linux-kernel - Re: [PATCH] x86/fred: Optimize the FRED entry by prioritizing high-probability event dispatching

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]

Message-ID:
 <TYZPR03MB8801DBB1463548FD1AD818D2D11B2@TYZPR03MB8801.apcprd03.prod.outlook.com>
Date: Fri, 17 Jan 2025 08:40:54 +0800
From: Ethan Zhao <etzhao@...look.com>
To: Nikolay Borisov <nik.borisov@...e.com>,
 Ethan Zhao <haifeng.zhao@...ux.intel.com>, linux-kernel@...r.kernel.org,
 stable@...r.kernel.org
Cc: tglx@...utronix.de, dave.hansen@...ux.intel.com, x86@...nel.org,
 hpa@...or.com, xin@...or.com, andrew.cooper3@...rix.com, mingo@...hat.com,
 bp@...en8.de
Subject: Re: [PATCH] x86/fred: Optimize the FRED entry by prioritizing
 high-probability event dispatching


On 1/16/2025 11:08 PM, Nikolay Borisov wrote:
>
>
> On 16.01.25 г. 8:51 ч., Ethan Zhao wrote:
>> External interrupts (EVENT_TYPE_EXTINT) and system calls 
>> (EVENT_TYPE_OTHER)
>> occur more frequently than other events in a typical system. 
>> Prioritizing
>> these events saves CPU cycles and optimizes the efficiency of 
>> performance-
>> critical paths.
>>
> <snip>
>
> Can you also include some performance numbers?

Of coz will do when timing is good :)

Thanks,
Ethan

>
>> Signed-off-by: Ethan Zhao <haifeng.zhao@...ux.intel.com>
>> ---
>> base commit: 619f0b6fad524f08d493a98d55bac9ab8895e3a6
>> ---
>>   arch/x86/entry/entry_fred.c | 25 +++++++++++++++++++------
>>   1 file changed, 19 insertions(+), 6 deletions(-)
>>
>> diff --git a/arch/x86/entry/entry_fred.c b/arch/x86/entry/entry_fred.c
>> index f004a4dc74c2..591f47771ecf 100644
>> --- a/arch/x86/entry/entry_fred.c
>> +++ b/arch/x86/entry/entry_fred.c
>> @@ -228,9 +228,18 @@ __visible noinstr void 
>> fred_entry_from_user(struct pt_regs *regs)
>>       /* Invalidate orig_ax so that syscall_get_nr() works correctly */
>>       regs->orig_ax = -1;
>>   -    switch (regs->fred_ss.type) {
>> -    case EVENT_TYPE_EXTINT:
>> +    if (regs->fred_ss.type == EVENT_TYPE_EXTINT)
>>           return fred_extint(regs);
>> +    else if (regs->fred_ss.type == EVENT_TYPE_OTHER)
>> +        return fred_other(regs);
>> +
>> +    /*
>> +     * Dispatch EVENT_TYPE_EXTINT and EVENT_TYPE_OTHER(syscall) type 
>> events
>> +     * first due to their high probability and let the compiler 
>> create binary search
>> +     * dispatching for the remaining events
>> +     */
>
> nit: At least to me it makes sense to have the comment above the 'if' 
> so that the flow is linear.
>
>> +
>> +    switch (regs->fred_ss.type) {
>>       case EVENT_TYPE_NMI:
>>           if (likely(regs->fred_ss.vector == X86_TRAP_NMI))
>>               return fred_exc_nmi(regs);
>> @@ -245,8 +254,6 @@ __visible noinstr void 
>> fred_entry_from_user(struct pt_regs *regs)
>>           break;
>>       case EVENT_TYPE_SWEXC:
>>           return fred_swexc(regs, error_code);
>> -    case EVENT_TYPE_OTHER:
>> -        return fred_other(regs);
>>       default: break;
>>       }
>>   @@ -260,9 +267,15 @@ __visible noinstr void 
>> fred_entry_from_kernel(struct pt_regs *regs)
>>       /* Invalidate orig_ax so that syscall_get_nr() works correctly */
>>       regs->orig_ax = -1;
>>   -    switch (regs->fred_ss.type) {
>> -    case EVENT_TYPE_EXTINT:
>> +    if (regs->fred_ss.type == EVENT_TYPE_EXTINT)
>>           return fred_extint(regs);
>> +
>> +    /*
>> +     * Dispatch EVENT_TYPE_EXTINT type event first due to its high 
>> probability
>> +     * and let the compiler do binary search dispatching for the 
>> other events
>> +     */
>> +
>> +    switch (regs->fred_ss.type) {
>>       case EVENT_TYPE_NMI:
>>           if (likely(regs->fred_ss.vector == X86_TRAP_NMI))
>>               return fred_exc_nmi(regs);
>