Re: Reviewing b58e4f6b3d7329.... - Samba

This is a discussion on Re: Reviewing b58e4f6b3d7329.... - Samba ; Karolin Seeger wrote: > Jerry, > > On Sat, Jun 28, 2008 at 09:31:01AM -0400, Gerald (Jerry) Carter wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> Karolin, >> >> I'm not sure this is correct. >> >> ...

+ Reply to Thread
Results 1 to 7 of 7

Thread: Re: Reviewing b58e4f6b3d7329....

  1. Re: Reviewing b58e4f6b3d7329....

    Karolin Seeger wrote:
    > Jerry,
    >
    > On Sat, Jun 28, 2008 at 09:31:01AM -0400, Gerald (Jerry) Carter wrote:
    >> -----BEGIN PGP SIGNED MESSAGE-----
    >> Hash: SHA1
    >>
    >> Karolin,
    >>
    >> I'm not sure this is correct.
    >>
    >> commit b58e4f6b3d73294d8448c0dff4341183c52e5b7c
    >> Author: Karolin Seeger
    >> Date: Mon Jun 16 15:21:28 2008 +0200

    .....
    >> Would you provide an exampole of what bug is it suposed to fix?
    >> Thanks.

    >
    > The problem was that sid_to_name requests were returning inconsistent values.
    > During name_to_sid requests, a reverse caching entry was added to the
    > Winbindd cache. Name_to_sid requests can be made for different
    > combinations of lower case and upper case as it depends on the user/client.
    > Following sid_to_name requests were taken out of the cache and returned the
    > name in the same notation as the name was given in a sid_to_name request
    > before.
    >
    > Example (without patch):


    .....snip....
    >
    > 'wbinfo -s' returns EXAMPLE\administrator although the real name is
    > EXAMPLE\Administrator.
    >
    > The patch ensures that no reverse caching entries are created and Winbindd
    > asks the DC for sid_to_name requests.


    Was there a functional bug here? The report is academic IMO. So
    we've traded have of the cache benefit for ensuring that
    Sid2name returns proper case. but smbd looks at names in case
    insenstive fashion and the NSS calls are normalized to lower
    case.

    > Please inform me as soon as possible if this patch should be removed
    > before we ship the final release.


    Unless there is a functional bug here, I would like to revert it.




    cheers, jerry


  2. Re: Reviewing b58e4f6b3d7329....

    On Mon, Jun 30, 2008 at 11:30:33AM -0500, Gerald (Jerry) Carter wrote:

    > Was there a functional bug here? The report is academic IMO. So
    > we've traded have of the cache benefit for ensuring that
    > Sid2name returns proper case. but smbd looks at names in case
    > insenstive fashion and the NSS calls are normalized to lower
    > case.
    >
    > > Please inform me as soon as possible if this patch should be removed
    > > before we ship the final release.

    >
    > Unless there is a functional bug here, I would like to revert it.


    Well, we should at least return consistent upper/lower case
    IMO. Not something that the user can influence.

    Volker

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.5 (GNU/Linux)

    iD8DBQFIaS8WUzqjrWwMRl0RAoBaAKCWK3YAK9FzGIZ1rNmrY0 yj2kIkuQCffyQE
    9R9OtqvKqG4dtgl8pzgey58=
    =X3oF
    -----END PGP SIGNATURE-----


  3. Re: Reviewing b58e4f6b3d7329....

    Jerry, Volker,

    On Mon, Jun 30, 2008 at 09:08:07PM +0200, Volker Lendecke wrote:
    > On Mon, Jun 30, 2008 at 11:30:33AM -0500, Gerald (Jerry) Carter wrote:
    >
    > > Was there a functional bug here? The report is academic IMO. So
    > > we've traded have of the cache benefit for ensuring that
    > > Sid2name returns proper case. but smbd looks at names in case
    > > insenstive fashion and the NSS calls are normalized to lower
    > > case.
    > >
    > > > Please inform me as soon as possible if this patch should be removed
    > > > before we ship the final release.

    > >
    > > Unless there is a functional bug here, I would like to revert it.


    I noticed a few hours ago that this patch is not in v3-2-stable. Sorry for
    the confusion. So time does not matter now...

    > Well, we should at least return consistent upper/lower case
    > IMO. Not something that the user can influence.


    +1

    Btw: It was a customer who was complaining about inconsistent results.
    So it's not too academic... ;-)

    Karolin

    --
    SerNet GmbH, Bahnhofsallee 1b, 37081 G÷ttingen
    phone: +49-551-370000-0, fax: +49-551-370000-9
    AG G÷ttingen, HRB 2816, GF: Dr. Johannes Loxen
    http://www.SerNet.DE, mailto: Info @ SerNet.DE


  4. Re: Reviewing b58e4f6b3d7329....

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    Volker Lendecke wrote:
    > On Mon, Jun 30, 2008 at 11:30:33AM -0500, Gerald (Jerry) Carter wrote:
    >
    >> Was there a functional bug here? The report is academic IMO. So
    >> we've traded have of the cache benefit for ensuring that
    >> Sid2name returns proper case. but smbd looks at names in case
    >> insenstive fashion and the NSS calls are normalized to lower
    >> case.
    >>
    >>> Please inform me as soon as possible if this patch should be removed
    >>> before we ship the final release.

    >> Unless there is a functional bug here, I would like to revert it.

    >
    > Well, we should at least return consistent upper/lower case
    > IMO. Not something that the user can influence.


    ok. I give up. Whatever you want to do. I still think dropping
    the cache entry is the wrong thing to do but I'll do a patch later.
    Just help me understand why throwing away valid cache information
    was necessary.

    I think my confusion is that this is an internal caching
    mechanism to me. What was the customer complaining about
    and what in their environment specially *broke*. or were they
    just confused by the output from wbinfo?

    So why didn't we do a proper fix as you propose normalize the case?
    Seems I'm not seeing the entire picture here.




    cheers, jerry
    - --
    ================================================== ===================
    Samba ------- http://www.samba.org
    Likewise Software --------- http://www.likewisesoftware.com
    "What man is a man who does not make the world better?" --Balian
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

    iD8DBQFIaUETIR7qMdg1EfYRAsjkAJ0QJwULpZ3YWZuIVaR5VM fAZvd69wCgsxRj
    ferinf+ySb5QicRIjiIYWaA=
    =+nn9
    -----END PGP SIGNATURE-----


  5. Re: Reviewing b58e4f6b3d7329....

    On Mon, Jun 30, 2008 at 03:24:51PM -0500, Gerald (Jerry) Carter wrote:

    > So why didn't we do a proper fix as you propose normalize the case?
    > Seems I'm not seeing the entire picture here.


    We certainly could always lower-case names returned from
    wbinfo -s & friends. Normalizing to what the DC exactly
    returned would be another round-trip at wbinfo -n time.

    Volker

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.5 (GNU/Linux)

    iD8DBQFIaUKvUzqjrWwMRl0RAthMAKCTvOMZUlJAXd3hnr4N0s/3n4UGoQCfRYp0
    j+4L2lIsfYmjbP4FMdYWpzc=
    =12N4
    -----END PGP SIGNATURE-----


  6. Re: Reviewing b58e4f6b3d7329....

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    Volker Lendecke wrote:
    > On Mon, Jun 30, 2008 at 11:30:33AM -0500, Gerald (Jerry) Carter wrote:
    >
    >> Was there a functional bug here? The report is academic IMO. So
    >> we've traded have of the cache benefit for ensuring that
    >> Sid2name returns proper case. but smbd looks at names in case
    >> insenstive fashion and the NSS calls are normalized to lower
    >> case.
    >>
    >>> Please inform me as soon as possible if this patch should be removed
    >>> before we ship the final release.

    >> Unless there is a functional bug here, I would like to revert it.

    >
    > Well, we should at least return consistent upper/lower case
    > IMO. Not something that the user can influence.


    ok. I give up. Whatever you want to do. I still think dropping
    the cache entry is the wrong thing to do but I'll do a patch later.
    Just help me understand why throwing away valid cache information
    was necessary.

    I think my confusion is that this is an internal caching
    mechanism to me. What was the customer complaining about
    and what in their environment specially *broke*. or were they
    just confused by the output from wbinfo?

    So why didn't we do a proper fix as you propose normalize the case?
    Seems I'm not seeing the entire picture here.




    cheers, jerry
    - --
    ================================================== ===================
    Samba ------- http://www.samba.org
    Likewise Software --------- http://www.likewisesoftware.com
    "What man is a man who does not make the world better?" --Balian
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

    iD8DBQFIaUtuIR7qMdg1EfYRAmt+AKDALsUVEwRknxF1IIdN+e 3yoJfkbACfQu/j
    Ipt0Ih+Evf+pBLuYM9IoUXA=
    =Spi0
    -----END PGP SIGNATURE-----


  7. Re: Reviewing b58e4f6b3d7329....

    On Mon, Jun 30, 2008 at 04:09:02PM -0500, Gerald (Jerry) Carter wrote:
    > ok. I give up. Whatever you want to do. I still think dropping
    > the cache entry is the wrong thing to do but I'll do a patch later.
    > Just help me understand why throwing away valid cache information
    > was necessary.


    Reverted.

    Karolin

    --
    SerNet GmbH, Bahnhofsallee 1b, 37081 G├Âttingen
    phone: +49-551-370000-0, fax: +49-551-370000-9
    AG G├Âttingen, HRB 2816, GF: Dr. Johannes Loxen
    http://www.SerNet.DE, mailto: Info @ SerNet.DE


    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v2.0.4-svn0 (GNU/Linux)

    iD8DBQFIaex4KGi9fisXk1ERAipSAKCZWaThLw9Tc6bD+hFI3r DYCy0DsACgiD13
    Ppfh02JX2eKs14soK0Khdf4=
    =MCtb
    -----END PGP SIGNATURE-----


+ Reply to Thread