[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: RPI: * failed to add service - already in use?
Similar issue with SDL2, this time on the latest 8.0_BETA:
Reading symbols from emulationstation...(no debugging symbols found)...done.
Starting program: /usr/pkg/bin/emulationstation
Program received signal SIGILL, Illegal instruction.
0x72bb7ae0 in _armv7_neon_probe () from /usr/lib/libcrypto.so.12
That function only gets called here in armv4cpuid.S
(I have applied your diff)
mrrc p15,1,r0,r1,c14 @ CNTVCT
__ARM_MAX_ARCH should be 6, not 7, since we are on armv6 and do not
have neon anyway.
I'll need to pick this up tomorrow now, but any ideas are welcome.
On 3 March 2018 at 10:02, Michael Blatt <michael.blatt%gmail.com@localhost> wrote:
> That looks like the right diff - I can't reproduce the issue until I
> get another SD card setup, as I've re-built the project on 8.0_BETA.
> On 3 March 2018 at 02:41, Jun Ebihara <jun%soum.co.jp@localhost> wrote:
>> need this ifdef?
>> diff -u armv4cpuid.S.1 armv4cpuid.S
>> --- armv4cpuid.S.1 2018-03-03 11:19:58.377064302 +0900
>> +++ armv4cpuid.S 2018-03-03 11:21:00.926429891 +0900
>> @@ -95,7 +95,9 @@
>> mrrc p15,1,r0,r1,c14 @ CNTVCT
>> .size _armv7_tick,.-_armv7_tick
>> +#if __ARM_MAX_ARCH__>=8
>> .global _armv8_aes_probe
>> .type _armv8_aes_probe,%function
>> From: Jun Ebihara <jun%soum.co.jp@localhost>
>> Subject: Re: RPI: * failed to add service - already in use?
>> Date: Thu, 01 Mar 2018 10:45:18 +0900 (JST)
>>> Thanx.I pickup and put 2017-12-16/2018-02-18 libcrypt.so.
>>> From: Michael Blatt <michael.blatt%gmail.com@localhost>
>>> Subject: Re: RPI: * failed to add service - already in use?
>>> Date: Wed, 28 Feb 2018 19:39:50 +0000
>>>> On that note: I've noticed the built-in unbound segfaults very quickly
>>>> upon doing lookups on earmv7hf, same codebase (8.99.12)
>>>> An analysis of the coredump suggests it's also on libcrypt and
>>>> references what looks like an armv8 instruction (I recall the name
>>>> containing 'armv8')
>>>> Perhaps it is related?
>>>> Cannot get more detail to send-pr as I've wiped the install, but
>>>> there's clearly a problem with libcrypt in -current, that isn't there
>>>> in 8.0_BETA.
Main Index |
Thread Index |