linux-parisc.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: John David Anglin <dave.anglin@bell.net>
To: Helge Deller <deller@gmx.de>, linux-parisc@vger.kernel.org
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>
Subject: Re: [PATCH 4/4] parisc: Reduce sigreturn trampoline to 3 instructions
Date: Tue, 16 Nov 2021 10:31:29 -0500	[thread overview]
Message-ID: <e6f05692-d275-589e-46aa-32e29765f373@bell.net> (raw)
In-Reply-To: <b2abc63f-a898-c925-1fc7-af5cf4930f6d@gmx.de>

I found the change with a regression search.

I don't know what's wrong with it.  I have reverted it for now.

Here are test results with 5.16.0-rc1+ #1 SMP Mon Nov 15 17:56:32 EST 2021 parisc64:
XPASS: conform/UNIX98/ndbm.h/linknamespace
XPASS: conform/XOPEN2K/ndbm.h/linknamespace
XPASS: conform/XOPEN2K8/ndbm.h/linknamespace
XPASS: conform/XPG42/ndbm.h/linknamespace
FAIL: dirent/tst-readdir64-compat
FAIL: elf/tst-audit2
UNSUPPORTED: io/tst-fallocate
UNSUPPORTED: io/tst-fallocate64
UNSUPPORTED: misc/tst-adjtimex
UNSUPPORTED: misc/tst-adjtimex-time64
UNSUPPORTED: misc/tst-clock_adjtime
UNSUPPORTED: misc/tst-clock_adjtime-time64
UNSUPPORTED: misc/tst-ntp_adjtime
UNSUPPORTED: misc/tst-ntp_adjtime-time64
UNSUPPORTED: misc/tst-pkey
UNSUPPORTED: nptl/test-cond-printers
UNSUPPORTED: nptl/test-condattr-printers
UNSUPPORTED: nptl/test-mutex-printers
UNSUPPORTED: nptl/test-mutexattr-printers
UNSUPPORTED: nptl/test-rwlock-printers
UNSUPPORTED: nptl/test-rwlockattr-printers
FAIL: nptl/tst-cleanupx4
FAIL: nptl/tst-cond25
FAIL: signal/tst-minsigstksz-5
UNSUPPORTED: stdlib/test-bz22786
FAIL: stdlib/tst-setcontext2
FAIL: stdlib/tst-setcontext7
UNSUPPORTED: stdlib/tst-strtod-overflow
UNSUPPORTED: string/tst-memmove-overflow
UNSUPPORTED: time/tst-clock_settime
UNSUPPORTED: time/tst-clock_settime-time64
UNSUPPORTED: time/tst-settimeofday
UNSUPPORTED: time/tst-settimeofday-time64
UNSUPPORTED: time/tst-y2039
Summary of test results:
       7 FAIL
    4173 PASS
      23 UNSUPPORTED
      16 XFAIL
       4 XPASS

Probably, Iecho $? should investigate why nptl/tst-cleanupx4 fails.  Might be another issue with signal delivery.

Dave
On 2021-11-16 8:54 a.m., Helge Deller wrote:
> Dave,
>
> Thanks for tracking this down!
>
> I wonder what's wrong with it.
> Based on the outcome I should either completely revert that patch or fix it.
>
> Helge
>
> On 11/15/21 23:25, John David Anglin wrote:
>> This change breaks signal delivery and causes various glibc tests to fail.
>>
>> commit 3e4a1aff2a97cb4fd7f0268e4b69e8c9d3641277:
>> dave@atlas:~/gnu/glibc/objdir$ env GCONV_PATH=/home/dave/gnu/glibc/objdir/iconvdata LOCPATH=/home/dave/gnu/glibc/objdir/localedata LC_ALL=C /home/dave/gnu/glibc/objdir/elf/ld.so.1 --library-path /home/dave/gnu/glibc/objdir:/home/dave/gnu/glibc/objdir/math:/home/dave/gnu/glibc/objdir/elf:/home/dave/gnu/glibc/objdir/dlfcn:/home/dave/gnu/glibc/objdir/nss:/home/dave/gnu/glibc/objdir/nis:/home/dave/gnu/glibc/objdir/rt:/home/dave/gnu/glibc/objdir/resolv:/home/dave/gnu/glibc/objdir/mathvec:/home/dave/gnu/glibc/objdir/support:/home/dave/gnu/glibc/objdir/crypt:/home/dave/gnu/glibc/objdir/nptl /home/dave/gnu/glibc/objdir/nptl/tst-cancelx4
>> got size 4512
>> in-time cancel test of 'read' successful
>> in-time cancel test of 'readv' successful
>> in-time cancel test of 'select' successful
>> in-time cancel test of 'pselect' successful
>> in-time cancel test of 'poll' successful
>> in-time cancel test of 'ppoll' successful
>> in-time cancel test of 'write' successful
>> in-time cancel test of 'writev' successful
>> in-time cancel test of 'sleep' successful
>> in-time cancel test of 'usleep' successful
>> in-time cancel test of 'nanosleep' successful
>> in-time cancel test of 'wait' successful
>> in-time cancel test of 'waitid' successful
>> in-time cancel test of 'waitpid' successful
>> in-time cancel test of 'sigpause' successful
>> in-time cancel test of 'sigsuspend' successful
>> in-time cancel test of 'sigwait' successful
>> in-time cancel test of 'sigwaitinfo' successful
>> in-time cancel test of 'sigtimedwait' successful
>> in-time cancel test of 'pause' successful
>> in-time cancel test of 'accept' successful
>> got size 4512
>> in-time cancel test of 'send' successful
>> in-time cancel test of 'recv' successful
>> in-time cancel test of 'recvfrom' successful
>> in-time cancel test of 'recvmsg' successful
>> in-time cancel test of 'msgrcv' successful
>> early cancel test of 'read' successful
>> early cancel test of 'readv' successful
>> early cancel test of 'select' successful
>> early cancel test of 'pselect' successful
>> early cancel test of 'poll' successful
>> early cancel test of 'ppoll' successful
>> early cancel test of 'write' successful
>> early cancel test of 'writev' successful
>> early cancel test of 'sleep' successful
>> early cancel test of 'usleep' successful
>> early cancel test of 'nanosleep' successful
>> early cancel test of 'wait' successful
>> early cancel test of 'waitid' successful
>> early cancel test of 'waitpid' successful
>> early cancel test of 'sigpause' successful
>> early cancel test of 'sigsuspend' successful
>> early cancel test of 'sigwait' successful
>> early cancel test of 'sigwaitinfo' successful
>> early cancel test of 'sigtimedwait' successful
>> early cancel test of 'pause' successful
>> early cancel test of 'accept' successful
>> got size 4512
>> early cancel test of 'send' successful
>> early cancel test of 'recv' successful
>> early cancel test of 'recvfrom' successful
>> early cancel test of 'recvmsg' successful
>> early cancel test of 'preadv' successful
>> early cancel test of 'preadv2' successful
>> early cancel test of 'pwritev' successful
>> early cancel test of 'pwritev2' successful
>> early cancel test of 'open' successful
>> early cancel test of 'close' successful
>> early cancel test of 'pread' successful
>> early cancel test of 'pwrite' successful
>> early cancel test of 'fsync' successful
>> early cancel test of 'fdatasync' successful
>> early cancel test of 'msync' successful
>> got size 4512
>> early cancel test of 'sendto' successful
>> early cancel test of 'sendmsg' successful
>> early cancel test of 'creat' successful
>> early cancel test of 'connect' successful
>> early cancel test of 'tcdrain' successful
>> early cancel test of 'msgrcv' successful
>> early cancel test of 'msgsnd' successful
>> dave@atlas:~/gnu/glibc/objdir$ echo $?
>> 0
>>
>> commit e4f2006f1287e7ea17660490569cff323772dac4:
>> dave@atlas:~/gnu/glibc/objdir$ env GCONV_PATH=/home/dave/gnu/glibc/objdir/iconvdata LOCPATH=/home/dave/gnu/glibc/objdir/localedata LC_ALL=C /home/dave/gnu/glibc/objdir/elf/ld.so.1 --library-path /home/dave/gnu/glibc/objdir:/home/dave/gnu/glibc/objdir/math:/home/dave/gnu/glibc/objdir/elf:/home/dave/gnu/glibc/objdir/dlfcn:/home/dave/gnu/glibc/objdir/nss:/home/dave/gnu/glibc/objdir/nis:/home/dave/gnu/glibc/objdir/rt:/home/dave/gnu/glibc/objdir/resolv:/home/dave/gnu/glibc/objdir/mathvec:/home/dave/gnu/glibc/objdir/support:/home/dave/gnu/glibc/objdir/crypt:/home/dave/gnu/glibc/objdir/nptl /home/dave/gnu/glibc/objdir/nptl/tst-cancelx4
>> got size 4512
>> cleanup handler not called for 'read'
>> cleanup handler not called for 'readv'
>> cleanup handler not called for 'select'
>> cleanup handler not called for 'pselect'
>> cleanup handler not called for 'poll'
>> cleanup handler not called for 'ppoll'
>> cleanup handler not called for 'write'
>> cleanup handler not called for 'writev'
>> cleanup handler not called for 'sleep'
>> cleanup handler not called for 'usleep'
>> cleanup handler not called for 'nanosleep'
>> cleanup handler not called for 'wait'
>> cleanup handler not called for 'waitid'
>> cleanup handler not called for 'waitpid'
>> cleanup handler not called for 'sigpause'
>> cleanup handler not called for 'sigsuspend'
>> cleanup handler not called for 'sigwait'
>> cleanup handler not called for 'sigwaitinfo'
>> cleanup handler not called for 'sigtimedwait'
>> cleanup handler not called for 'pause'
>> cleanup handler not called for 'accept'
>> got size 4512
>> cleanup handler not called for 'send'
>> cleanup handler not called for 'recv'
>> cleanup handler not called for 'recvfrom'
>> cleanup handler not called for 'recvmsg'
>> cleanup handler not called for 'msgrcv'
>> early cancel test of 'read' successful
>> early cancel test of 'readv' successful
>> early cancel test of 'select' successful
>> early cancel test of 'pselect' successful
>> early cancel test of 'poll' successful
>> early cancel test of 'ppoll' successful
>> early cancel test of 'write' successful
>> early cancel test of 'writev' successful
>> early cancel test of 'sleep' successful
>> early cancel test of 'usleep' successful
>> early cancel test of 'nanosleep' successful
>> early cancel test of 'wait' successful
>> early cancel test of 'waitid' successful
>> early cancel test of 'waitpid' successful
>> early cancel test of 'sigpause' successful
>> early cancel test of 'sigsuspend' successful
>> early cancel test of 'sigwait' successful
>> early cancel test of 'sigwaitinfo' successful
>> early cancel test of 'sigtimedwait' successful
>> early cancel test of 'pause' successful
>> early cancel test of 'accept' successful
>> got size 4512
>> early cancel test of 'send' successful
>> early cancel test of 'recv' successful
>> early cancel test of 'recvfrom' successful
>> early cancel test of 'recvmsg' successful
>> early cancel test of 'preadv' successful
>> early cancel test of 'preadv2' successful
>> early cancel test of 'pwritev' successful
>> early cancel test of 'pwritev2' successful
>> early cancel test of 'open' successful
>> early cancel test of 'close' successful
>> early cancel test of 'pread' successful
>> early cancel test of 'pwrite' successful
>> early cancel test of 'fsync' successful
>> early cancel test of 'fdatasync' successful
>> early cancel test of 'msync' successful
>> got size 4512
>> early cancel test of 'sendto' successful
>> early cancel test of 'sendmsg' successful
>> early cancel test of 'creat' successful
>> early cancel test of 'connect' successful
>> early cancel test of 'tcdrain' successful
>> early cancel test of 'msgrcv' successful
>> early cancel test of 'msgsnd' successful
>> dave@atlas:~/gnu/glibc/objdir$ echo $?
>> 1
>>
>> Dave
>>
>> On 2021-09-08 4:44 p.m., Helge Deller wrote:
>>> We can move the INSN_LDI_R20 instruction into the branch delay slot.
>>>
>>> Signed-off-by: Helge Deller <deller@gmx.de>
>>> ---
>>>    arch/parisc/include/asm/rt_sigframe.h |  2 +-
>>>    arch/parisc/kernel/signal.c           | 13 ++++++-------
>>>    arch/parisc/kernel/signal32.h         |  2 +-
>>>    3 files changed, 8 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/arch/parisc/include/asm/rt_sigframe.h b/arch/parisc/include/asm/rt_sigframe.h
>>> index 2b3010ade00e..4b9e3d707571 100644
>>> --- a/arch/parisc/include/asm/rt_sigframe.h
>>> +++ b/arch/parisc/include/asm/rt_sigframe.h
>>> @@ -2,7 +2,7 @@
>>>    #ifndef _ASM_PARISC_RT_SIGFRAME_H
>>>    #define _ASM_PARISC_RT_SIGFRAME_H
>>>
>>> -#define SIGRETURN_TRAMP 4
>>> +#define SIGRETURN_TRAMP 3
>>>    #define SIGRESTARTBLOCK_TRAMP 5
>>>    #define TRAMP_SIZE (SIGRETURN_TRAMP + SIGRESTARTBLOCK_TRAMP)
>>>
>>> diff --git a/arch/parisc/kernel/signal.c b/arch/parisc/kernel/signal.c
>>> index 46b1050640b8..bbfe23c40c01 100644
>>> --- a/arch/parisc/kernel/signal.c
>>> +++ b/arch/parisc/kernel/signal.c
>>> @@ -288,22 +288,21 @@ setup_rt_frame(struct ksignal *ksig, sigset_t *set, struct pt_regs *regs,
>>>           already in userspace. The first words of tramp are used to
>>>           save the previous sigrestartblock trampoline that might be
>>>           on the stack. We start the sigreturn trampoline at
>>> -       SIGRESTARTBLOCK_TRAMP+X. */
>>> +       SIGRESTARTBLOCK_TRAMP. */
>>>        err |= __put_user(in_syscall ? INSN_LDI_R25_1 : INSN_LDI_R25_0,
>>>                &frame->tramp[SIGRESTARTBLOCK_TRAMP+0]);
>>> -    err |= __put_user(INSN_LDI_R20,
>>> -            &frame->tramp[SIGRESTARTBLOCK_TRAMP+1]);
>>>        err |= __put_user(INSN_BLE_SR2_R0,
>>> +            &frame->tramp[SIGRESTARTBLOCK_TRAMP+1]);
>>> +    err |= __put_user(INSN_LDI_R20,
>>>                &frame->tramp[SIGRESTARTBLOCK_TRAMP+2]);
>>> -    err |= __put_user(INSN_NOP, &frame->tramp[SIGRESTARTBLOCK_TRAMP+3]);
>>>
>>> -    start = (unsigned long) &frame->tramp[0];
>>> -    end = (unsigned long) &frame->tramp[TRAMP_SIZE];
>>> +    start = (unsigned long) &frame->tramp[SIGRESTARTBLOCK_TRAMP+0];
>>> +    end = (unsigned long) &frame->tramp[SIGRESTARTBLOCK_TRAMP+3];
>>>        flush_user_dcache_range_asm(start, end);
>>>        flush_user_icache_range_asm(start, end);
>>>
>>>        /* TRAMP Words 0-4, Length 5 = SIGRESTARTBLOCK_TRAMP
>>> -     * TRAMP Words 5-9, Length 4 = SIGRETURN_TRAMP
>>> +     * TRAMP Words 5-7, Length 3 = SIGRETURN_TRAMP
>>>         * So the SIGRETURN_TRAMP is at the end of SIGRESTARTBLOCK_TRAMP
>>>         */
>>>        rp = (unsigned long) &frame->tramp[SIGRESTARTBLOCK_TRAMP];
>>> diff --git a/arch/parisc/kernel/signal32.h b/arch/parisc/kernel/signal32.h
>>> index f166250f2d06..a5bdbb5678b7 100644
>>> --- a/arch/parisc/kernel/signal32.h
>>> +++ b/arch/parisc/kernel/signal32.h
>>> @@ -36,7 +36,7 @@ struct compat_regfile {
>>>            compat_int_t rf_sar;
>>>    };
>>>
>>> -#define COMPAT_SIGRETURN_TRAMP 4
>>> +#define COMPAT_SIGRETURN_TRAMP 3
>>>    #define COMPAT_SIGRESTARTBLOCK_TRAMP 5
>>>    #define COMPAT_TRAMP_SIZE (COMPAT_SIGRETURN_TRAMP + \
>>>                    COMPAT_SIGRESTARTBLOCK_TRAMP)
>>> --
>>> 2.31.1
>>>
>>


-- 
John David Anglin  dave.anglin@bell.net


  reply	other threads:[~2021-11-16 15:31 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-08 20:44 [PATCH 1/4] parisc: Drop strnlen_user() in favour of generic version Helge Deller
2021-09-08 20:44 ` [PATCH 2/4] parisc: Drop useless debug info and comments from signal.c Helge Deller
2021-09-08 20:44 ` [PATCH 3/4] parisc: Check user signal stack trampoline is inside TASK_SIZE Helge Deller
2021-09-08 20:44 ` [PATCH 4/4] parisc: Reduce sigreturn trampoline to 3 instructions Helge Deller
2021-11-15 22:25   ` John David Anglin
2021-11-16 13:54     ` Helge Deller
2021-11-16 15:31       ` John David Anglin [this message]
2021-11-16 17:08         ` John David Anglin
2021-11-16 19:02           ` Helge Deller
2021-11-16 19:12             ` John David Anglin
2021-09-08 21:26 ` [PATCH 1/4] parisc: Drop strnlen_user() in favour of generic version Arnd Bergmann
2021-09-08 21:40   ` Helge Deller
2021-09-08 22:08     ` Arnd Bergmann
2021-09-09  2:03       ` Helge Deller
2021-09-09  6:51         ` Arnd Bergmann

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=e6f05692-d275-589e-46aa-32e29765f373@bell.net \
    --to=dave.anglin@bell.net \
    --cc=James.Bottomley@HansenPartnership.com \
    --cc=deller@gmx.de \
    --cc=linux-parisc@vger.kernel.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 NNTP newsgroup(s).