Linux PARISC architecture development
 help / color / mirror / Atom feed
From: Helge Deller <deller@gmx.de>
To: John David Anglin <dave.anglin@bell.net>,
	linux-parisc@vger.kernel.org,
	James Bottomley <James.Bottomley@hansenpartnership.com>
Subject: Re: [PATCH][RFC] parisc: Use local tlb purges only on PA2.0 machines
Date: Sun, 25 Sep 2022 22:11:08 +0200	[thread overview]
Message-ID: <1de28183-2252-04dc-9838-d695161a01ed@gmx.de> (raw)
In-Reply-To: <1063fc54-738a-a1b9-fb53-8202287805c7@bell.net>

On 9/25/22 22:02, John David Anglin wrote:
> On 2022-09-25 2:57 p.m., Helge Deller wrote:
>> +#ifdef CONFIG_PA20
>> +#define ALT_COND_PACACHE    ALT_COND_ALWAYS
>> +#else
>> +#define ALT_COND_PACACHE    ALT_COND_NO_SMP
>> +#endif
>> +
>>   ENTRY_CFI(flush_tlb_all_local)
>>       /*
>>        * The pitlbe and pdtlbe instructions should only be used to
>> @@ -539,15 +545,10 @@ ENTRY_CFI(copy_user_page_asm)
>>
>>       /* Purge any old translations */
>>
>> -#ifdef CONFIG_PA20
>> -    pdtlb,l        %r0(%r28)
>> -    pdtlb,l        %r0(%r29)
>> -#else
>>   0:    pdtlb        %r0(%r28)
>>   1:    pdtlb        %r0(%r29)
>> -    ALTERNATIVE(0b, 0b+4, ALT_COND_NO_SMP, INSN_PxTLB)
>> -    ALTERNATIVE(1b, 1b+4, ALT_COND_NO_SMP, INSN_PxTLB)
>> -#endif
>> +    ALTERNATIVE(0b, 0b+4, ALT_COND_PACACHE, INSN_PxTLB)
>> +    ALTERNATIVE(1b, 1b+4, ALT_COND_PACACHE, INSN_PxTLB)
> This doesn't look correct.  If ALT_COND_PACACHE is defined as ALT_COND_NO_SMP, the pdtlb
> instructions will be converted to pdtlb,l instructions when running UP.  These are not supported
> on PA 1.1. 

Your concern is correct, but there is an additonal check in the alternative-coding,
which prevents enabling the local flag if we're not running on a PA2.0 CPU.
So, those ALTERNATIVE() macros will only apply on PA2.0 machines.

> When running SMP, we have pdtlb purges.  For these, we need bus serialization.
> 
> These also don't look correct:
> 
> #define pdtlb(sr, addr) asm volatile("pdtlb 0(%%sr%0,%1)" \
>                          ALTERNATIVE(ALT_COND_NO_SMP, INSN_PxTLB) \
>                          : : "i"(sr), "r" (addr) : "memory")
> #define pitlb(sr, addr) asm volatile("pitlb 0(%%sr%0,%1)" \
>                          ALTERNATIVE(ALT_COND_NO_SMP, INSN_PxTLB) \
>                          ALTERNATIVE(ALT_COND_NO_SPLIT_TLB, INSN_NOP) \
>                          : : "i"(sr), "r" (addr) : "memory")
> 
> We only have local purge on PA 2.0 machines.

Same as above.

Helge

  reply	other threads:[~2022-09-25 20:11 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-25 18:57 [PATCH][RFC] parisc: Use local tlb purges only on PA2.0 machines Helge Deller
2022-09-25 20:02 ` John David Anglin
2022-09-25 20:11   ` Helge Deller [this message]
2022-09-25 20:28     ` John David Anglin
2022-09-27 20:06       ` Helge Deller
2022-09-27 22:06         ` John David Anglin
2022-09-28  9:25           ` Helge Deller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1de28183-2252-04dc-9838-d695161a01ed@gmx.de \
    --to=deller@gmx.de \
    --cc=James.Bottomley@hansenpartnership.com \
    --cc=dave.anglin@bell.net \
    --cc=linux-parisc@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox