linux-kernel - Re: [PATCH v3 1/2] perf util: Check for fused instruction

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 for Android: free password hash cracker in your pocket

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

Date:   Fri, 7 Jul 2017 09:09:51 +0800
From:   "Jin, Yao" <yao.jin@...ux.intel.com>
To:     Arnaldo Carvalho de Melo <acme@...nel.org>
Cc:     jolsa@...nel.org, peterz@...radead.org, mingo@...hat.com,
        alexander.shishkin@...ux.intel.com, Linux-kernel@...r.kernel.org,
        ak@...ux.intel.com, kan.liang@...el.com, yao.jin@...el.com
Subject: Re: [PATCH v3 1/2] perf util: Check for fused instruction

>> +
>> +static bool x86__ins_is_fused(char *cpuid, const char *ins1, const char *ins2)
>> +{
>> +	unsigned int family, model, stepping;
>> +	int ret;
>> +
>> +	/*
>> +	 * cpuid = "GenuineIntel,family,model,stepping"
>> +	 */
>> +	ret = sscanf(cpuid, "%*[^,],%u,%u,%u", &family, &model, &stepping);
> So, looking at the next patch, that uses this, I see that you'll call
> this everytime that jump arrow will be printed, why not do this when
> doing the initial disassembly, and having this info cached in the struct
> ins or disasm_line (haven't looked closely to provide exact instructions
> on how to do it)?

Yes, you're right. It doesn't need to do this model checking each time.

I will move the model checking to other place and record the model in 
annotate_browser.

> Even more, you could do this model checking just once per disassembly,
> then use it as you go reading the disassembly lines, marking them as
> fused/not fused and then at jump arror printing just look at a flag, no?
>
> - Arnaldo
>

For this, I have another consideration.

The fused instruction pair consists of 2 instructions. One is jump, the 
other is "CMP/TEST/...".

If I check the fused instruction pair once per disassembly, I need to 
check if one instruction is jump and the other is "CMP/TEST/..." for all 
instructions.

While in annotate_browser__draw_current_jump(), it has already checked 
if a disasm_line is a valid jump, so the jump checking will be duplicated.

The current implementation is just performing the fused checking when 
user moves the cursor on the jump instruction. It doesn't need to add 
additional jump checking.

So the current way may be acceptable as well?

Thanks
Jin Yao