype for 1.0 Make it take an address instead of a PFN. Pass the virtqueue pointer. In virtio 1.0, more information has to be configured in the device. Also call virtio_setup_queue() after the information has been filled in. ok mlarkin@ | |
|
203af93c | 2019-01-19 16:23:46 | virtio: Introduce defines for config(8) flags | |
|
090e143c | 2019-01-19 16:21:00 | viomb: tweak feature negotiation VIRTIO_F_RING_INDIRECT_DESC is always negotiated by the transport driver, no need to specify it explicitly. VIRTIO_BALLOON_F_MUST_TELL_HOST is not offered but is handled in the code. Offer it during negotiation, too. | |
|
5acd3de4 | 2019-01-10 18:22:17 | Fix previous It accidentally contained a part of a different diff. | |
|
fbc65ea9 | 2019-01-10 18:11:42 | Move some PCI-specific defines to a new virtio_pcireg.h file Also add some virtio 1.0 status and feature bits | |
|
e578ff6a | 2019-01-10 18:06:56 | Remove some more code if VIRTIO_DEBUG is 0 | |
|
1dd6effc | 2019-01-10 18:05:43 | Move some common defines to virtiovar.h And fix some comments | |
|
6164d80c | 2019-01-08 16:24:09 | viocon: Remove obsolete handling of sc_intrhand | |
|
c30b7943 | 2019-01-08 16:23:01 | virtio: Don't include feature strings without VIRTIO_DEBUG They are only used if VIRTIO_DEBUG is enabled. This should make the kernel slightly smaller. | |
 |
c14cae50 | 2018-12-05 18:02:51 | Correctly disable pvclock(4) on old hardware that lack a stable clock I falsely assumed that the KVM_FEATURE_CLOCKSOURCE_STABLE_BIT indicates that the actual clock values are stable, but it turned out that this isn't always the case. To detect if the clock value is stable, we now read it once in pvclock_attach() and check for the PVCLOCK_FLAG_TSC_STABLE flag. This needs further investigation. Reported and fix tested by johnw.mail at gmail.com OK chris@ phessler@ | |
 |
c9ec5b84 | 2018-11-24 13:12:29 | only attach pvclock(4) inside a KVM guest tested by brynet@ OK reyk@ | |
|
22d921f9 | 2018-11-23 12:37:40 | Add the pvclock(4) guest driver for paravirtual clocks This improves timekeeping on KVM guests as it runs much better than the virtualized acpihpet or acpitimer timecounters and the invtsc is not always available. Many thanks to Janne Johansson, landry@, and benno@ for testing amd64/i386. OK mlarkin@ phessler@ | |
|
126b881f | 2018-09-25 13:46:44 | Insert a workaround for per-ifp ioctl being called w/o NET_LOCK(). An example of such code path is vlan_destroy() where we don't want to grab the lock because this driver is already NET_LOCK()-free. The real solution to this problem is to stop holding the NET_LOCK() around driver *_ioctl() routines. Problem reported & fix tested by Élie Bouttier, ok visa@, bluhm@ | |
|
5f7616cc | 2018-08-24 16:07:01 | Add common Ethernet interface attributes to vio(4)'s build config. OK deraadt@ | |
|
dbe7e165 | 2018-05-17 12:32:33 | Let hvn_iff handle promisc mode activation Tested by Daniel Wade, <Daniel2 ! Wade at ge ! com>, thanks! | |
 |
9e9abf5b | 2018-04-28 15:44:59 | replace add_*_randomness with enqueue_randomness() this gets rid of the source annotation which doesn't really add anything other than adding complexitiy. randomess is generally good enough that the few extra bits that the source type would add are not worth it. ok mikeb@ deraadt@ | |
|
769cb144 | 2018-03-11 14:31:34 | Workaround an Azure UDP transmit checksum offloading bug According to https://reviews.freebsd.org/D12429, UDP transmit checksum offloading doesn't work in Azure when sizeof(IP hdr + UDP hdr + payload) is greater than 1420 and don't fragment bit isn't set in the IP header. Instead of peeking into the packet and attempting to fix the fallout, we restrict UDP checksum offloading to NDIS versions newer than the one used in Azure (6.30) in an attempt to keep the feature enabled. Unfortunately, there's no way to tell Azure apart from a standalone Hyper-V to make this check more precise. | |
|
c72a87aa | 2018-03-08 05:33:56 | Implement the VMWare vmbackup protocol using vfs_stall(). This lets you clone running guests and take disk-only snapshots, most of the time ending up with a clean filesystem. ok dlg@ deraadt@ | |
|
f0b002d0 | 2018-02-27 08:44:58 | Release the netlock when sleeping for control messages in in vioioctl(). Prevents a deadlock reported by aalm@ | |
|
3e676399 | 2018-02-19 08:59:52 | Remove almost unused `flags' argument of suser(). The account flag `ASU' will no longer be set but that makes suser() mpsafe since it no longer mess with a per-process field. No objection from millert@, ok tedu@, bluhm@ | |
|
220dd317 | 2018-01-21 18:54:46 | Release interrupt source if the interrupt task is already scheduled Tested by Imre Oolberg, thanks! | |
|
b5e9546a | 2018-01-21 18:51:19 | Improve panic message | |
|
b47fcd48 | 2018-01-20 20:03:45 | Disable TCP and UDP transmit checksum offloading In certain configurations, transmit checksum offloading doesn't appear to work correctly, preventing correct TCP and UDP operation. Some of these issues involving VLAN tagging are documented by the Xen project. Problem reported and fix tested by Imre Oolberg, thanks! | |
|
15c2e150 | 2018-01-18 11:43:20 | Fix an off-by-one uncovered by the recent free(9) change From canacar@, thanks! |