Message ID | 56BDA67E.8060404@gmx.de |
---|---|
State | Not Applicable |
Headers | show |
Yann, is the patch o.k. for you? Shall I resend it using git? Please let me know. Regards, Andreas Am 12.02.2016 um 10:31 schrieb Andreas Ehmanns: > Yann, > indeed a killall -9 worked. So I reworked the patch. According to > Thomas' recommendation I added the OK/FAIL output as done in the > dropbox init script. > Please have a look at the attached patch. If it's o.k., I'll resend it > using git. > > Regards, > Andreas > > Am 04.02.2016 um 21:50 schrieb Yann E. MORIN: >> Andreas, All, >> >> On 2016-02-04 09:31 +0100, Andreas Ehmanns spake thusly: >>> Dear Yann, >>> please find below the rework of the nfsd init script. >>> The changes are very small. Please note that a change from >>> >>> kill -9 `pidof nfsd` 2>/dev/null >>> >>> to killall was not possible since killall did not kill the nfsd >>> threads. It >>> simply did not work and I couldn't figure out the root cause. >> What about using something like; >> rpc.nfsd 0 >> killall -9 nfsd >> >> From 'man rpc.nfsd': >> In particular rpc.nfsd 0 will stop all threads and thus close any >> open connections. >> >> (maybe with a little sleep in-between the two calls to let things settle >> down?) >> >> Care to have a look, please? >> >>> Let me know what you think about. If it's o.k. for you, I re-send >>> the patch >>> using git. >> Well, it's already a nice improvement. ;-) >> >> Please, consider the above suggestion. Adapt the script if required, and >> respin with git, please. >> >> Thanks! :-) >> >> Regards, >> Yann E. MORIN. >> >>> From 13fa8728220cfe80b7ab3d48837183088f4094d5 Mon Sep 17 00:00:00 2001 >>> From: Andreas Ehmanns <universeII@gmx.de> >>> Date: Thu, 4 Feb 2016 09:23:35 +0100 >>> Subject: [PATCH 1/1] Reordered kill, echo and pid file remove >>> >>> Signed-off-by: Andreas Ehmanns <universeII@gmx.de> >>> --- >>> package/nfs-utils/S60nfs | 7 ++++--- >>> 1 file changed, 4 insertions(+), 3 deletions(-) >>> >>> diff --git a/package/nfs-utils/S60nfs b/package/nfs-utils/S60nfs >>> index 5639b8f..0bd5a44 100755 >>> --- a/package/nfs-utils/S60nfs >>> +++ b/package/nfs-utils/S60nfs >>> @@ -38,8 +38,8 @@ start() { >>> printf "Starting NFS mountd: " >>> rpc.mountd >>> - echo "done" >>> touch /var/lock/subsys/nfs >>> + echo "done" >>> } >>> stop() { >>> @@ -54,14 +54,15 @@ stop() { >>> printf "Shutting down NFS services: " >>> /usr/sbin/exportfs -au >>> - rm -f /var/lock/subsys/nfs >>> killall -q rpc.statd >>> + rm -f /var/lock/subsys/nfs >>> echo "done" >>> printf "Stopping NFS statd: " >>> killall -q rpc.statd >>> - echo "done" >>> + rm -f /var/run/rpc.statd.pid >>> rm -f /var/lock/subsys/nfslock >>> + echo "done" >>> } >>> # See how we were called. >>> -- >>> 2.1.4 >>> >> > > > > _______________________________________________ > buildroot mailing list > buildroot@busybox.net > http://lists.busybox.net/mailman/listinfo/buildroot
On 02/25/16 21:40, Andreas Ehmanns wrote: > Yann, > is the patch o.k. for you? Shall I resend it using git? Yes, please send using git. Patches sent as attachments have an even bigger chance of being ignored than other patches. However, I can already say that it should be two separate patches: one changing "done" to OK/FAIL, and a second one doing the functional changes. Regards, Arnout > Please let me know. > > Regards, > Andreas > > Am 12.02.2016 um 10:31 schrieb Andreas Ehmanns: >> Yann, >> indeed a killall -9 worked. So I reworked the patch. According to Thomas' >> recommendation I added the OK/FAIL output as done in the dropbox init script. >> Please have a look at the attached patch. If it's o.k., I'll resend it using git. [snip]
Am 26.02.2016 um 00:30 schrieb Arnout Vandecappelle: > On 02/25/16 21:40, Andreas Ehmanns wrote: >> Yann, >> is the patch o.k. for you? Shall I resend it using git? > Yes, please send using git. Patches sent as attachments have an even bigger > chance of being ignored than other patches. > > However, I can already say that it should be two separate patches: one changing > "done" to OK/FAIL, and a second one doing the functional changes. > > Regards, > Arnout Yann, Arnout, can someone of you please check what happened with this patch? It's more than one month and a half ago that I sent it and it seems that it is still not incorporated? Is there something from my side what is missing? Additionally I sent a set of patches for the netsnmp init script one day before which are also not incorporated? Can you please have a look? Regards, Andreas > >> Please let me know. >> >> Regards, >> Andreas >> >> Am 12.02.2016 um 10:31 schrieb Andreas Ehmanns: >>> Yann, >>> indeed a killall -9 worked. So I reworked the patch. According to Thomas' >>> recommendation I added the OK/FAIL output as done in the dropbox init script. >>> Please have a look at the attached patch. If it's o.k., I'll resend it using git. > [snip] >
On 04/15/16 10:37, Andreas Ehmanns wrote: > Am 26.02.2016 um 00:30 schrieb Arnout Vandecappelle: >> On 02/25/16 21:40, Andreas Ehmanns wrote: >>> Yann, >>> is the patch o.k. for you? Shall I resend it using git? >> Yes, please send using git. Patches sent as attachments have an even bigger >> chance of being ignored than other patches. >> >> However, I can already say that it should be two separate patches: one changing >> "done" to OK/FAIL, and a second one doing the functional changes. >> >> Regards, >> Arnout > > Yann, Arnout, > can someone of you please check what happened with this patch? It's more than > one month and a half ago that I sent it and it seems that it is still not > incorporated? We are unfortunately swamped with patches that need review, and we don't have enough people doing reviews. And we're not willing to forego on quality by accepting patches without review. We will get to your patches eventually, because they are tracked in patchwork. Note that the oldest pending patch is 9 months old, to give you an idea of what you can expect. What you can do to speed things up: - Help out with reviewing and testing. Look at some (non-trivial) patches that are relevant to you (e.g. affecting a package that you use), download and test it, then give it your Tested-by tag. You can also go over the changes, look for any mistakes that you can identify, and give feedback or give it your Reviewed-by tag. Obviously, your reviews will not immediately carry the same weight as those of long-time buildroot developers, but they are really helpful. - Reply to your patches with 'ping', so they appear on the top of the stack again. Note that this doesn't help that much because we're still swamped. But it does at least tell us that you will be paying attention if we have any feedback to give. Regards, Arnout > Is there something from my side what is missing? > > Additionally I sent a set of patches for the netsnmp init script one day before > which are also not incorporated? > Can you please have a look? > > Regards, > Andreas > >> >>> Please let me know. >>> >>> Regards, >>> Andreas >>> >>> Am 12.02.2016 um 10:31 schrieb Andreas Ehmanns: >>>> Yann, >>>> indeed a killall -9 worked. So I reworked the patch. According to Thomas' >>>> recommendation I added the OK/FAIL output as done in the dropbox init script. >>>> Please have a look at the attached patch. If it's o.k., I'll resend it using >>>> git. >> [snip] >> >
Am 15.04.2016 um 11:00 schrieb Arnout Vandecappelle: > > > On 04/15/16 10:37, Andreas Ehmanns wrote: >> Am 26.02.2016 um 00:30 schrieb Arnout Vandecappelle: >>> On 02/25/16 21:40, Andreas Ehmanns wrote: >>>> Yann, >>>> is the patch o.k. for you? Shall I resend it using git? >>> Yes, please send using git. Patches sent as attachments have an >>> even bigger >>> chance of being ignored than other patches. >>> >>> However, I can already say that it should be two separate patches: >>> one changing >>> "done" to OK/FAIL, and a second one doing the functional changes. >>> >>> Regards, >>> Arnout >> >> Yann, Arnout, >> can someone of you please check what happened with this patch? It's >> more than >> one month and a half ago that I sent it and it seems that it is still >> not >> incorporated? > > We are unfortunately swamped with patches that need review, and we > don't have enough people doing reviews. And we're not willing to > forego on quality by accepting patches without review. > > We will get to your patches eventually, because they are tracked in > patchwork. Note that the oldest pending patch is 9 months old, to give > you an idea of what you can expect. Arnout, I wasn't aware of that. So it seems that I was a little bit too impatient. Thanks for the clarification and I will have a look if I can follow your recommendation and review some patches. Regards, Andreas > > > What you can do to speed things up: > > - Help out with reviewing and testing. Look at some (non-trivial) > patches that are relevant to you (e.g. affecting a package that you > use), download and test it, then give it your Tested-by tag. You can > also go over the changes, look for any mistakes that you can identify, > and give feedback or give it your Reviewed-by tag. Obviously, your > reviews will not immediately carry the same weight as those of > long-time buildroot developers, but they are really helpful. > > - Reply to your patches with 'ping', so they appear on the top of the > stack again. Note that this doesn't help that much because we're still > swamped. But it does at least tell us that you will be paying > attention if we have any feedback to give. > > > Regards, > Arnout > >> Is there something from my side what is missing? >> >> Additionally I sent a set of patches for the netsnmp init script one >> day before >> which are also not incorporated? >> Can you please have a look? >> >> Regards, >> Andreas >> >>> >>>> Please let me know. >>>> >>>> Regards, >>>> Andreas >>>> >>>> Am 12.02.2016 um 10:31 schrieb Andreas Ehmanns: >>>>> Yann, >>>>> indeed a killall -9 worked. So I reworked the patch. According to >>>>> Thomas' >>>>> recommendation I added the OK/FAIL output as done in the dropbox >>>>> init script. >>>>> Please have a look at the attached patch. If it's o.k., I'll >>>>> resend it using >>>>> git. >>> [snip] >>> >> >
On 04/15/16 11:09, Andreas Ehmanns wrote: > Am 15.04.2016 um 11:00 schrieb Arnout Vandecappelle: >> >> >> On 04/15/16 10:37, Andreas Ehmanns wrote: >>> Am 26.02.2016 um 00:30 schrieb Arnout Vandecappelle: >>>> On 02/25/16 21:40, Andreas Ehmanns wrote: >>>>> Yann, >>>>> is the patch o.k. for you? Shall I resend it using git? >>>> Yes, please send using git. Patches sent as attachments have an even bigger >>>> chance of being ignored than other patches. >>>> >>>> However, I can already say that it should be two separate patches: one >>>> changing >>>> "done" to OK/FAIL, and a second one doing the functional changes. >>>> >>>> Regards, >>>> Arnout >>> >>> Yann, Arnout, >>> can someone of you please check what happened with this patch? It's more than >>> one month and a half ago that I sent it and it seems that it is still not >>> incorporated? >> >> We are unfortunately swamped with patches that need review, and we don't have >> enough people doing reviews. And we're not willing to forego on quality by >> accepting patches without review. >> >> We will get to your patches eventually, because they are tracked in >> patchwork. Note that the oldest pending patch is 9 months old, to give you an >> idea of what you can expect. > > Arnout, > I wasn't aware of that. So it seems that I was a little bit too impatient. I think your impatience is very reasonable. This point is discussed at about every buildroot developer meeting. Asking people to do more reviews is about the only thing we have come up with. If you have other ideas, feel free to launch them. Well, actually, we do have one other solution: the Buildroot Summer Camp, where a bunch of core developers come together for 4-5 days for a long hacking session. Last year we managed to get about 300 patches applied, and developed some important new features like the kernel-module infra, better support for systemd, ... > > Thanks for the clarification and I will have a look if I can follow your > recommendation and review some patches. Thanks! That's really appreciated! Regards, Arnout > > Regards, > Andreas > >> >> >> What you can do to speed things up: >> >> - Help out with reviewing and testing. Look at some (non-trivial) patches that >> are relevant to you (e.g. affecting a package that you use), download and test >> it, then give it your Tested-by tag. You can also go over the changes, look >> for any mistakes that you can identify, and give feedback or give it your >> Reviewed-by tag. Obviously, your reviews will not immediately carry the same >> weight as those of long-time buildroot developers, but they are really helpful. >> >> - Reply to your patches with 'ping', so they appear on the top of the stack >> again. Note that this doesn't help that much because we're still swamped. But >> it does at least tell us that you will be paying attention if we have any >> feedback to give. >> >> >> Regards, >> Arnout >> >>> Is there something from my side what is missing? >>> >>> Additionally I sent a set of patches for the netsnmp init script one day before >>> which are also not incorporated? >>> Can you please have a look? >>> >>> Regards, >>> Andreas >>> >>>> >>>>> Please let me know. >>>>> >>>>> Regards, >>>>> Andreas >>>>> >>>>> Am 12.02.2016 um 10:31 schrieb Andreas Ehmanns: >>>>>> Yann, >>>>>> indeed a killall -9 worked. So I reworked the patch. According to Thomas' >>>>>> recommendation I added the OK/FAIL output as done in the dropbox init script. >>>>>> Please have a look at the attached patch. If it's o.k., I'll resend it using >>>>>> git. >>>> [snip] >>>> >>> >> >
From 3e631fae53dce78d6b2f10cb91805f2eee34e34d Mon Sep 17 00:00:00 2001 From: Andreas Ehmanns <universeII@gmx.de> Date: Fri, 12 Feb 2016 10:19:43 +0100 Subject: [PATCH 1/1] Remorked nfsd init script and added OK/FAIL output. Signed-off-by: Andreas Ehmanns <universeII@gmx.de> --- package/nfs-utils/S60nfs | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/package/nfs-utils/S60nfs b/package/nfs-utils/S60nfs index 5639b8f..4b510c6 100755 --- a/package/nfs-utils/S60nfs +++ b/package/nfs-utils/S60nfs @@ -25,20 +25,20 @@ start() { # Start daemons. printf "Starting NFS statd: " rpc.statd + [ $? = 0 ] && echo "OK" || echo "FAIL" touch /var/lock/subsys/nfslock - echo "done" printf "Starting NFS services: " /usr/sbin/exportfs -r - echo "done" + [ $? = 0 ] && echo "OK" || echo "FAIL" printf "Starting NFS daemon: " rpc.nfsd ${NR_THREADS} - echo "done" + [ $? = 0 ] && echo "OK" || echo "FAIL" printf "Starting NFS mountd: " rpc.mountd - echo "done" + [ $? = 0 ] && echo "OK" || echo "FAIL" touch /var/lock/subsys/nfs } @@ -46,21 +46,21 @@ stop() { # Stop daemons. printf "Shutting down NFS mountd: " killall -q rpc.mountd - echo "done" + [ $? = 0 ] && echo "OK" || echo "FAIL" printf "Shutting down NFS daemon: " - kill -9 `pidof nfsd` 2>/dev/null - echo "done" + killall -q -9 nfsd 2>/dev/null + [ $? = 0 ] && echo "OK" || echo "FAIL" printf "Shutting down NFS services: " /usr/sbin/exportfs -au - rm -f /var/lock/subsys/nfs - killall -q rpc.statd - echo "done" + [ $? = 0 ] && echo "OK" || echo "FAIL" printf "Stopping NFS statd: " killall -q rpc.statd - echo "done" + [ $? = 0 ] && echo "OK" || echo "FAIL" + rm -f /var/lock/subsys/nfs + rm -f /var/run/rpc.statd.pid rm -f /var/lock/subsys/nfslock } -- 2.1.4