Linux SNPS ARC Archive on lore.kernel.org
 help / color / Atom feed
From: Vineet Gupta <Vineet.Gupta1@synopsys.com>
To: Alistair Francis <alistair23@gmail.com>,
	Joseph Myers <joseph@codesourcery.com>
Cc: Florian Weimer <fweimer@redhat.com>,
	GNU C Library <libc-alpha@sourceware.org>,
	Arnd Bergmann <arnd@arndb.de>,
	Palmer Dabbelt <palmerdabbelt@google.com>,
	Zong Li <zongbox@gmail.com>,
	Alistair Francis <alistair.francis@wdc.com>,
	Adhemerval Zanella <adhemerval.zanella@linaro.org>,
	"Maciej W. Rozycki" <macro@wdc.com>,
	arcml <linux-snps-arc@lists.infradead.org>
Subject: switching ARC to 64-bit time_t (Re: [RFC v6 07/23] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64)
Date: Tue, 18 Feb 2020 23:05:28 +0000
Message-ID: <8a9784b3-fc52-adc3-4595-33142b059388@synopsys.com> (raw)
In-Reply-To: <CAKmqyKOgFRfMOws_-48GqMnuS3ygmN9A4DzLg6UtEjRXVoM82A@mail.gmail.com>

Hi Alistair, Arnd

On 2/14/20 2:39 PM, Alistair Francis wrote:
> On Tue, Feb 11, 2020 at 5:30 PM Joseph Myers <joseph@codesourcery.com> wrote:
>> On Tue, 11 Feb 2020, Alistair Francis wrote:
>>
>>>>> diff --git a/sysdeps/unix/sysv/linux/riscv/bits/typesizes.h b/sysdeps/unix/sysv/linux/riscv/bits/typesizes.h
>>>>> new file mode 100644
>>>>> index 0000000000..0da3bdeb5d
>>>>> --- /dev/null
>>>>> +++ b/sysdeps/unix/sysv/linux/riscv/bits/typesizes.h
>>>> I was hoping newer arches could simply use the asm-generic one ?
>>> We need to specify that RV32 uses a 64-bit time_t. The generic ones
>>> don't do that for 32-bit arches.
>> Since it seems we'd like future 32-bit ports of glibc to use 64-bit time
>> and offsets, we should make that as easy as possible.
>>
>> That is, you need an RISC-V-specific bits/timesize.h.  But you shouldn't
>> need an RISC-V-specific bits/typesizes.h - rather, make the linux/generic
>> one do the right thing for __TIME_T_TYPE based on bits/timesize.h.  And
>> have some other header that 32-bit linux/generic ports can use to say
>> whether they use the 64-bit offset/stat/statfs interface, that
>> bits/typesizes.h can use together with its existing __LP64__ check, and
>> make the definitions of __OFF_T_TYPE etc. check that as well, and then you
>> shouldn't need an RISC-V-specific bits/typesizes.h - the RISC-V-specific
>> headers should be strictly minimal.  (No architecture-specific
>> bits/time64.h headers should be needed in any case.)
> Ok, I have updated this. I'll send the patch once my "Always use
> 32-bit time_t for certain syscalls" series is in (the headers are
> changed in that series).

I guess you haven't pushed changes yet, which don't make full copy of typesizes.h ?

Anyhow I have a version based on your prior next branch where I switched ARC to
64-bit time_t - things work fine in general but I see some additional failures
with the testsuite.

Consider io/test-stat2.c which calls stat() and stat64() and compares the results:
it now fails for ctime mismatch

| ...
| st_atime: [72] 644245094405576070 vs [72] 644245094405576070  OK
| st_mtime: [88] 1975684956160000000 vs [88] 1975684956160000000  OK
| st_ctime: [104] 0 vs [104] 2306351876938924035  FAIL


In kernel asm-generic stat64 has 32-bit ctime (secs)

    struct stat64 {
    ...
        int                   st_atime;            <-- offset 72
        unsigned int    st_atime_nsec;
        int                   st_mtime;            <-- offset *80*
        unsigned int    st_mtime_nsec;
        int                   st_ctime;            <-- offset 92
        unsigned int    st_ctime_nsec;
    ...
    };

In glibc, we have 64-bit time_t based timestamps so the structure diverges with
kernel counterpart from time fields onwards.

    __extension__ typedef __int64_t __time_t;        <-- ARC switched to 64-bit time_t

    struct timespec
    {
      __time_t tv_sec;    <-- 8
      long int tv_nsec;    <-- 4
      int: 32;                   <-- 4
    };

    struct stat64
      {
    ...
        struct timespec st_atim;    <-- offset 72
        struct timespec st_mtim;    <-- offset *88*
        struct timespec st_ctim;    <-- offset 104
        int __glibc_reserved[2];
      };

However glibc stat64()wrapper calls
sysdeps/unix/sysv/linux/generic/wordsize-32/xstat64.c

which for ARC is just doing a pass thru syscall because we do have __NR_fstatat64
- hence the issues I see.

It needs itemized copy

    __xstat64 (int vers, const char *name, struct stat64 *buf)
    {
    #ifdef __NR_fstatat64
          return INLINE_SYSCALL (fstatat64, 4, AT_FDCWD, name, buf, 0);
    #else
    ....
          int rc = INLINE_SYSCALL (statx, 5, AT_FDCWD, name, AT_NO_AUTOMOUNT,
                               STATX_BASIC_STATS, &tmp);
        __cp_stat64_statx (buf, &tmp);
    #endif

An the reason this all works on RISCV is that your kernel doesn't define
__ARCH_WANT_STAT64 -> lacks __NR_statat64 and instead uses the statx call which
does itemized copy and would work fine when copying from 32-bits time (in kernel)
to 64-bits container in glibc. Is this is right understanding or am I missing
something here.

How do I build a latest RISCV 32-bit kernel + userland - do you have a buildroot
branch somewhere that I can build / test with qemu ?

Thx,
-Vineet



_______________________________________________
linux-snps-arc mailing list
linux-snps-arc@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-snps-arc

  reply index

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <cover.1578824547.git.alistair.francis@wdc.com>
     [not found] ` <4e95f95966d8d7c6a8339160dc62d81c1f6a1bfb.1578824547.git.alistair.francis@wdc.com>
2020-02-12  0:14   ` [RFC v6 07/23] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Vineet Gupta
2020-02-12  0:14     ` Alistair Francis
2020-02-12  1:30       ` Joseph Myers
2020-02-14 22:39         ` Alistair Francis
2020-02-18 23:05           ` Vineet Gupta [this message]
2020-02-18 23:13             ` switching ARC to 64-bit time_t (Re: [RFC v6 07/23] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64) Joseph Myers
2020-02-19 23:09               ` Lukasz Majewski
2020-02-19 23:11             ` Lukasz Majewski
2020-02-20  8:31               ` Arnd Bergmann
2020-02-20  9:37                 ` Lukasz Majewski
2020-02-20 12:37                   ` Arnd Bergmann
2020-02-20 13:14                     ` Lukasz Majewski
2020-02-20 14:44                       ` Arnd Bergmann
2020-02-20 15:42                         ` Lukasz Majewski
2020-02-20 16:08                           ` Arnd Bergmann
2020-02-20 16:31                             ` Lukasz Majewski
2020-02-24  2:48                         ` Viresh Kumar
2020-02-21 19:56                     ` Alistair Francis
2020-02-22  8:42                   ` Arnd Bergmann
2020-02-24  9:00                     ` Lukasz Majewski
2020-02-24  9:46                       ` Andreas Schwab
2020-02-24 10:14                         ` Lukasz Majewski
2020-02-24 10:23                           ` Andreas Schwab
2020-02-24 10:36                             ` Lukasz Majewski
2020-02-24 10:42                               ` Andreas Schwab
2020-02-24 11:13                                 ` Lukasz Majewski
2020-02-24 12:41                                   ` Lukasz Majewski
2020-02-25  0:03                               ` Joseph Myers
2020-02-25 11:39                                 ` Lukasz Majewski
2020-02-25 14:36                                   ` Joseph Myers
2020-02-26 13:18                                     ` Lukasz Majewski
2020-02-26 14:48                                       ` Joseph Myers
2020-02-26 16:28                                         ` Lukasz Majewski
2020-02-25  9:03                       ` Arnd Bergmann
2020-02-20 16:27                 ` Helmut Grohne
2020-03-26  0:25                   ` ARC rebootstrap prereq (was Re: switching ARC to 64-bit time_t ) Vineet Gupta
2020-03-26  5:54                     ` Helmut Grohne
2020-03-26 11:51                       ` Alexey Brodkin
2020-03-26 12:24                         ` Helmut Grohne
2020-03-26 12:53                           ` Alexey Brodkin
2020-03-26 14:28                             ` Helmut Grohne
2020-03-26 19:04                               ` Lennart Sorensen
2020-02-12  1:42       ` [RFC v6 07/23] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Vineet Gupta
2020-02-12 12:58         ` Arnd Bergmann
2020-02-19  0:56   ` Vineet Gupta
2020-02-19  1:03     ` Alistair Francis
2020-02-19  1:31       ` Vineet Gupta
2020-02-19  8:30         ` Andreas Schwab
2020-02-19 18:42           ` Vineet Gupta
2020-02-19 23:18             ` Lukasz Majewski
2020-02-20  0:26               ` Vineet Gupta
2020-02-20  0:46                 ` Joseph Myers
2020-02-20  8:24                   ` Arnd Bergmann
2020-02-20 10:28                     ` Lukasz Majewski
2020-02-20 14:14                     ` Joseph Myers

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=8a9784b3-fc52-adc3-4595-33142b059388@synopsys.com \
    --to=vineet.gupta1@synopsys.com \
    --cc=adhemerval.zanella@linaro.org \
    --cc=alistair.francis@wdc.com \
    --cc=alistair23@gmail.com \
    --cc=arnd@arndb.de \
    --cc=fweimer@redhat.com \
    --cc=joseph@codesourcery.com \
    --cc=libc-alpha@sourceware.org \
    --cc=linux-snps-arc@lists.infradead.org \
    --cc=macro@wdc.com \
    --cc=palmerdabbelt@google.com \
    --cc=zongbox@gmail.com \
    /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

Linux SNPS ARC Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-snps-arc/0 linux-snps-arc/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-snps-arc linux-snps-arc/ https://lore.kernel.org/linux-snps-arc \
		linux-snps-arc@lists.infradead.org
	public-inbox-index linux-snps-arc

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.infradead.lists.linux-snps-arc


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git