Re: rpc: use monotime for timeouts

2017-12-14 Thread Todd C. Miller
On Thu, 14 Dec 2017 18:00:08 +0100, Jeremie Courreges-Anglas wrote: > Here's another diff on top to remove two useless temporary variables, > and unify the names of the remaining ones. Looks correct, OK millert@ - todd

Re: rpc: use monotime for timeouts

2017-12-14 Thread Jeremie Courreges-Anglas
Diff committed, thank you Scott. Here's another diff on top to remove two useless temporary variables, and unify the names of the remaining ones. ok? Index: clnt_tcp.c === --- clnt_tcp.c.orig +++ clnt_tcp.c @@ -385,7 +385,7 @@ sta

Re: rpc: use monotime for timeouts

2017-12-13 Thread Jeremie Courreges-Anglas
On Tue, Dec 12 2017, Jeremie Courreges-Anglas wrote: > On Sun, Dec 10 2017, Scott Cheloha wrote: >> Hi, >> >> These timeouts in sunrpc need to be based on the monotonic >> clock to avoid a race with adjtime(2), settimeofday(2), etc. >> >> There are obvious possible improvements here and elsewhere

Re: rpc: use monotime for timeouts

2017-12-11 Thread Jeremie Courreges-Anglas
On Sun, Dec 10 2017, Scott Cheloha wrote: > Hi, > > These timeouts in sunrpc need to be based on the monotonic > clock to avoid a race with adjtime(2), settimeofday(2), etc. > > There are obvious possible improvements here and elsewhere > in sunrpc. Here especially the time-related variable names

rpc: use monotime for timeouts

2017-12-09 Thread Scott Cheloha
Hi, These timeouts in sunrpc need to be based on the monotonic clock to avoid a race with adjtime(2), settimeofday(2), etc. There are obvious possible improvements here and elsewhere in sunrpc. Here especially the time-related variable names could be made more descriptive, the various BSD time m