Re: openbmc qemu vs Cedric qemu is different

Re: openbmc qemu vs Cedric qemu is different

[Joel Stanley]

Hello Chris,

On Thu, 2016-07-14 at 13:27 -0500, Chris Austen wrote:
> Hi Joel, 
> last night you indicated that our openbmc/qemu had enough for
> networking the palmetto-bmc. I don't think that is 100% true. I
> believe Cedric's repo has some additional changes that should be
> integrated back in to openbmc/qemu. Hopefully you two can figure out
> what commits are needed to get us back to one working repo. 

I don't recall claming that. I think you're confusing me with Andrew.

I'll let Andrew clarify the situation.

Cheers,

Joel

> 
> 
> causten@causten-VirtualBox:~/gitstuff/obmcqemu/build/arm-softmmu
> $ ./qemu-system-arm -m 256 -M palmetto-bmc -nographic \
> > -drive file=~/palmetto-qemu/flash-palmetto,format=raw,if=mtd \
> > -net nic,macaddr=C0:FF:EE:00:00:02,model=ftgmac100 \
> > -net bridge,id=net0,helper=/usr/lib/qemu-bridge-helper,br=virbr0
> qemu-system-arm: Unsupported NIC model: ftgmac100 
> 
> 
> causten@causten-VirtualBox:~/gitstuff/obmcqemu/build/arm-softmmu
> ./qemu-system-arm -m 256 -M palmetto-bmc -nographic -kernel
> ~/palmetto-qemu/cuImage-palmetto.bin -initrd ~/palmetto-qemu/obmc-
> phosphor-image-palmetto.cpio.gz
> ...
> ftgmac100: NCSI interface down
> 
> Phosphor OpenBMC (Phosphor OpenBMC Project Reference Distro) 0.1.0
> palmetto ttyS4
> 
> palmetto login: root
> Password: 
> root@palmetto:~# ifconfig
> eth0 Link encap:Ethernet HWaddr DA:74:6E:BE:E9:CD 
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:0 errors:0 dropped:0 overruns:0 frame:0
> TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000 
> RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
> 
> lo Link encap:Local Loopback 
> inet addr:127.0.0.1 Mask:255.0.0.0
> UP LOOPBACK RUNNING MTU:65536 Metric:1
> RX packets:160 errors:0 dropped:0 overruns:0 frame:0
> TX packets:160 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1 
> RX bytes:12160 (11.8 KiB) TX bytes:12160 (11.8 KiB)
> 
> 
> 
> 
> 
> 
> 
> Chris Austen
> POWER Systems Enablement Manager 
> (512) 286-5184 (T/L: 363-5184)

Re: openbmc qemu vs Cedric qemu is different

[Cédric Le Goater]

On 07/15/2016 03:42 AM, Joel Stanley wrote:
> Hello Chris,
> 
> On Thu, 2016-07-14 at 13:27 -0500, Chris Austen wrote:
>> Hi Joel, 
>> last night you indicated that our openbmc/qemu had enough for
>> networking the palmetto-bmc. I don't think that is 100% true. I
>> believe Cedric's repo has some additional changes that should be
>> integrated back in to openbmc/qemu. Hopefully you two can figure out
>> what commits are needed to get us back to one working repo. 

Here is a quick summary of the differences, my branch compared to 
openbmc/qemu:


* qemu base 

  closer to mainline.

* smc

  adds some minor code clean ups (merged in mainline now)

* vic

  misses some nice-to-have properties. 

* watchdog

  adds a fix for uboot (linux still needs fixing)
  
* timer

  in sync
  
* scu

  adds minor fixes

* ast2500

  adds more support (can boot uboot with the vbar hack)

* ftgmac100

  adds a new 'aspeed' property to change the end of ring buffer bit 
  in the network adpater model for aspeed socs. This is needed for 
  the linux 4.6 ftgmac100 network driver which uses the 'correct'
  bits. the ast2500 flash images will be fine but there is a problem 
  for older ast2400 flash images which have a 4.4 kernel using the 
  old bits definition. 

  We could need to make the 'aspeed' property configurable from the 
  qemu command line to support these. to be discussed 

So, qemu should behave the same in both branches. 


Support for ast2400/linux 4.6 is broken in my branch and a quick 
would be :

	https://github.com/legoater/qemu/commit/e14da799c81447f9a045999de454cf4490905bf4

That will drop support for the images with a 4.4 kernel. Can we ?


qemu-2.7-rc0 will have some more of our code merged. we will update
the openbmc/qemu branch at that time. 

I am mostly disconnected next week, so expect latency in the answers :)

Cheers,

C.

Re: openbmc qemu vs Cedric qemu is different

[Andrew Jeffery]

[-- Attachment #1: Type: text/plain, Size: 1847 bytes --]

On Fri, 2016-07-15 at 11:12 +0930, Joel Stanley wrote:
> Hello Chris,
> 
> On Thu, 2016-07-14 at 13:27 -0500, Chris Austen wrote:
> > 
> > Hi Joel, 
> > last night you indicated that our openbmc/qemu had enough for
> > networking the palmetto-bmc. I don't think that is 100% true. I
> > believe Cedric's repo has some additional changes that should be
> > integrated back in to openbmc/qemu. Hopefully you two can figure out
> > what commits are needed to get us back to one working repo. 
> I don't recall claming that. I think you're confusing me with Andrew.

Sounds about right.

> 
> I'll let Andrew clarify the situation.
> 
> Cheers,
> 
> Joel
> 
> > 
> > 
> > 
> > causten@causten-VirtualBox:~/gitstuff/obmcqemu/build/arm-softmmu
> > $ ./qemu-system-arm -m 256 -M palmetto-bmc -nographic \
> > > 
> > > -drive file=~/palmetto-qemu/flash-palmetto,format=raw,if=mtd \
> > > -net nic,macaddr=C0:FF:EE:00:00:02,model=ftgmac100 \
> > > -net bridge,id=net0,helper=/usr/lib/qemu-bridge-helper,br=virbr0
> > qemu-system-arm: Unsupported NIC model: ftgmac100 

Cedric's outlined the differences between his github repo and
openbmc/qemu, but I wouldn't have thought the differences would account
for this. Have you confirmed you're running the right qemu-system-arm
binary?

I believe I tested the networking before pushing Cedric's changes to
openbmc/qemu around 3 weeks ago, but I could be mistaken. Having said
that, as Cedric outlined, there may be some broken behaviour due to
fixes for ASPEED's movement of the EDO[RT]R bits in the MAC, so we
should resolve any issues as soon as possible.

Unfortunately I haven't kept pace with Cedric's work as I've been
distracted with other problems, but I intend on making up what ground I
can after I send my kernel patches upstream.

Cheers,

Andrew

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: openbmc qemu vs Cedric qemu is different

[Chris Austen]

[-- Attachment #1.1: Type: text/plain, Size: 2447 bytes --]


Sounds like you have a workable plan.  Thank you.  The test team will run
off of https://github.com/legoater/qemu (branch:aspeed) until the
openbmc/qemu gets the needed updates.


Chris Austen
POWER Systems Enablement Manager
(512) 286-5184 (T/L: 363-5184)



From:	Andrew Jeffery <andrew@aj.id.au>
To:	Joel Stanley <joel@jms.id.au>, Chris Austen/Austin/IBM@IBMUS,
            openbmc@lists.ozlabs.org
Cc:	Cedric Le Goater <clegoater@fr.ibm.com>
Date:	07/15/2016 05:20 AM
Subject:	Re: openbmc qemu vs Cedric qemu is different



On Fri, 2016-07-15 at 11:12 +0930, Joel Stanley wrote:
> Hello Chris,
>
> On Thu, 2016-07-14 at 13:27 -0500, Chris Austen wrote:
> >
> > Hi Joel,
> > last night you indicated that our openbmc/qemu had enough for
> > networking the palmetto-bmc. I don't think that is 100% true. I
> > believe Cedric's repo has some additional changes that should be
> > integrated back in to openbmc/qemu. Hopefully you two can figure out
> > what commits are needed to get us back to one working repo.
> I don't recall claming that. I think you're confusing me with Andrew.

Sounds about right.

>
> I'll let Andrew clarify the situation.
>
> Cheers,
>
> Joel
>
> >
> >
> >
> > causten@causten-VirtualBox:~/gitstuff/obmcqemu/build/arm-softmmu
> > $ ./qemu-system-arm -m 256 -M palmetto-bmc -nographic \
> > >
> > > -drive file=~/palmetto-qemu/flash-palmetto,format=raw,if=mtd \
> > > -net nic,macaddr=C0:FF:EE:00:00:02,model=ftgmac100 \
> > > -net bridge,id=net0,helper=/usr/lib/qemu-bridge-helper,br=virbr0
> > qemu-system-arm: Unsupported NIC model: ftgmac100

Cedric's outlined the differences between his github repo and
openbmc/qemu, but I wouldn't have thought the differences would account
for this. Have you confirmed you're running the right qemu-system-arm
binary?

I believe I tested the networking before pushing Cedric's changes to
openbmc/qemu around 3 weeks ago, but I could be mistaken. Having said
that, as Cedric outlined, there may be some broken behaviour due to
fixes for ASPEED's movement of the EDO[RT]R bits in the MAC, so we
should resolve any issues as soon as possible.

Unfortunately I haven't kept pace with Cedric's work as I've been
distracted with other problems, but I intend on making up what ground I
can after I send my kernel patches upstream.

Cheers,

Andrew[attachment "signature.asc" deleted by Chris Austen/Austin/IBM]

[-- Attachment #1.2: Type: text/html, Size: 3632 bytes --]

[-- Attachment #2: graycol.gif --]
[-- Type: image/gif, Size: 105 bytes --]

Re: openbmc qemu vs Cedric qemu is different

[Cédric Le Goater]

On 07/15/2016 05:58 PM, Chris Austen wrote:
> Sounds like you have a workable plan. Thank you. The test team will run 
> off of https://github.com/legoater/qemu  (branch:aspeed) until the 
> openbmc/qemu gets the needed updates.


FYI, I just updated this branch. Changes are :

  - rebase on v2.7.0-rc0
  - end of n/w support for linux 4.4. the fgtmac100 now uses 
    the aspeed end of ring buffer bits to be in sync with 
    linux 4.6 
  - rework of ast2500 support to remove duplicated code. I should
    send to mainline shortly for review.

I gave them a quick test with these images:

    https://openpower.xyz/job/openbmc-build/distro=ubuntu,target=palmetto/lastSuccessfulBuild/artifact/images/palmetto/flash-palmetto
    https://openpower.xyz/job/openbmc-build/distro=ubuntu,target=evb-ast2500/lastSuccessfulBuild/artifact/images/evb-ast2500/flash-evb-ast2500

the ast2500 still needs a little hack in uboot which Joel included in 
v2016.07-aspeed-openbmc :
	
	https://github.com/shenki/u-boot/commit/a0ca4ecbeae4ae9632ebc67bf5318dd2ea09c94f

I need to dig that one to add VBAR support the ast2500 qemu guest. 

Cheers,

C.


> Chris Austen
> POWER Systems Enablement Manager
> (512) 286-5184 (T/L: 363-5184)
> 
> Inactive hide details for Andrew Jeffery ---07/15/2016 05:20:17 AM---On Fri, 2016-07-15 at 11:12 +0930, Joel Stanley wrote: > HAndrew Jeffery ---07/15/2016 05:20:17 AM---On Fri, 2016-07-15 at 11:12 +0930, Joel Stanley wrote: > Hello Chris,
> 
> From: Andrew Jeffery <andrew@aj.id.au>
> To: Joel Stanley <joel@jms.id.au>, Chris Austen/Austin/IBM@IBMUS, openbmc@lists.ozlabs.org
> Cc: Cedric Le Goater <clegoater@fr.ibm.com>
> Date: 07/15/2016 05:20 AM
> Subject: Re: openbmc qemu vs Cedric qemu is different
> 
> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
> 
> 
> 
> On Fri, 2016-07-15 at 11:12 +0930, Joel Stanley wrote:
>> Hello Chris,
>>
>> On Thu, 2016-07-14 at 13:27 -0500, Chris Austen wrote:
>> >
>> > Hi Joel, 
>> > last night you indicated that our openbmc/qemu had enough for
>> > networking the palmetto-bmc. I don't think that is 100% true. I
>> > believe Cedric's repo has some additional changes that should be
>> > integrated back in to openbmc/qemu. Hopefully you two can figure out
>> > what commits are needed to get us back to one working repo. 
>> I don't recall claming that. I think you're confusing me with Andrew.
> 
> Sounds about right.
> 
>>
>> I'll let Andrew clarify the situation.
>>
>> Cheers,
>>
>> Joel
>>
>> >
>> >
>> >
>> > causten@causten-VirtualBox:~/gitstuff/obmcqemu/build/arm-softmmu
>> > $ ./qemu-system-arm -m 256 -M palmetto-bmc -nographic \
>> > >
>> > > -drive file=~/palmetto-qemu/flash-palmetto,format=raw,if=mtd \
>> > > -net nic,macaddr=C0:FF:EE:00:00:02,model=ftgmac100 \
>> > > -net bridge,id=net0,helper=/usr/lib/qemu-bridge-helper,br=virbr0
>> > qemu-system-arm: Unsupported NIC model: ftgmac100 
> 
> Cedric's outlined the differences between his github repo and
> openbmc/qemu, but I wouldn't have thought the differences would account
> for this. Have you confirmed you're running the right qemu-system-arm
> binary?
> 
> I believe I tested the networking before pushing Cedric's changes to
> openbmc/qemu around 3 weeks ago, but I could be mistaken. Having said
> that, as Cedric outlined, there may be some broken behaviour due to
> fixes for ASPEED's movement of the EDO[RT]R bits in the MAC, so we
> should resolve any issues as soon as possible.
> 
> Unfortunately I haven't kept pace with Cedric's work as I've been
> distracted with other problems, but I intend on making up what ground I
> can after I send my kernel patches upstream.
> 
> Cheers,
> 
> Andrew[attachment "signature.asc" deleted by Chris Austen/Austin/IBM]
> 

Re: openbmc qemu vs Cedric qemu is different

[Andrew Jeffery]

[-- Attachment #1: Type: text/plain, Size: 951 bytes --]

On Wed, 2016-07-27 at 12:26 +0200, Cédric Le Goater wrote:
> On 07/15/2016 05:58 PM, Chris Austen wrote:
> > 
> > Sounds like you have a workable plan. Thank you. The test team will run 
> > off of https://github.com/legoater/qemu  (branch:aspeed) until the 
> > openbmc/qemu gets the needed updates.
> 
> FYI, I just updated this branch. Changes are :
> 
>   - rebase on v2.7.0-rc0
>   - end of n/w support for linux 4.4. the fgtmac100 now uses 
>     the aspeed end of ring buffer bits to be in sync with 
>     linux 4.6 
>   - rework of ast2500 support to remove duplicated code. I should
>     send to mainline shortly for review.

I've just merged Cedric's work above into openbmc/qemu master. Tested
with a Palmetto flash image and kernel/rootfs from Witherspoon,
everything looks good.

Chris: Any documentation you have can now point there. Let me know if
there are any issues.

Cheers,

Andrew

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: openbmc qemu vs Cedric qemu is different

[Cédric Le Goater]

On 07/27/2016 02:30 PM, Andrew Jeffery wrote:
> On Wed, 2016-07-27 at 12:26 +0200, Cédric Le Goater wrote:
>> On 07/15/2016 05:58 PM, Chris Austen wrote:
>>>
>>> Sounds like you have a workable plan. Thank you. The test team will run 
>>> off of https://github.com/legoater/qemu  (branch:aspeed) until the 
>>> openbmc/qemu gets the needed updates.
>>
>> FYI, I just updated this branch. Changes are :
>>
>>   - rebase on v2.7.0-rc0
>>   - end of n/w support for linux 4.4. the fgtmac100 now uses 
>>     the aspeed end of ring buffer bits to be in sync with 
>>     linux 4.6 
>>   - rework of ast2500 support to remove duplicated code. I should
>>     send to mainline shortly for review.
> 
> I've just merged Cedric's work above into openbmc/qemu master. Tested
> with a Palmetto flash image and kernel/rootfs from Witherspoon,
> everything looks good.

Thanks Andrew.

C.

> Chris: Any documentation you have can now point there. Let me know if
> there are any issues.
> 
> Cheers,
> 
> Andrew
> 

Re: openbmc qemu vs Cedric qemu is different

[Chris Austen]

[-- Attachment #1.1: Type: text/plain, Size: 1510 bytes --]


Tested on my servers, great job merging back in to openbmc/qemu.  I've
updated my cheatsheet commit to reflect QEMU operations.


Chris Austen
POWER Systems Enablement Manager
(512) 286-5184 (T/L: 363-5184)



From:	Cédric Le Goater <clg@kaod.org>
To:	Andrew Jeffery <andrew@aj.id.au>, Chris Austen/Austin/IBM@IBMUS
Cc:	Joel Stanley <joel@jms.id.au>, openbmc@lists.ozlabs.org
Date:	07/27/2016 07:37 AM
Subject:	Re: openbmc qemu vs Cedric qemu is different



On 07/27/2016 02:30 PM, Andrew Jeffery wrote:
> On Wed, 2016-07-27 at 12:26 +0200, Cédric Le Goater wrote:
>> On 07/15/2016 05:58 PM, Chris Austen wrote:
>>>
>>> Sounds like you have a workable plan. Thank you. The test team will run

>>> off of https://github.com/legoater/qemu  (branch:aspeed) until the
>>> openbmc/qemu gets the needed updates.
>>
>> FYI, I just updated this branch. Changes are :
>>
>>   - rebase on v2.7.0-rc0
>>   - end of n/w support for linux 4.4. the fgtmac100 now uses
>>     the aspeed end of ring buffer bits to be in sync with
>>     linux 4.6
>>   - rework of ast2500 support to remove duplicated code. I should
>>     send to mainline shortly for review.
>
> I've just merged Cedric's work above into openbmc/qemu master. Tested
> with a Palmetto flash image and kernel/rootfs from Witherspoon,
> everything looks good.

Thanks Andrew.

C.

> Chris: Any documentation you have can now point there. Let me know if
> there are any issues.
>
> Cheers,
>
> Andrew
>



[-- Attachment #1.2: Type: text/html, Size: 2741 bytes --]

[-- Attachment #2: graycol.gif --]
[-- Type: image/gif, Size: 105 bytes --]