aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* wintun: destroy devinfolist after usageJason A. Donenfeld2019-05-101-0/+1
|
* wintun: registry: replace REG_NOTIFY with NOTIFYSimon Rozman2019-05-103-13/+9
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* wintun: IpConfig is a MULTI_SZ, and fix errorsJason A. Donenfeld2019-05-102-49/+72
|
* wintun: poll for device keyJason A. Donenfeld2019-05-101-9/+10
| | | | It's actually pretty hard to guess where it is.
* wintun: fix scope of error objectJason A. Donenfeld2019-05-101-5/+6
|
* wintun: wait for interface registry key on device creationSimon Rozman2019-05-107-133/+534
| | | | | | By using RegNotifyChangeKeyValue(). Also disable dead gateway detection. Signed-off-by: Simon Rozman <simon@rozman.si>
* conn: remove scope when sanity checking IP address formatJason A. Donenfeld2019-05-091-1/+7
|
* wintun: fix GUID leading zero paddingSimon Rozman2019-05-091-1/+1
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* mod: update depsJason A. Donenfeld2019-05-032-11/+14
|
* setupapi: safer aliasing of slice typesJason A. Donenfeld2019-05-035-23/+40
|
* wintun: work around GetInterface staleness bugJason A. Donenfeld2019-05-032-18/+30
|
* wintun: don't retry when not creatingJason A. Donenfeld2019-05-022-9/+26
| | | | | | The only time we're trying to counteract the race condition is when we're creating a driver. When we're simply looking up all drivers, it doesn't make sense to retry.
* wintun: try harder to open registry keyJason A. Donenfeld2019-04-291-2/+2
| | | | | This sucks. Can we please find a deterministic way of doing this instead?
* go.mod: use vendored winioJason A. Donenfeld2019-04-292-1/+4
|
* tun: freebsd: work around numerous kernel panics on shutdownJason A. Donenfeld2019-04-231-23/+82
| | | | | | | | | There are numerous race conditions. But even this will crash it: while true; do ifconfig tun0 create; ifconfig tun0 destroy; done It seems like LLv6 is related, which we're not using anyway, so explicitly disable it on the interface.
* setupapi: Fix struct size mismatchesSimon Rozman2019-04-194-8/+27
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* setupapi: actually fix padding by rounding up to sizeof(void*)Jason A. Donenfeld2019-04-191-1/+1
|
* tun: darwin: write routeSocket variable in helperJason A. Donenfeld2019-04-191-2/+4
| | | | Otherwise the race detector "complains".
* main: revise warningsJason A. Donenfeld2019-04-192-25/+17
|
* device: send: check packet length before freeing elementJason A. Donenfeld2019-04-181-3/+3
|
* setupapi: Revise DrvInfoDetailData struct size calculationSimon Rozman2019-04-182-3/+7
| | | | | | | Go adds trailing padding to DrvInfoDetailData struct in GOARCH=386 which confuses SetupAPI expecting exactly sizeof(SP_DRVINFO_DETAIL_DATA). Signed-off-by: Simon Rozman <simon@rozman.si>
* windows: use proper constants from updated x/sysJason A. Donenfeld2019-04-135-25/+28
|
* conn: linux: RTA_MARK has moved to x/sysJason A. Donenfeld2019-04-131-1/+1
|
* tun: windows: Adopt new error codes returned by WintunSimon Rozman2019-04-111-3/+2
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: do not sleep after OPERATION_ABORTED on writeSimon Rozman2019-04-111-2/+1
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* main_windows: use proper version constantJason A. Donenfeld2019-04-091-1/+1
|
* version: put version in right place0.0.20190409Jason A. Donenfeld2019-04-092-4/+1
|
* version: bump snapshotJason A. Donenfeld2019-04-091-1/+1
|
* tun: windows: do not sleep after OPERATION_ABORTEDJason A. Donenfeld2019-04-091-2/+1
|
* tun: windows: Retry R/W on ERROR_OPERATION_ABORTEDSimon Rozman2019-04-041-18/+44
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: Attempt to reopen handle on all errorsSimon Rozman2019-04-031-15/+9
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* wintun: rename device using undocumented API that netsh.exe usesJason A. Donenfeld2019-04-012-1/+41
|
* wintun: add more retry loopsJason A. Donenfeld2019-04-013-30/+56
|
* tun: windows: cancel ongoing reads on closing and delete after closeJason A. Donenfeld2019-03-261-4/+11
| | | | | This reverts commit 52ec440d7977fad966002c3710ed9df957943407 and adds some spice.
* wintun: query for NetCfgInstanceId several timesJason A. Donenfeld2019-03-221-5/+16
|
* tun: windows: delete interface before deleting file handlesJason A. Donenfeld2019-03-221-9/+4
|
* tun: windows: Make adapter rename asynchronousSimon Rozman2019-03-222-60/+13
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: Adapter devices renamed to WINTUN<LUID Index>Simon Rozman2019-03-224-52/+63
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: Increase unavailable adapter timeout to 30secSimon Rozman2019-03-221-1/+1
| | | | | | 5 seconds was too short when debugging. Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: Make writing persistent tooSimon Rozman2019-03-221-15/+17
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: Fix paused adapter testSimon Rozman2019-03-221-1/+1
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* tun: windows: just open two file handlesJason A. Donenfeld2019-03-213-162/+83
|
* receive: implement flush semanticsJason A. Donenfeld2019-03-2112-98/+147
|
* tun: windows: add dummy overlapped events backJason A. Donenfeld2019-03-211-2/+21
| | | | These seem basically wrong to me, but we get crashes without them.
* tun: windows: use new constants in sysJason A. Donenfeld2019-03-203-13/+15
|
* uapi: remove unhelpful log messagesJason A. Donenfeld2019-03-201-7/+0
|
* wintun: Use native Win32 API for I/OSimon Rozman2019-03-215-38/+173
| | | | Signed-off-by: Simon Rozman <simon@rozman.si>
* uapi: report endpoint errorJason A. Donenfeld2019-03-191-1/+1
|
* tun: windows: wintun does iocpJason A. Donenfeld2019-03-182-142/+49
|
* tun: windows: temporary hack for forcing MTUJason A. Donenfeld2019-03-131-1/+8
|