[gPXE-devel] Attempting to add support for Broadcom BCM5755M (14e4:1673) ethernet adapter

Tue Jan 3 07:42:49 EST 2012

Hi,

Success!  I think it is working now.  I used the following command along
with the snapshot:
make DEBUG=tg3_main,tg3_hw,tg3_phy

Serial output follows:
-------------------------------------
ok

gPXE 1.0.1+ -- Open Source Boot Firmware -- http://etherboot.org
Features: AoE HTTP iSCSI DNS TFTP bzImage COMBOOT ELF Multiboot PXE PXEXT

Link is down
net0: 00:1c:23:26:82:e7 on PCI09:00.0 (open)
  [Link:down, TX:0 TXE:0 RX:0 RXE:0]
  [Link status: The socket is not connected (0x38086001)]
Waiting for link-up on net0...Link is up at 1000 Mbps, full duplex
Flow control is off for TX and off for RX
 ok
DHCP (net0 00:1c:23:26:82:e7)... ok
net0: 172.27.25.204/255.255.254.0 gw 172.27.24.1
No filename or root path specified
No more network devices

-------------------------------------

How about you?

br,
Quinn

On Tue, Jan 3, 2012 at 1:14 PM, Quinn Plattel <qiet72 at gmail.com> wrote:

> Should probably forward this to the mailing list in case somebody else has
> a solution. :-)
>
>
> ---------- Forwarded message ----------
> From: Quinn Plattel <qiet72 at gmail.com>
> Date: Tue, Jan 3, 2012 at 1:04 PM
> Subject: Re: [gPXE-devel] Attempting to add support for Broadcom BCM5755M
> (14e4:1673) ethernet adapter
> To: Chris Pitzel <chris at pitzel.dyndns.org>
>
>
> Hi,
>
> I found out my problem.  I didn't have Thomas's changes, I was missing the
> tg3_main.c file in my source tree, so I download his snapshot instead here:
>
> http://git.etherboot.org/people/meteger/gpxe.git/snapshot/951eee4530cf318a21d0a3ee71f2fa5edca16fc2.tar.gz
>
> Now I can build without errors, but there is not much debugging
> information:
>
> -------------------------------
>
> dma test buffer, virt: 0x9a070 phys: 0x7fe11000
> ok
>
>
>
> gPXE 1.0.1+ -- Open Source Boot Firmware -- http://etherboot.org
> Features: AoE HTTP iSCSI DNS TFTP bzImage COMBOOT ELF Multiboot PXE PXEXT
>
>
> ----------------------------------------
>
> I don't think it is detecting the card, I'll check the pci ids in the
> source file.
>
> br,
> Quinn
>
>
> On Tue, Jan 3, 2012 at 12:16 PM, Chris Pitzel <chris at pitzel.dyndns.org>wrote:
>
>> I did git clone git://git.etherboot.org/scm/gpxe.git****
>>
>> ** **
>>
>> And then i applied the tg3 driver commit from Thomas (
>> http://git.etherboot.org/people/meteger/gpxe.git/commit/951eee4530cf318a21d0a3ee71f2fa5edca16fc2
>> ).****
>>
>>
>> ****
>>
>> And nothing else.****
>>
>> ** **
>>
>> I’m a total newbie to git; this might be a dumb question, but doesn’t
>> “git clone” pull the latest tree at the time its run?  I’ve been working
>> off of the ipxe code until our convo when you inspired me to give the gpxe
>> tg3 driver a shot, so i git cloned literally yesterday.****
>>
>> ** **
>>
>> (on that note...i hope the ipxe and gpxe groups can sort out their
>> differences because its rather counterproductive what’s going on right
>> now....)****
>>
>> ** **
>>
>> ** **
>>
>> ** **
>>
>> *From:* Quinn Plattel [mailto:qiet72 at gmail.com]
>> *Sent:* Tuesday, January 03, 2012 5:08 AM
>>
>> *To:* Chris Pitzel; gpxe-devel at etherboot.org; gpxe at etherboot.org
>> *Subject:* Re: [gPXE-devel] Attempting to add support for Broadcom
>> BCM5755M (14e4:1673) ethernet adapter****
>>
>> ** **
>>
>> Hi,
>>
>> Did you do a git pull?
>>
>> git clone git://git.etherboot.org/scm/gpxe.git
>> git pull
>>
>>
>> br,
>> Quinn****
>>
>> On Tue, Jan 3, 2012 at 11:55 AM, Chris Pitzel <chris at pitzel.dyndns.org>
>> wrote:****
>>
>> Strange; builds here just fine (ignore my sloppiness in being root).  ***
>> *
>>
>>  ****
>>
>> root at pitzel:~/gpxe/src# make bin/14e41673.usb****
>>
>>   [LD] bin/14e41673.hd.tmp****
>>
>>   [BIN] bin/14e41673.hd.bin****
>>
>>   [ZINFO] bin/14e41673.hd.zinfo****
>>
>>   [ZBIN] bin/14e41673.hd.zbin****
>>
>>   [FINISH] bin/14e41673.hd****
>>
>>   [FINISH] bin/14e41673.usb****
>>
>> rm bin/14e41673.hd.zbin bin/14e41673.hd bin/14e41673.hd.zinfo
>> bin/14e41673.hd.b
>> in****
>>
>> root at pitzel:~/gpxe/src# make bin/14e41673.usb
>> DEBUG=tg3_main:3,tg3_hw,tg3_phy****
>>
>>   [BUILD] bin/tg3_main.dbg3.o****
>>
>>   [BUILD] bin/tg3_hw.dbg1.o****
>>
>>   [BUILD] bin/tg3_phy.dbg1.o****
>>
>>   [AR] bin/blib.a****
>>
>> ar: creating bin/blib.a****
>>
>>   [LD] bin/14e41673.hd.tmp****
>>
>>   [BIN] bin/14e41673.hd.bin****
>>
>>   [ZINFO] bin/14e41673.hd.zinfo****
>>
>>   [ZBIN] bin/14e41673.hd.zbin****
>>
>>   [FINISH] bin/14e41673.hd****
>>
>>   [FINISH] bin/14e41673.usb****
>>
>> rm bin/14e41673.hd.zbin bin/14e41673.hd bin/14e41673.hd.zinfo
>> bin/14e41673.hd.b
>> in****
>>
>> root at pitzel:~/gpxe/src#****
>>
>>  ****
>>
>>  ****
>>
>>  ****
>>
>>  ****
>>
>> *From:* Quinn Plattel [mailto:qiet72 at gmail.com]
>> *Sent:* Tuesday, January 03, 2012 4:46 AM****
>>
>>
>> *To:* Chris Pitzel; gpxe-devel at etherboot.org; gpxe at etherboot.org
>> *Subject:* Re: [gPXE-devel] Attempting to add support for Broadcom
>> BCM5755M (14e4:1673) ethernet adapter****
>>
>>  ****
>>
>> Nope, same issue:
>>
>> $ make bin/14E41673.usb DEBUG=tg3_main:3,tg3_hw,tg3_phy****
>>
>>   [BUILD] bin/usbdisk.o
>>   [OBJCOPY] bin/usbdisk.bin
>> make: *** No rule to make target `bin/tg3_main.dbg3.o', needed by
>> `bin/blib.a'. Stop.
>>
>> All of the following commands gives the same error:
>> make bin/14E41673.usb DEBUG=tg3_main:3,tg3_hw,tg3_phy
>> make bin/14e41673.usb DEBUG=tg3_main:3,tg3_hw,tg3_phy****
>>
>> make bin/14E41673.usb DEBUG=14E41673:3,tg3_hw,tg3_phy
>> make bin/14e41673.usb DEBUG=14e41673:3,tg3_hw,tg3_phy
>>
>> br,
>> Quinn****
>>
>> On Tue, Jan 3, 2012 at 11:36 AM, Chris Pitzel <chris at pitzel.dyndns.org>
>> wrote:****
>>
>> Yeah iSCSI works great on the undi driver.  Wanted to get native bcm5755
>> / tg3 iscsi working simply so I could take the laptop outside of my
>> dhcp/tftp environment and still use it to http/tftp-load stuff.****
>>
>>  ****
>>
>> Looks like you’re working with the (new tg3-driver patched) gpxe code
>> tree.  I think your error relates to how make interprets the tg3_main.usb
>> ‘target’.  ****
>>
>>  ****
>>
>> For example, if you go “make bin/80861043.usb”, it will build only the
>> code for PCI vendor 0x8086 (Intel), and PCI ID 0x1043, into the usb image.
>> ****
>>
>>  ****
>>
>> In your case, if you want a pure bcm5755m driver build, wouldn’t you do,
>> “make bin/14E41673.usb DEBUG=tg3_main,tg3_hw,tg3_phy” ???  That’s my
>> thought (but its 4:35am here....).****
>>
>>  ****
>>
>> Chris****
>>
>>  ****
>>
>>  ****
>>
>> *From:* Quinn Plattel [mailto:qiet72 at gmail.com]
>> *Sent:* Tuesday, January 03, 2012 4:26 AM
>> *To:* Chris Pitzel; gpxe-devel at etherboot.org; gpxe at etherboot.org****
>>
>>
>> *Subject:* Re: [gPXE-devel] Attempting to add support for Broadcom
>> BCM5755M (14e4:1673) ethernet adapter****
>>
>>  ****
>>
>> Hi,
>>
>> I am currently testing his code now, but I currently have compile issues
>> when I try to turn on debugging.
>> Regarding iSCSI - i have a fair bit of experience there.  I have a
>> gpxe/pxe/tftpboot/iscsi implementation currently working with Ubuntu.
>> Check out etherboot.org for my iscsi recipe. :-)
>>
>> Do you happen to know why I get this error:
>>
>> $ make bin/tg3_main.usb DEBUG=tg3_main:3,tg3_hw,tg3_phy
>>   [BUILD] bin/usbdisk.o
>>   [OBJCOPY] bin/usbdisk.bin
>> make: *** No rule to make target `bin/tg3_main.dbg3.o', needed by
>> `bin/blib.a'. Stop.
>>
>>
>> br,
>> Quinn****
>>
>> On Tue, Jan 3, 2012 at 11:01 AM, Chris Pitzel <chris at pitzel.dyndns.org>
>> wrote:****
>>
>> I tried it out Thomas’ tg3 code as well on gPXE (Dell Latitude D630 and
>> D830 laptops).  It seemed to be a little bit more stable during set-up, but
>> not good enough to make iSCSI work, and tftp booting would time-out.****
>>
>>  ****
>>
>> I did hack at the iPXE code, and found that, with the bcm5755m, if I
>> replaced the PCI ID (and a few other variables) of the BCM5787 with the
>> bcm5755m PCI ID, it sort of worked beter.    Enough to tftp and http boot,
>> but no-go on the iSCSI.  I think its a ring buffer issue; the bcm5755m has
>> a unique way of handling ring buffers compared to the other members of the
>> family.  Even in the (linux) kernel tg3 driver, the amount of special
>> handling for the BCM5755 is considerable.****
>>
>>  ****
>>
>> Have you tested iSCSI?  Or just tftp and http?  ****
>>
>>  ****
>>
>> Chris****
>>
>>  ****
>>
>>  ****
>>
>>  ****
>>
>> *From:* Quinn Plattel [mailto:qiet72 at gmail.com]
>> *Sent:* Tuesday, January 03, 2012 3:48 AM
>> *To:* Chris Pitzel
>> *Subject:* Re: [gPXE-devel] Attempting to add support for Broadcom
>> BCM5755M (14e4:1673) ethernet adapter****
>>
>>  ****
>>
>> Hi,
>>
>> I just tested Thomas's newest tg3 code.  It is now working properly with
>> the Dell Laptop that didn't work in the previous version. :-)
>>
>> br,
>> Quinn****
>>
>> On Mon, Jan 2, 2012 at 2:45 PM, Quinn Plattel <qiet72 at gmail.com> wrote:**
>> **
>>
>> Check the list, Thomas Miletich has updated the code for the tg3 driver
>> and is calling for testers.
>>
>> br,
>> Quinn****
>>
>> On Sun, Jan 1, 2012 at 12:42 AM, Chris Pitzel <chris at pitzel.dyndns.org>
>> wrote:****
>>
>> Hello.  I was just reviewing your thread on the gPXE list concerning the
>> Broadcom BCM5755M (presumably from a Dell laptop).****
>>
>>  ****
>>
>> I’ve basically replicated your efforts on the latest iPXE build to the
>> same extent as your attempt.  There appears to be a slight difference in
>> how the BCM5755M is handled in the mainline (Linux) kernel tg3 driver.  *
>> ***
>>
>>  ****
>>
>> Have you pursued any additional line of inquiry?   ****
>>
>>  ****
>>
>> Chris****
>>
>> Saskatchewan, Canada****
>>
>>  ****
>>
>>  ****
>>
>>
>>
>> ****
>>
>>
>> --
>> Best regards/Med venlig hilsen,
>> Quinn Plattel****
>>
>>
>>
>>
>> --
>> Best regards/Med venlig hilsen,
>> Quinn Plattel****
>>
>>
>>
>>
>> --
>> Best regards/Med venlig hilsen,
>> Quinn Plattel****
>>
>>
>>
>>
>> --
>> Best regards/Med venlig hilsen,
>> Quinn Plattel****
>>
>>
>>
>>
>> --
>> Best regards/Med venlig hilsen,
>> Quinn Plattel****
>>
>
>
>
> --
> Best regards/Med venlig hilsen,
> Quinn Plattel
>
>
>
> --
> Best regards/Med venlig hilsen,
> Quinn Plattel
>

-- 
Best regards/Med venlig hilsen,
Quinn Plattel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://etherboot.org/pipermail/gpxe-devel/attachments/20120103/45c2648a/attachment-0001.html>