netdev - Re: BCM4401 LAN

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] [day] [month] [year] [list]

Date:   Thu, 24 Nov 2022 12:12:03 +0100
From:   Francois Romieu <romieu@...zoreil.com>
To:     Ioannis Barkas <jnyb.de@...il.com>
Cc:     netdev@...r.kernel.org, Michael Chan <michael.chan@...adcom.com>
Subject: Re: BCM4401 LAN

(MAINTAINERS file suggests to Cc: Michael Chan Cced as b44 maintainer)

Ioannis Barkas <jnyb.de@...il.com> :
[...]
> I resurrected some old PCs at job and found one with BCM4401 LAN. This
> LAN had some issues as I read in various OSes so I gave it a try using
> live Ubuntu 22.10. It worked fine though I got this when I plugged in
> the Ethernet cable:
> [  517.928052] b44 ssb0:0 eth0: Link is up at 100 Mbps, full duplex
> [  517.928066] b44 ssb0:0 eth0: Flow control is off for TX and off for RX
> [  517.928246] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
> [  520.697405] ------------[ cut here ]------------
> [  520.697417] b44 0000:0a:0b.0: DMA addr 0x00000000766d9402+42
> overflow (mask 3fffffff, bus limit 0).
> [  520.697438] WARNING: CPU: 1 PID: 1187 at kernel/dma/direct.h:97
> dma_direct_map_page+0x1f1/0x200
> [  520.697456] Modules linked in: ntfs3 snd_seq_dummy snd_hrtimer
> binfmt_misc zfs(PO) zunicode(PO) zzstd(O) zlua(O) zavl(PO) icp(PO)
> zcommon(PO) znvpair(PO) spl(O) snd_intel8x0 snd_ac97_codec ac97_bus
[...]
> This is the LAN device:
> 0a:0b.0 Ethernet controller [0200]: Broadcom Inc. and subsidiaries
> BCM4401-B0 100Base-TX [14e4:170c] (rev 02)
>     Subsystem: Lenovo BCM4401-B0 100Base-TX [17aa:1004]
>     Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
> ParErr- Stepping- SERR+ FastB2B- DisINTx-
>     Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort-
> <TAbort- <MAbort- >SERR- <PERR- INTx-
>     Latency: 32
>     Interrupt: pin A routed to IRQ 23
>     Region 0: Memory at d0000000 (32-bit, non-prefetchable) [size=8K]
>     Capabilities: [40] Power Management version 2
>         Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA
> PME(D0+,D1+,D2+,D3hot+,D3cold+)
>         Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=2 PME-
>     Kernel driver in use: b44
>     Kernel modules: b44
[...]
> Does anything look awfully wrong in the trace?

See kernel/dma/direct.c::dma_direct_map_resource for a WARN_ON_ONCE
message related to the comment below in drivers/net/ethernet/broadcom/b44.c

static netdev_tx_t b44_start_xmit(struct sk_buff *skb, struct net_device *dev)
[...]
        mapping = dma_map_single(bp->sdev->dma_dev, skb->data, len, DMA_TO_DEVICE);
        if (dma_mapping_error(bp->sdev->dma_dev, mapping) || mapping + len > DMA_BIT_MASK(30)) {
                struct sk_buff *bounce_skb;

                /* Chip can't handle DMA to/from >1GB, use bounce buffer */
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                if (!dma_mapping_error(bp->sdev->dma_dev, mapping))
                        dma_unmap_single(bp->sdev->dma_dev, mapping, len, 
                                             DMA_TO_DEVICE);

It's a bit less than optimal performance-wise but it is expected to work.

This one was rather trivial but please avoid requests related to giant
out-of-tree (zfs) module tainted kernel in the future.

-- 
Ueimor