Re: bce(4) on IBM BladeCenter HS22

On Mon, Jan 18, 2010 at 09:14:39PM +0100, Oliver Fromme wrote:
> Pyun YongHyeon <pyunyh_at_gmail.com> wrote:
>  > Ok, now let's open new thread for bce(4) issues.
> 
> Ok, it took a little longer because I re-installed one of the
> blades with 9-current as of today.
> 
> Again, these are "HS22" blades installed in an IBM BladeCenter.
> There are two bce(4) interfaces that don't attach correctly.
> The problem seems to be the same as in PRs kern/136417 and
> kern/139761 (which are duplicates of each other).
> 
> Here's the excerpt from dmesg (verbose kernel):
> 
>  bce0: <Broadcom NetXtreme II BCM5709 1000Base-SX (C0)> mem 0x92000000-0x93ffffff irq 30 at device 0.0 on pci16
>  bce0: attempting to allocate 1 MSI vectors (16 supported)
>  msi: routing MSI IRQ 260 to local APIC 0 vector 112
>  bce0: using IRQ 260 for MSI
>  bce0: /usr/src/sys/dev/bce/if_bce.c(1097): No PHY found on child MII bus!
>  device_attach: bce0 attach returned 6
>  bce1: <Broadcom NetXtreme II BCM5709 1000Base-SX (C0)> mem 0x94000000-0x95ffffff irq 37 at device 0.1 on pci16
>  bce1: attempting to allocate 1 MSI vectors (16 supported)
>  msi: routing MSI IRQ 260 to local APIC 0 vector 112
>  bce1: using IRQ 260 for MSI
>  bce1: /usr/src/sys/dev/bce/if_bce.c(1097): No PHY found on child MII bus!
>  device_attach: bce1 attach returned 6
> 
> Attaching the driver fails, so ifconfig doesn't list anything.
> The ports are fiber, i.e. they should appear as 1000baseSX.
> 
> This is what pciconf -lcv says:
> 
>  bce0_at_pci0:16:0:0:	class=0x020000 card=0x03701014 chip=0x163a14e4 rev=0x20 hdr=0x00
>     vendor     = 'Broadcom Corporation'
>     device     = 'NetXtreme II BCM5709S Gigabit Ethernet'
>     class      = network
>     subclass   = ethernet
>     cap 01[48] = powerspec 3  supports D0 D3  current D0
>     cap 03[50] = VPD
>     cap 05[58] = MSI supports 16 messages, 64 bit 
>     cap 11[a0] = MSI-X supports 9 messages in map 0x10
>     cap 10[ac] = PCI-Express 2 endpoint max data 256(512) link x4(x4)
>  bce1_at_pci0:16:0:1:	class=0x020000 card=0x03701014 chip=0x163a14e4 rev=0x20 hdr=0x00
>     vendor     = 'Broadcom Corporation'
>     device     = 'NetXtreme II BCM5709S Gigabit Ethernet'
>     class      = network
>     subclass   = ethernet
>     cap 01[48] = powerspec 3  supports D0 D3  current D0
>     cap 03[50] = VPD
>     cap 05[58] = MSI supports 16 messages, 64 bit 
>     cap 11[a0] = MSI-X supports 9 messages in map 0x10
>     cap 10[ac] = PCI-Express 2 endpoint max data 256(512) link x4(x4)
> 
> And the relevant parts from devinfo -rv:
> 
>  pcib6 pnpinfo vendor=0x8086 device=0x340e subvendor=0x1014 subdevice=0x340e class=0x060400 at slot=7 function=0 handle=\_SB_.PCI0.ETH1
>    pci16
>      bce0 pnpinfo vendor=0x14e4 device=0x163a subvendor=0x1014 subdevice=0x0370 class=0x020000 at slot=0 function=0
>          I/O memory addresses:
>              0x92000000-0x93ffffff
>      bce1 pnpinfo vendor=0x14e4 device=0x163a subvendor=0x1014 subdevice=0x0370 class=0x020000 at slot=0 function=1
>          Interrupt request lines:
>              260
>          I/O memory addresses:
>              0x94000000-0x95ffffff
> 
> If you need more information, or want me to test any patches,
> please let me know.
> 

Ok, it seems brgphy(4) does not know the PHY. Let's see what PHY
you have. Please apply attached patch and let me know the output of
the patch. It will show PHYID1 and PHYID2 value.