| Message ID | ZKpkDG2kTWVFSNiZ@eldamar.lan (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Always run rpc-pipefs-generator generator (was: Re: Why keep var-lib-nfs-rpc_pipefs.mount around?) | expand |
On 9 Jul 2023, at 3:38, Salvatore Bonaccorso wrote: > Hi Steve, ... > FWIW, in Debian we have applied the respective change. The idea would > be to only depend on a single mechanism for setting up the mounts > rather than a combination of the two (the generator and the static > mount unit). For this reason we have applied the attached patch, and > are not installing the units that we will let the generator produce, > that is var-lib-nfs-rpc_pipefs.mount and rpc_pipefs.target > > We in Debian for long have diverged too much from you upstream, > causing that we lacked behind several new upstream version and stuck > with old versions in stable releases. We want to avoid running into > that again in future. So if this make sense to you, would you apply > the same (or as you prefer similar) change to you upstream? > > On one side so you could apply Andreas Hasenack patch, secondly > installing the var-lib-nfs-rpc_pipefs.mount and rpc_pipefs.target > could be dropped (note no changes to the other units needed as the > repsective needed dependencies are generated by the systemd > generator). > > Ben, Andreas, please add what else is needed from your point of view > please! I don't think I've seen the PATCH land on the list addressed to nfs-utils maintainer yet, but I could have missed it. Otherwise it looks sane to me, but I could be missing some upstream case. > Thanks a lot for considering this. If you have any suggestion further > how we can unify the Debian downstream to you upstream, let us know > please. At Red Hat, we use "upstream first" as a leading principle. If this change makes sense for upstream, send Adreas' patch along and I am sure Steve D will consider it or let us know why its not acceptible for upstream. Ben
Hi, On Mon, Jul 10, 2023 at 10:39:43AM -0400, Benjamin Coddington wrote: > On 9 Jul 2023, at 3:38, Salvatore Bonaccorso wrote: > > > Hi Steve, > > ... > > > FWIW, in Debian we have applied the respective change. The idea would > > be to only depend on a single mechanism for setting up the mounts > > rather than a combination of the two (the generator and the static > > mount unit). For this reason we have applied the attached patch, and > > are not installing the units that we will let the generator produce, > > that is var-lib-nfs-rpc_pipefs.mount and rpc_pipefs.target > > > > We in Debian for long have diverged too much from you upstream, > > causing that we lacked behind several new upstream version and stuck > > with old versions in stable releases. We want to avoid running into > > that again in future. So if this make sense to you, would you apply > > the same (or as you prefer similar) change to you upstream? > > > > On one side so you could apply Andreas Hasenack patch, secondly > > installing the var-lib-nfs-rpc_pipefs.mount and rpc_pipefs.target > > could be dropped (note no changes to the other units needed as the > > repsective needed dependencies are generated by the systemd > > generator). > > > > Ben, Andreas, please add what else is needed from your point of view > > please! > > I don't think I've seen the PATCH land on the list addressed to nfs-utils > maintainer yet, but I could have missed it. > > Otherwise it looks sane to me, but I could be missing some upstream case. > > > Thanks a lot for considering this. If you have any suggestion further > > how we can unify the Debian downstream to you upstream, let us know > > please. > > At Red Hat, we use "upstream first" as a leading principle. If this change > makes sense for upstream, send Adreas' patch along and I am sure Steve D will > consider it or let us know why its not acceptible for upstream. Andreas, could you sent a proper patchset please, so upstream can have a look at it for inclusion? Regards, Salvatore
On it, need to refresh some knowledge and think with an upstream hat on now :) On Mon, Jul 24, 2023 at 6:12 AM Salvatore Bonaccorso <carnil@debian.org> wrote: > > Hi, > > On Mon, Jul 10, 2023 at 10:39:43AM -0400, Benjamin Coddington wrote: > > On 9 Jul 2023, at 3:38, Salvatore Bonaccorso wrote: > > > > > Hi Steve, > > > > ... > > > > > FWIW, in Debian we have applied the respective change. The idea would > > > be to only depend on a single mechanism for setting up the mounts > > > rather than a combination of the two (the generator and the static > > > mount unit). For this reason we have applied the attached patch, and > > > are not installing the units that we will let the generator produce, > > > that is var-lib-nfs-rpc_pipefs.mount and rpc_pipefs.target > > > > > > We in Debian for long have diverged too much from you upstream, > > > causing that we lacked behind several new upstream version and stuck > > > with old versions in stable releases. We want to avoid running into > > > that again in future. So if this make sense to you, would you apply > > > the same (or as you prefer similar) change to you upstream? > > > > > > On one side so you could apply Andreas Hasenack patch, secondly > > > installing the var-lib-nfs-rpc_pipefs.mount and rpc_pipefs.target > > > could be dropped (note no changes to the other units needed as the > > > repsective needed dependencies are generated by the systemd > > > generator). > > > > > > Ben, Andreas, please add what else is needed from your point of view > > > please! > > > > I don't think I've seen the PATCH land on the list addressed to nfs-utils > > maintainer yet, but I could have missed it. > > > > Otherwise it looks sane to me, but I could be missing some upstream case. > > > > > Thanks a lot for considering this. If you have any suggestion further > > > how we can unify the Debian downstream to you upstream, let us know > > > please. > > > > At Red Hat, we use "upstream first" as a leading principle. If this change > > makes sense for upstream, send Adreas' patch along and I am sure Steve D will > > consider it or let us know why its not acceptible for upstream. > > Andreas, could you sent a proper patchset please, so upstream can have > a look at it for inclusion? > > Regards, > Salvatore
Description: Always run the generator Run the generator even if the pipefs-directory setting is the default one. Author: Andreas Hasenack <andreas@canonical.com> Bug-Ubuntu: https://bugs.launchpad.net/bugs/1971935 Bug-Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1014429 Forwarded: https://lore.kernel.org/linux-nfs/EE39279C-4E40-48C8-ABC9-707EB1AD6D79@redhat.com/ Last-Update: 2022-07-12 --- This patch header follows DEP-3: http://dep.debian.net/deps/dep3/ diff --git a/systemd/rpc-pipefs-generator.c b/systemd/rpc-pipefs-generator.c index c24db567..7c42431f 100644 --- a/systemd/rpc-pipefs-generator.c +++ b/systemd/rpc-pipefs-generator.c @@ -139,9 +139,6 @@ int main(int argc, char *argv[]) s = conf_get_str("general", "pipefs-directory"); if (!s) exit(0); - if (strlen(s) == strlen(RPC_PIPEFS_DEFAULT) && - strcmp(s, RPC_PIPEFS_DEFAULT) == 0) - exit(0); if (is_non_pipefs_mountpoint(s)) exit(1);