Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 23 Sep 2016 08:38:30 +0200
From:      Andrea Venturoli <ml@netfence.it>
To:        Ralf Mardorf <ralf.mardorf@rocketmail.com>, "freebsd-questions@freebsd.org" <freebsd-questions@freebsd.org>
Subject:   Re: firefox crash
Message-ID:  <5d9a7a6b-b951-20e4-5557-f9bcd03924b2@netfence.it>
In-Reply-To: <20160922100646.26ef2406@archlinux.localdomain>
References:  <d5cb94ad-9dc0-7102-1f58-cd6eedb72521@hiwaay.net> <20160919172525.2857fcee@archlinux.localdomain> <de6d50e8-f63e-d303-4075-f8fe3dd13c00@hiwaay.net> <20160920053641.e9075ee2.freebsd@edvax.de> <810b54d9-9769-eb99-8809-ea22a09f6d65@hiwaay.net> <1e50daef-c0e6-813b-2295-be8d413b4eb5@hiwaay.net> <20160921075019.3e68d68b@archlinux.localdomain> <1bce8b64-b993-6e1e-2fee-f7c3cfe0c4ca@netfence.it> <20160922100646.26ef2406@archlinux.localdomain>

next in thread | previous in thread | raw e-mail | index | archive | help
On 09/22/16 10:06, Ralf Mardorf via freebsd-questions wrote:

(I'm resending this because I don't see it on the list; please forgive 
me if you receive this twice).




> You noticed the follow-up?

Yes, I think my problem is solved (no crash in two days)!



> Well, a practical workaround is always to build the port yourself using
> different OPTIONs :-)

That's what I've always done: however I did not know which option to 
tweak until I saw your message yesterday.



> Jokes aside, your point is of course valid. It seems there are still a
> large enough number of i386 CPUs in production that can't work with
> the SSE/SSE2-optimisations in ffmpeg.

My CPU is actually 64b; it's running i386 since it was an upgrade to an 
old system, built more than 10 (possibly even 15) years ago. I just 
never felt an urge to convert since it's working fine enough for me.
However I notice an ever-growing number of ports that don't seem to get 
too much attention on this architecture.



> We might need to drop these
> optimisations in the set of default OPTIONs on i386.

I guess so; knowing my CPU should support SSE I've always enabled it, 
however. Maybe some warning is better?



> In your particular case: What exact CPU is this?

  From dmesg:
> CPU: AMD Phenom(tm) II X4 925 Processor (2800.23-MHz 686-class CPU)
>   Origin="AuthenticAMD"  Id=0x100f42  Family=0x10  Model=0x4  Stepping=2
>   Features=0x178bfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,MMX,FXSR,SSE,SSE2,HTT>
>   Features2=0x802009<SSE3,MON,CX16,POPCNT>
>   AMD Features=0xee500800<SYSCALL,NX,MMX+,FFXSR,Page1GB,RDTSCP,LM,3DNow!+,3DNow!>
>   AMD Features2=0x37ff<LAHF,CMP,SVM,ExtAPIC,CR8,ABM,SSE4A,MAS,Prefetch,OSVW,IBS,SKINIT,WDT>
>   SVM: NP,NRIP,NAsids=64
>   TSC: P-state invariant

Is this known to be buggy (and perhaps fixable via a BIOS upgrade)? Or 
is it a ffmpeg bug? Clang "feature"? ...?



> Could you confirm
> that building an ffmpeg package without the SSE OPTION resolves the
> issue for you?

Yes.
I wasn't able to watch Youtube before; I can now (not that I care that 
much, but it's a quite sure test :).
I had FireFox crashing every few minutes (as soon as I landed on a page 
with a nice animated advertisement that AdBlock Plus would let through); 
now, as I said, it hasn't crashed in the last two days.



   bye & Thanks
	av.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5d9a7a6b-b951-20e4-5557-f9bcd03924b2>