netdev - Re: Experimental fix for MSI-X issue on r8169

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]

Date:   Tue, 21 Aug 2018 16:32:45 -0700 (PDT)
From:   David Miller <davem@...emloft.net>
To:     hkallweit1@...il.com
Cc:     jian-hong@...lessm.com, steved424@...il.com, gogen@...root.org,
        netdev@...r.kernel.org
Subject: Re: Experimental fix for MSI-X issue on r8169

From: Heiner Kallweit <hkallweit1@...il.com>
Date: Tue, 21 Aug 2018 23:19:04 +0200

> That's what I get on my system (RTL8168E-VL). In your case you'll come
> only till the first suspend.
> 
> [    3.743404] r8169 0000:03:00.0: MSI-X entry: context probe: fee01004 0 40ef 1

On probe, MSI-X is masked (ie. disabled) and is configured to use:

	address: 0x00000000fee01004
	data:    0x00000000000040ef

> [   29.539250] r8169 0000:03:00.0: MSI-X entry: context suspend: fee02004 0 4028 0

At suspend time, MSI-X is unmasked (ie. enabled) and is configured to use:

	address: 0x00000000fee01004
	data:    0x0000000000004028

> [   29.837457] r8169 0000:03:00.0: MSI-X entry: context resume: fee01004 0 402b 0

At reume time, MSI-X is unmasked (ie. enabled) and is configured to use:

	address: 0x00000000fee01004
	data:    0x000000000000402b

> [   36.921370] r8169 0000:03:00.0: MSI-X entry: context suspend: fee01004 0 402b 0

Second suspend:

	address: 0x00000000fee01004
	data:    0x000000000000402b

> [   37.239407] r8169 0000:03:00.0: MSI-X entry: context resume: fee01004 0 402b 0

Second resume:

	address: 0x00000000fee01004
	data:    0x000000000000402b

And this all looks normal.  The data field is changing when you first up the device
and interrupts are enabled.  This is where the request_irq happens, the MSI
vector is allocated, and that vector number is written to the data field of the
MSI-X entry.

It looks like this (re-)allocation of MSI vectors happens on every resume as well.

And that's why the data field changes each resume.