This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [PATCH 4/6] linux: Use 32-bit time_t for itimerval

From: Lukasz Majewski <lukma at denx dot de>
To: Andreas Schwab <schwab at suse dot de>, Alistair Francis <alistair dot francis at wdc dot com>, alistair23 at gmail dot com, Joseph Myers <joseph at codesourcery dot com>
Cc: libc-alpha at sourceware dot org, Adhemerval Zanella <adhemerval dot zanella at linaro dot org>, Carlos O'Donell <carlos at redhat dot com>, Florian Weimer <fweimer at redhat dot com>
Date: Tue, 4 Feb 2020 17:20:42 +0100
Subject: Re: [PATCH 4/6] linux: Use 32-bit time_t for itimerval
References: <20200203183153.11635-1-alistair.francis@wdc.com> <20200203183153.11635-5-alistair.francis@wdc.com> <20200204154908.39cfb08e@jawa> <mvmmu9ymlpy.fsf@suse.de>

Hi Andreas, Alistair,

> On Feb 04 2020, Lukasz Majewski wrote:
> 
> > I've dug to the @sysnames@ variable set in configure in the source
> > of glibc. I'm wondering how RV32 modifies it (and how it differs
> > from ARM32)?  
> 
> It is explicitly added through Implies (git grep -e wordsize-32 --
> '*/Implies*')

Thanks for pointing this out.

It seems like no riscv (rv32) sets this particular patch.

Only it is set for csky and nios2:
sysdeps/unix/sysv/linux/csky/Implies:unix/sysv/linux/generic/wordsize-32
sysdeps/unix/sysv/linux/nios2/Implies:unix/sysv/linux/generic/wordsize-32

However, it seems like I could:

1. Remove setitimer, getitimer, getrusage from
sysdeps/unix/syscalls.list

2. Extend Imply file for ARM32 (sysdeps/unix/sysv/linux/arm/Implies) by
adding
unix/sysv/linux/generic/wordsize-32

Unfortunately, this brings some issues with overflow header

../sysdeps/unix/sysv/linux/generic/wordsize-32/overflow.h:42:12: error:
'struct stat' has no member named '__st_ino_pad'; did you mean
'st_ino'? |    if (buf->__st_ino_pad == 0 && buf->__st_size_pad == 0 |
           ^~~~~~~~~~~~ |             st_ino
                                                                                                                                                                                                         

Maybe it would be better to not introduce setitimer, getitimer and
getrusage in:
sysdeps/unix/sysv/linux/generic/wordsize-32/

but instead in:
sysdeps/unix/sysv/linux/


So, it would be widely reusable as for example the __getitimer64 is
aliased anyway to getitimer for __TIMESIZE == 64 and __WORDSIZE==64 ?

(The same approach was taken with __timerfd_settime64 conversion:
https://sourceware.org/git/?p=glibc.git;a=commit;h=eae22432723b877354291aca4dbbfde5891dad59


> 
> Andreas.
> 




Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

Attachment: pgpXdJubc1Zpr.pgp
Description: OpenPGP digital signature

References:
- [PATCH 0/6] Always use 32-bit time_t for certain syscalls
  - From: Alistair Francis
- [PATCH 4/6] linux: Use 32-bit time_t for itimerval
  - From: Alistair Francis
- Re: [PATCH 4/6] linux: Use 32-bit time_t for itimerval
  - From: Lukasz Majewski
- Re: [PATCH 4/6] linux: Use 32-bit time_t for itimerval
  - From: Andreas Schwab

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]