Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <786f58e8-aa61-d439-c9bb-4a27599d2aa5@gmx.de>
Date:   Tue, 24 May 2022 18:44:59 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.9.0
Subject: Re: [PATCH AUTOSEL 5.10 2/8] parisc: Disable debug code regarding
 cache flushes in handle_nadtlb_fault()
Content-Language: en-US
To:     Sasha Levin <sashal@kernel.org>, linux-kernel@vger.kernel.org,
        stable@vger.kernel.org
Cc:     John David Anglin <dave.anglin@bell.net>,
        James.Bottomley@HansenPartnership.com, akpm@linux-foundation.org,
        zhengqi.arch@bytedance.com, linux-parisc@vger.kernel.org
References: <20220524160035.827109-1-sashal@kernel.org>
 <20220524160035.827109-2-sashal@kernel.org>
From:   Helge Deller <deller@gmx.de>
In-Reply-To: <20220524160035.827109-2-sashal@kernel.org>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hello Sascha,

On 5/24/22 18:00, Sasha Levin wrote:
> From: John David Anglin <dave.anglin@bell.net>
>
> [ Upstream commit 67c35a3b646cc68598ff0bb28de5f8bd7b2e81b3 ]
>
> Change the "BUG" to "WARNING" and disable the message because it trigger=
s
> occasionally in spite of the check in flush_cache_page_if_present.

Please drop this patch from the backporting-queue (v5.10, v5.15 and v5.17)=
.
It's not necessary since the warning will only trigger on v5.18 on machine=
s
with PA8800/PA8900 processors.

Thanks.
Helge


> The pte value extracted for the "from" page in copy_user_highpage is rac=
y and
> occasionally the pte is cleared before the flush is complete.  I assume =
that
> the page is simultaneously flushed by flush_cache_mm before the pte is c=
leared
> as nullifying the fdc doesn't seem to cause problems.
>
> I investigated various locking scenarios but I wasn't able to find a way=
 to
> sequence the flushes.  This code is called for every COW break and locks=
 impact
> performance.
>
> This patch is related to the bigger cache flush patch because we need th=
e pte
> on PA8800/PA8900 to flush using the vma context.
> I have also seen this from copy_to_user_page and copy_from_user_page.
>
> The messages appear infrequently when enabled.
>
> Signed-off-by: John David Anglin <dave.anglin@bell.net>
> Signed-off-by: Helge Deller <deller@gmx.de>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  arch/parisc/mm/fault.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/arch/parisc/mm/fault.c b/arch/parisc/mm/fault.c
> index 5faa3cff4738..2472780d4039 100644
> --- a/arch/parisc/mm/fault.c
> +++ b/arch/parisc/mm/fault.c
> @@ -22,6 +22,8 @@
>
>  #include <asm/traps.h>
>
> +#define DEBUG_NATLB 0
> +
>  /* Various important other fields */
>  #define bit22set(x)		(x & 0x00000200)
>  #define bits23_25set(x)		(x & 0x000001c0)
> @@ -449,8 +451,8 @@ handle_nadtlb_fault(struct pt_regs *regs)
>  		fallthrough;
>  	case 0x380:
>  		/* PDC and FIC instructions */
> -		if (printk_ratelimit()) {
> -			pr_warn("BUG: nullifying cache flush/purge instruction\n");
> +		if (DEBUG_NATLB && printk_ratelimit()) {
> +			pr_warn("WARNING: nullifying cache flush/purge instruction\n");
>  			show_regs(regs);
>  		}
>  		if (insn & 0x20) {