Hi! ---- Does anyone know out-of-the-head whether tmpfs supports ACLs - and if "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs etc.) are supported by tmpfs ? ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;)
Norm Jacobs
2009-Sep-16 02:49 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfs support ?
I have some vague recollection that tmpfs doesn''t support ACLs snd it
appears to be so...
ZFS
opensolaris% touch /var/tmp/bar
opensolaris% chmod A=user:lp:r:deny /var/tmp/bar
opensolaris%
TMPFS
opensolaris% touch /tmp/bar
opensolaris% chmod A=user:lp:r:deny /tmp/bar
chmod: ERROR: Failed to set ACL: Operation not supported
opensolaris%
-Norm
Roland Mainz wrote:> Hi!
>
> ----
>
> Does anyone know out-of-the-head whether tmpfs supports ACLs - and if
> "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft
ACLs
> etc.) are supported by tmpfs ?
>
> ----
>
> Bye,
> Roland
>
>
Mark Shellenbaum
2009-Sep-16 03:00 UTC
[zfs-discuss] Which kind of ACLs does tmpfs support ?
Roland Mainz wrote:> Hi! > > ---- > > Does anyone know out-of-the-head whether tmpfs supports ACLs - and if > "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs > etc.) are supported by tmpfs ? >tmpfs does not support ACLs see _PC_ACL_ENABLED in [f]pathconf(2). You can query the file system for what type of ACLs it supports.> ---- > > Bye, > Roland >
Roland Mainz
2009-Sep-16 03:07 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfs support ?
Norm Jacobs wrote:> Roland Mainz wrote: > > Does anyone know out-of-the-head whether tmpfs supports ACLs - and if > > "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs > > etc.) are supported by tmpfs ? > > I have some vague recollection that tmpfs doesn''t support ACLs snd it > appears to be so...Is there any RFE which requests the implementation of NFSv4-like ACLs for tmpfs yet ?> ZFS > > opensolaris% touch /var/tmp/bar > opensolaris% chmod A=user:lp:r:deny /var/tmp/bar > opensolaris% > > TMPFS > > opensolaris% touch /tmp/bar > opensolaris% chmod A=user:lp:r:deny /tmp/bar > chmod: ERROR: Failed to set ACL: Operation not supported > opensolaris%Ok... does that mean that I have to create a ZFS filesystem to actually test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any other options ? [1]=The idea is to have a test module which checks whether ACL operations work correctly, however the testing framework must only run as normal, unpriviledged user... ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;)
Interesting question takes a few minutes to test... http://docs.sun.com/app/docs/doc/819-2252/acl-5?l=en&a=view&q=acl%285%29+ http://docs.sun.com/app/docs/doc/819-2239/chmod-1?l=en&a=view ZFS [tp47565@norton:] df . Filesystem size used avail capacity Mounted on rpool/export/home/tp47565 16G 1.2G 9.7G 11% /export/home/tp47565 [tp47565@norton:] touch file.3 [tp47565@norton:] ls -v file.3 -rw-r----- 1 tp47565 staff 0 Sep 16 15:02 file.3 0:owner@:execute:deny 1:owner@:read_data/write_data/append_data/write_xattr/write_attributes /write_acl/write_owner:allow 2:group@:write_data/append_data/execute:deny 3:group@:read_data:allow 4:everyone@:read_data/write_data/append_data/write_xattr/execute /write_attributes/write_acl/write_owner:deny 5:everyone@:read_xattr/read_attributes/read_acl/synchronize:allow [tp47565@norton:] chmod A+user:lp:read_data:deny file.3 [tp47565@norton:] ls -v file.3 -rw-r-----+ 1 tp47565 staff 0 Sep 16 15:02 file.3 0:user:lp:read_data:deny 1:owner@:execute:deny 2:owner@:read_data/write_data/append_data/write_xattr/write_attributes /write_acl/write_owner:allow 3:group@:write_data/append_data/execute:deny 4:group@:read_data:allow 5:everyone@:read_data/write_data/append_data/write_xattr/execute /write_attributes/write_acl/write_owner:deny 6:everyone@:read_xattr/read_attributes/read_acl/synchronize:allow [tp47565@norton:] Let''s try the new ACLs on tmpfs [tp47565@norton:] cd /tmp [tp47565@norton:] df . Filesystem size used avail capacity Mounted on swap 528M 12K 528M 1% /tmp [tp47565@norton:] grep swap /etc/vfstab swap - /tmp tmpfs - yes - /dev/zvol/dsk/rpool/swap - - swap - no - [tp47565@norton:] [tp47565@norton:] touch file.3 [tp47565@norton:] ls -v file.3 -rw-r----- 1 tp47565 staff 0 Sep 16 14:58 file.3 0:user::rw- 1:group::r-- #effective:r-- 2:mask:rwx 3:other:--- [tp47565@norton:] [tp47565@norton:] chmod A+user:lp:read_data:deny file.3 chmod: ERROR: ACL type''s are different [tp47565@norton:] So tmpfs does not support the new ACLs Do I have to do the old way as well? Roland Mainz wrote: Hi! ---- Does anyone know out-of-the-head whether tmpfs supports ACLs - and if "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs etc.) are supported by tmpfs ? ---- Bye, Roland -- Trevor Pretty | Technical Account Manager | +64 9 639 0652 | +64 21 666 161 Eagle Technology Group Ltd. Gate D, Alexandra Park, Greenlane West, Epsom Private Bag 93211, Parnell, Auckland www.eagle.co.nz This email is confidential and may be legally privileged. If received in error please destroy and immediately notify us. _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Ian Collins
2009-Sep-16 03:16 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfs support ?
Roland Mainz wrote:> Norm Jacobs wrote: > >> Roland Mainz wrote: >> >>> Does anyone know out-of-the-head whether tmpfs supports ACLs - and if >>> "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs >>> etc.) are supported by tmpfs ? >>> >> I have some vague recollection that tmpfs doesn''t support ACLs snd it >> appears to be so... >> > > Is there any RFE which requests the implementation of NFSv4-like ACLs > for tmpfs yet ? > > >> ZFS >> >> opensolaris% touch /var/tmp/bar >> opensolaris% chmod A=user:lp:r:deny /var/tmp/bar >> opensolaris% >> >> TMPFS >> >> opensolaris% touch /tmp/bar >> opensolaris% chmod A=user:lp:r:deny /tmp/bar >> chmod: ERROR: Failed to set ACL: Operation not supported >> opensolaris% >> > > Ok... does that mean that I have to create a ZFS filesystem to actually > test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any > other options ?Use function interposition. I am currently updating an ACL manipulation application and I use mocks for the acl get/set functions. -- Ian.
Roland Mainz
2009-Sep-16 03:23 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Ian Collins wrote:> Roland Mainz wrote: > > Norm Jacobs wrote: > >> Roland Mainz wrote: > >>> Does anyone know out-of-the-head whether tmpfs supports ACLs - and if > >>> "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs > >>> etc.) are supported by tmpfs ? > >>> > >> I have some vague recollection that tmpfs doesn''t support ACLs snd it > >> appears to be so... > > > > Is there any RFE which requests the implementation of NFSv4-like ACLs > > for tmpfs yet ? > > > >> ZFS > >> > >> opensolaris% touch /var/tmp/bar > >> opensolaris% chmod A=user:lp:r:deny /var/tmp/bar > >> opensolaris% > >> > >> TMPFS > >> > >> opensolaris% touch /tmp/bar > >> opensolaris% chmod A=user:lp:r:deny /tmp/bar > >> chmod: ERROR: Failed to set ACL: Operation not supported > >> opensolaris% > >> > > > > Ok... does that mean that I have to create a ZFS filesystem to actually > > test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any > > other options ? > Use function interposition.Umpf... the matching code is linked with -Bdirect ... AFAIK I can''t interpose library functions linked with this option, right ? ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;)
Ian Collins
2009-Sep-16 03:29 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Roland Mainz wrote:> Ian Collins wrote: > >> Roland Mainz wrote: >> >>> Norm Jacobs wrote: >>> >>>> Roland Mainz wrote: >>>> >>>>> Does anyone know out-of-the-head whether tmpfs supports ACLs - and if >>>>> "yes" - which type(s) of ACLs (e.g. NFSv4/ZFS, old POSIX draft ACLs >>>>> etc.) are supported by tmpfs ? >>>>> >>>>> >>>> I have some vague recollection that tmpfs doesn''t support ACLs snd it >>>> appears to be so... >>>> >>> Is there any RFE which requests the implementation of NFSv4-like ACLs >>> for tmpfs yet ? >>> >>> >>>> ZFS >>>> >>>> opensolaris% touch /var/tmp/bar >>>> opensolaris% chmod A=user:lp:r:deny /var/tmp/bar >>>> opensolaris% >>>> >>>> TMPFS >>>> >>>> opensolaris% touch /tmp/bar >>>> opensolaris% chmod A=user:lp:r:deny /tmp/bar >>>> chmod: ERROR: Failed to set ACL: Operation not supported >>>> opensolaris% >>>> >>>> >>> Ok... does that mean that I have to create a ZFS filesystem to actually >>> test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any >>> other options ? >>> >> Use function interposition. >> > > Umpf... the matching code is linked with -Bdirect ... AFAIK I can''t > interpose library functions linked with this option, right ? > >I never build test harnesses with explicit ld options (I use the C or C++ compiler for linking). There is a note about interposition in the ld man page description of -B. -- Ian.
Robert Thurlow
2009-Sep-16 03:35 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfs support ?
Roland Mainz wrote:> Ok... does that mean that I have to create a ZFS filesystem to actually > test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any > other options ?By all means, test with ZFS. But it''s easy to do that: # mkfile 64m /zpool.file # zpool create test /zpool.file # zfs list test 67.5K 27.4M 18K /test Rob T
Roland Mainz
2009-Sep-16 04:14 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Robert Thurlow wrote:> Roland Mainz wrote: > > > Ok... does that mean that I have to create a ZFS filesystem to actually > > test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any > > other options ? > > By all means, test with ZFS. But it''s easy to do that: > > # mkfile 64m /zpool.file > # zpool create test /zpool.file > # zfs list > test 67.5K 27.4M 18K /testI know... but AFAIK this requires "root" priviledges which the test suite won''t have... ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;)
Ali Bahrami
2009-Sep-16 04:17 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Roland Mainz wrote:> Umpf... the matching code is linked with -Bdirect ... AFAIK I can''t > interpose library functions linked with this option, right ? >You could set LD_NODIRECT to defeat direct bindings --- see ld.so.1(1). However, I agree with the thought that it would be easier to just have a ZFS filesystem to test against. - Ali
Alan Coopersmith
2009-Sep-16 04:20 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Roland Mainz wrote:> Robert Thurlow wrote: >> Roland Mainz wrote: >> >>> Ok... does that mean that I have to create a ZFS filesystem to actually >>> test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any >>> other options ? >> By all means, test with ZFS. But it''s easy to do that: >> >> # mkfile 64m /zpool.file >> # zpool create test /zpool.file >> # zfs list >> test 67.5K 27.4M 18K /test > > I know... but AFAIK this requires "root" priviledges which the test > suite won''t have...If the test suite is going to be running on nv_128 or later, then you are guaranteed to have a zfs filesystem, since root must be zfs then (since the only install method will be IPS, which requires zfs root). Until then you could just document to run it on a system with a zfs filesystem available. -- -Alan Coopersmith- alan.coopersmith at sun.com Sun Microsystems, Inc. - X Window System Engineering
Norm Jacobs
2009-Sep-16 04:22 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Roland Mainz wrote:> Robert Thurlow wrote: > >> Roland Mainz wrote: >> >> >>> Ok... does that mean that I have to create a ZFS filesystem to actually >>> test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any >>> other options ? >>> >> By all means, test with ZFS. But it''s easy to do that: >> >> # mkfile 64m /zpool.file >> # zpool create test /zpool.file >> # zfs list >> test 67.5K 27.4M 18K /test >> > > I know... but AFAIK this requires "root" priviledges which the test > suite won''t have... > > ---- > > Bye, > Roland > >You can delegate the ability to create ZFS filesystems. -Norm
Ian Collins
2009-Sep-16 04:58 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Roland Mainz wrote:> Robert Thurlow wrote: > >> Roland Mainz wrote: >> >> >>> Ok... does that mean that I have to create a ZFS filesystem to actually >>> test ([1]) an application which modifies ZFS/NFSv4 ACLs or are there any >>> other options ? >>> >> By all means, test with ZFS. But it''s easy to do that: >> >> # mkfile 64m /zpool.file >> # zpool create test /zpool.file >> # zfs list >> test 67.5K 27.4M 18K /test >> > > I know... but AFAIK this requires "root" priviledges which the test > suite won''t have... >It is also more difficult to test error conditions. Unless you really have to, don''t use fancy link options for a test harness so you can easily interpose (or simply define) mock library calls. If you stick with regular dynamic linking, you can simply define your mock/stub library functions in your test code. Easy. -- Ian.
Joerg Schilling
2009-Sep-16 12:28 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Alan Coopersmith <Alan.Coopersmith at Sun.COM> wrote:> If the test suite is going to be running on nv_128 or later, then > you are guaranteed to have a zfs filesystem, since root must be > zfs then (since the only install method will be IPS, which requires > zfs root). Until then you could just document to run it on a > system with a zfs filesystem available.Are you kidding? An installer that depends on the availability of a specific root filesystem is based on a design bug. What is the reason? Is there still no pkgrm command in IPS? If Yes, then it is obvious that IPS is not yet ready for production use. J?rg -- EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin js at cs.tu-berlin.de (uni) joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Darren J Moffat
2009-Sep-16 12:45 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Joerg Schilling wrote:> Alan Coopersmith <Alan.Coopersmith at Sun.COM> wrote: > >> If the test suite is going to be running on nv_128 or later, then >> you are guaranteed to have a zfs filesystem, since root must be >> zfs then (since the only install method will be IPS, which requires >> zfs root). Until then you could just document to run it on a >> system with a zfs filesystem available. > > Are you kidding?pkg will install just fine onto other filesystems eg: $ pkg image-create -a opensolaris.org=http://pkg.opensolaris.org/dev /tmp/darrenm/myimage $ pkg -R /tmp/darrenm/myimage install SUNWmkcd /tmp is tmpfs on this sytem and note that the prompt is $ not # pkg doesn''t even require you be root. Whats more pkg isn''t even specific to Solaris it works on other UNIX like systems and even on Windows. So it is clearly false that it requires ZFS since that doesn''t even exist on all the other platforms. Don''t confuse the decisions made for the OpenSolaris 2009.06 releases of only supporting and allowing install to ZFS root with the capabilities of the package system.> An installer that depends on the availability of a specific root filesystem > is based on a design bug.> What is the reason? Is there still no pkgrm command in IPS? > If Yes, then it is obvious that IPS is not yet ready for production use.See the man page for pkg: pkg uninstall [-nrvq] [--no-index] package... If you want to discuss pkg capabilities more please continue this discussion on pkg-discuss at opensolaris.org where the pkg experts will be happy to engage. -- Darren J Moffat
Alan Coopersmith
2009-Sep-16 14:29 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
Joerg Schilling wrote:> Alan Coopersmith <Alan.Coopersmith at Sun.COM> wrote: > >> If the test suite is going to be running on nv_128 or later, then >> you are guaranteed to have a zfs filesystem, since root must be >> zfs then (since the only install method will be IPS, which requires >> zfs root). Until then you could just document to run it on a >> system with a zfs filesystem available. > > Are you kidding?No, though as Darren clarified, it''s the IPS-based installers and pkg image-update mechanism that rely on ZFS, since boot environments are managed using ZFS snapshots/clones, much as Live Upgrade does on ZFS systems.> An installer that depends on the availability of a specific root filesystem > is based on a design bug.The installer used in Solaris 2.0 through the original release of 10 required UFS as the root filesystem - that wasn''t a design bug, just the way it was designed.> Is there still no pkgrm command in IPS?pkgrm is the command for SVR4 packages. pkg uninstall is the IPS equivalent and has of course been there since the first public release of IPS. Perhaps you should try using IPS before telling the world how broken your understanding of it is. -- -Alan Coopersmith- alan.coopersmith at sun.com Sun Microsystems, Inc. - X Window System Engineering
David Dyer-Bennet
2009-Sep-16 17:06 UTC
[zfs-discuss] [osol-discuss] Which kind of ACLs does tmpfssupport?
On Wed, September 16, 2009 09:29, Alan Coopersmith wrote:> The installer used in Solaris 2.0 through the original release of 10 > required > UFS as the root filesystem - that wasn''t a design bug, just the way it was > designed.If there are multiple filesystems available, an installer that forces root to be one specific one of them is suboptimal. "Design bug" is a bit too absolute; sometimes it''s quite hard, sometimes the number of people who might reasonably want anything but one of the choices is very small, etc. In the current case, the use of snapshots to handle keeping boot environments straight is just so elegant, it makes some sense to me to have an installation approach that requires it. If future filesystems also support snapshots, then it''ll start to look more like a bug again if the installer forces you to use one specific one. -- David Dyer-Bennet, dd-b at dd-b.net; http://dd-b.net/ Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/ Photos: http://dd-b.net/photography/gallery/ Dragaera: http://dragaera.info