netdev - Re: [PATCH net] Revert "r8169: don't try to disable interrupts if NAPI is, scheduled already"

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: <3d6e07fc-cef7-4480-a12d-1d70930cba1b@gmail.com>
Date: Mon, 20 May 2024 16:55:34 +0100
From: Ken Milmore <ken.milmore@...il.com>
To: Heiner Kallweit <hkallweit1@...il.com>, Eric Dumazet
 <edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>,
 Jakub Kicinski <kuba@...nel.org>, David Miller <davem@...emloft.net>,
 Realtek linux nic maintainers <nic_swsd@...ltek.com>
Cc: "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH net] Revert "r8169: don't try to disable interrupts if
 NAPI is, scheduled already"

I concur, Heiner's patch is correct. Sorry about the noise.

On 20/05/2024 14:50, Heiner Kallweit wrote:
> On 15.05.2024 08:18, Heiner Kallweit wrote:
>> This reverts commit 7274c4147afbf46f45b8501edbdad6da8cd013b9.
>>
>> Ken reported that RTL8125b can lock up if gro_flush_timeout has the
>> default value of 20000 and napi_defer_hard_irqs is set to 0.
>> In this scenario device interrupts aren't disabled, what seems to
>> trigger some silicon bug under heavy load. I was able to reproduce this
>> behavior on RTL8168h. Fix this by reverting 7274c4147afb.
>>
>> Fixes: 7274c4147afb ("r8169: don't try to disable interrupts if NAPI is scheduled already")
>> Cc: stable@...r.kernel.org
>> Reported-by: Ken Milmore <ken.milmore@...il.com>
>> Signed-off-by: Heiner Kallweit <hkallweit1@...il.com>
>> ---
> 
> This patch was mistakenly set to "changes requested".
> The replies from Ken provide additional details on how interrupt mask
> and status register behave on these chips. However the patch itself
> is correct and should be applied.
>