linux-kernel - Re: [PATCH] sh: fix syscall tracing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 10 Sep 2020 09:37:51 -0400
From:   Rich Felker <dalias@...c.org>
To:     Rob Landley <rob@...dley.net>
Cc:     John Paul Adrian Glaubitz <glaubitz@...sik.fu-berlin.de>,
        linux-sh@...r.kernel.org,
        Michael Karcher <kernel@...rcher.dialup.fu-berlin.de>,
        linux-kernel@...r.kernel.org,
        Yoshinori Sato <ysato@...rs.sourceforge.jp>
Subject: Re: [PATCH] sh: fix syscall tracing

On Thu, Sep 10, 2020 at 06:02:05AM -0500, Rob Landley wrote:
> On 9/10/20 4:55 AM, John Paul Adrian Glaubitz wrote:
> > Hi Rich!
> > 
> > On 9/7/20 7:44 PM, Rich Felker wrote:
> >>> Can we still get this merged as a hotfix for 5.9?
> >>
> >> Yes, fixes for regressions in the same release cycle are in-scope (the
> >> whole point of having -rc's). I have at least one other fix that needs
> >> to go in too and was just giving it a little time to make sure
> >> everything's ok now and that there are no more.
> > 
> > Let me know if there is anything else left for testing.
> 
> Could you also merge the fix the build break, ala:
> 
> > The vmlinux image is a current vanilla Linux kernel using an initramfs filesystem:
> > 
> >   make ARCH=sh CROSS_COMPILE=sh2eb-linux-muslfdpic- j2_defconfig vmlinux
> > 
> > And trying to do that in current git dies with:
> > 
> >   CC      init/version.o
> > In file included from ./include/linux/spinlock.h:318,
> >                  from ./arch/sh/include/asm/smp.h:11,
> >                  from ./include/linux/smp.h:82,
> >                  from ./include/linux/lockdep.h:14,
> >                  from ./include/linux/rcupdate.h:29,
> >                  from ./include/linux/rculist.h:11,
> >                  from ./include/linux/pid.h:5,
> >                  from ./include/linux/sched.h:14,
> >                  from ./include/linux/utsname.h:6,
> >                  from init/version.c:14:
> > ./include/linux/spinlock_api_smp.h: In function '__raw_spin_trylock':
> > ./include/linux/spinlock_api_smp.h:90:3: error: implicit declaration of function
> > 'spin_acquire'; did you mean 'xchg_acquire'? [-Werror=implicit-function-declaration]
> >    90 |   spin_acquire(&lock->dep_map, 0, 1, _RET_IP_);
> >       |   ^~~~~~~~~~~~
> >       |   xchg_acquire
> > ./include/linux/spinlock_api_smp.h:90:21: error: 'raw_spinlock_t' {aka 'struct
> > raw_spinlock'} has no member named 'dep_map'
> >    90 |   spin_acquire(&lock->dep_map, 0, 1, _RET_IP_);
> >       |                     ^~
> > 
> > And so on and so forth for pages. I bisected it to:
> > 
> > commit 0cd39f4600ed4de859383018eb10f0f724900e1b
> > Author: Peter Zijlstra <peterz@...radead.org>
> > Date:   Thu Aug 6 14:35:11 2020 +0200
> > 
> >     locking/seqlock, headers: Untangle the spaghetti monster
> 
> Which I reported to Rich on the 2nd and he had me test a one line patch fixing
> it (adding an extra #include) on the 3rd, but I just did a fresh pull and the
> j2_defconfig build still broke a week later.

Yes, that's presently the other regression fix I have queued for the
second pull request.

Rich