public inbox for gcc-help@gcc.gnu.org
 help / color / mirror / Atom feed
* mfix-cortex-a53
@ 2024-03-04 13:09 Christer Solskogen
  2024-03-07 15:01 ` mfix-cortex-a53 Kyrylo Tkachov
  0 siblings, 1 reply; 5+ messages in thread
From: Christer Solskogen @ 2024-03-04 13:09 UTC (permalink / raw)
  To: gcc-help

The mfix-cortex-a53 options(-mfix-cortex-a53-835769 /
-mno-fix-cortex-a53-835769 and -mfix-cortex-a53-843419/
-mno-fix-cortex-a53-843419), which of them are used by default? I mean, 
when none of them are used.

-- 
chs


^ permalink raw reply	[flat|nested] 5+ messages in thread

* RE: mfix-cortex-a53
  2024-03-04 13:09 mfix-cortex-a53 Christer Solskogen
@ 2024-03-07 15:01 ` Kyrylo Tkachov
  2024-03-07 16:50   ` mfix-cortex-a53 Christer Solskogen
  0 siblings, 1 reply; 5+ messages in thread
From: Kyrylo Tkachov @ 2024-03-07 15:01 UTC (permalink / raw)
  To: Christer Solskogen, gcc-help

Hello,

> -----Original Message-----
> From: Gcc-help <gcc-help-bounces+kyrylo.tkachov=arm.com@gcc.gnu.org> On
> Behalf Of Christer Solskogen via Gcc-help
> Sent: Monday, March 4, 2024 1:09 PM
> To: gcc-help@gcc.gnu.org
> Subject: mfix-cortex-a53
> 
> The mfix-cortex-a53 options(-mfix-cortex-a53-835769 /
> -mno-fix-cortex-a53-835769 and -mfix-cortex-a53-843419/
> -mno-fix-cortex-a53-843419), which of them are used by default? I mean,
> when none of them are used.

By default the fixes will not be applied i.e. the behaviour will be equivalent to -mno-fix-cortex-a53-835769 and -mno-fix-cortex-a53-843419.
However, you (or your compiler vendor) can configure GCC during GCC compile time with the --enable-fix-cortex-a53-843419 or --enable-fix-cortex-a53-835769 options.
In that case the fixes will be applied automatically i.e. behave as if -mfix-cortex-a53-843419 or -mfix-cortex-a53-835769 were specified.
You can find out if your GCC distribution enables these configuration options automatically by running gcc with the "-v" option, and seeing if any of those two --enable-fix* options appear in the configuration options.

Thanks,
Kyrill 

> 
> --
> chs


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: mfix-cortex-a53
  2024-03-07 15:01 ` mfix-cortex-a53 Kyrylo Tkachov
@ 2024-03-07 16:50   ` Christer Solskogen
  0 siblings, 0 replies; 5+ messages in thread
From: Christer Solskogen @ 2024-03-07 16:50 UTC (permalink / raw)
  To: gcc-help

On 3/7/2024 4:01 PM, Kyrylo Tkachov via Gcc-help wrote:
> Hello,
> 
>> -----Original Message-----
>> From: Gcc-help <gcc-help-bounces+kyrylo.tkachov=arm.com@gcc.gnu.org> On
>> Behalf Of Christer Solskogen via Gcc-help
>> Sent: Monday, March 4, 2024 1:09 PM
>> To: gcc-help@gcc.gnu.org
>> Subject: mfix-cortex-a53
>>
>> The mfix-cortex-a53 options(-mfix-cortex-a53-835769 /
>> -mno-fix-cortex-a53-835769 and -mfix-cortex-a53-843419/
>> -mno-fix-cortex-a53-843419), which of them are used by default? I mean,
>> when none of them are used.
> 
> By default the fixes will not be applied i.e. the behaviour will be equivalent to -mno-fix-cortex-a53-835769 and -mno-fix-cortex-a53-843419.
> However, you (or your compiler vendor) can configure GCC during GCC compile time with the --enable-fix-cortex-a53-843419 or --enable-fix-cortex-a53-835769 options.
> In that case the fixes will be applied automatically i.e. behave as if -mfix-cortex-a53-843419 or -mfix-cortex-a53-835769 were specified.
> You can find out if your GCC distribution enables these configuration options automatically by running gcc with the "-v" option, and seeing if any of those two --enable-fix* options appear in the configuration options.
> 

Thank you!

Are there any good reason for NOT having them enabled?

-- 
chs



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: mfix-cortex-a53
  2024-03-07 19:09 mfix-cortex-a53 Wilco Dijkstra
@ 2024-03-08  6:41 ` Christer Solskogen
  0 siblings, 0 replies; 5+ messages in thread
From: Christer Solskogen @ 2024-03-08  6:41 UTC (permalink / raw)
  To: gcc-help

On 3/7/2024 8:09 PM, Wilco Dijkstra via Gcc-help wrote:
> Hi,
> 
>> Are there any good reason for NOT having them enabled?
> 
> Yes, there is a codesize and performance impact (and it can be
> nasty if you have a hot loop that crosses a page).
> 
> So only use the option if you're using Cortex-A53 and you're
> unlucky enough to have an early revision with the bug.
> It is absolutely not needed for any newer core.
> 

Thanks a lot!

-- 
chs



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: mfix-cortex-a53
@ 2024-03-07 19:09 Wilco Dijkstra
  2024-03-08  6:41 ` mfix-cortex-a53 Christer Solskogen
  0 siblings, 1 reply; 5+ messages in thread
From: Wilco Dijkstra @ 2024-03-07 19:09 UTC (permalink / raw)
  To: christer.solskogen; +Cc: gcc-help

Hi,

> Are there any good reason for NOT having them enabled?

Yes, there is a codesize and performance impact (and it can be
nasty if you have a hot loop that crosses a page).

So only use the option if you're using Cortex-A53 and you're
unlucky enough to have an early revision with the bug.
It is absolutely not needed for any newer core.

Cheers,
Wilco

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2024-03-08  6:41 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-03-04 13:09 mfix-cortex-a53 Christer Solskogen
2024-03-07 15:01 ` mfix-cortex-a53 Kyrylo Tkachov
2024-03-07 16:50   ` mfix-cortex-a53 Christer Solskogen
2024-03-07 19:09 mfix-cortex-a53 Wilco Dijkstra
2024-03-08  6:41 ` mfix-cortex-a53 Christer Solskogen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).