public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: Joseph Myers <joseph@codesourcery.com>
Cc: libc-alpha@sourceware.org
Subject: Re: [PATCH v2 4/7] login: Add 64-bit time support to utmp/utmpx
Date: Wed, 5 Aug 2020 17:35:25 -0300	[thread overview]
Message-ID: <d7f3866a-a6d2-a0c1-3f29-cac2a0f33d16@linaro.org> (raw)
In-Reply-To: <alpine.DEB.2.21.2008051922120.1829@digraph.polyomino.org.uk>



On 05/08/2020 16:28, Joseph Myers wrote:
> On Wed, 5 Aug 2020, Adhemerval Zanella via Libc-alpha wrote:
> 
>>   - New file names are defined for _PATH_UTMP, _PATH_WTMP, and
>>     _PATH_UTMP_DB (if applicable) and used as default for the new 2.33
>>     utmp{x} symbols.
> 
> On architectures that already have 64-bit time in these files, I think it 
> would be best to keep the same names as at present (and likewise for 
> lastlog).  (I'm not particularly concerned with whether such architectures 
> get (unnecessary) new symbol versions, though ideally we'd avoid code size 
> increase by ensuring there is no translation logic.)  If any such 
> architectures are using a different layout from the new generic one, 
> despite using 64-bit times, maybe they could keep their own version of the 
> layout to avoid needing multiple versions of the files?
> 
> Additions to the NEWS file are also needed to discuss these changes and 
> what distributions should do on glibc upgrade.
> 

I thought keeping the same filename for architectures that already
have 64-bit ut_tv member.  However, currently only s390 fits in this
category and it would make it an outlier.  It means that glibc on s390
would be the only ABI with a different _PATH_UTMP than the default one,
which I see only adds extra complexity in the transition (distro would
need to handle s390 as an exception, this will need to be properly
documented not only in glibc, it will require an arch-specific s390
path.h). 

I am not really against keep the same _PATH_UTMP and _PATH_WTMP, but
I think keeping arch-neutral tends to cause less confusion (which I 
foresee it would happen with this change anyway).

Regarding the new symbols versions, the patch already handles it 
for s390: neither on s390 nor on s390x new symbols are created.  And
utmp{x} layout is also preserved (the ABI has arch-specific 
struct_utmp.h and struct_utmpx.h). The only change is indeed new
_PATH_UTMP and _PATH_WTMP names.

I will work on a NEWS entry.


  reply	other threads:[~2020-08-05 20:35 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-05 18:59 [PATCH v2 1/7] login: Move gnu utmpx to default implementation Adhemerval Zanella
2020-08-05 18:59 ` [PATCH v2 2/7] Consolidate and simplify internal utmp definitions Adhemerval Zanella
2020-08-05 18:59 ` [PATCH v2 3/7] support: Add 'touch' command Adhemerval Zanella
2020-08-05 19:08   ` DJ Delorie
2020-08-05 18:59 ` [PATCH v2 4/7] login: Add 64-bit time support to utmp/utmpx Adhemerval Zanella
2020-08-05 19:28   ` Joseph Myers
2020-08-05 20:35     ` Adhemerval Zanella [this message]
2020-08-05 20:43       ` Joseph Myers
2020-08-05 20:54         ` Adhemerval Zanella
2020-08-05 18:59 ` [PATCH v2 5/7] linux: Add pwrite64_nocancel Adhemerval Zanella
2020-08-05 18:59 ` [PATCH v2 6/7] login: Use 64-bit time on struct lastlog Adhemerval Zanella
2020-08-05 19:56   ` Florian Weimer
2020-08-05 20:36     ` Adhemerval Zanella
2020-08-05 18:59 ` [PATCH v2 7/7] Remove __WORDSIZE_TIME64_COMPAT32 Adhemerval Zanella

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=d7f3866a-a6d2-a0c1-3f29-cac2a0f33d16@linaro.org \
    --to=adhemerval.zanella@linaro.org \
    --cc=joseph@codesourcery.com \
    --cc=libc-alpha@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).