About the link to wiki in reversible.html

About the link to wiki in reversible.html

[Hui Zhu]

Hi guys,

I found that in http://www.gnu.org/software/gdb/news/reversible.html
See the wiki page here. link to
http://sourceware.org/gdb/wiki/ReversibleDebugging
But this is an old link,  the current wiki link to reverse debug is
http://sourceware.org/gdb/wiki/ReverseDebug

Could someone help me change it or tell me how to change it?
Thanks a lot.

Best Regards,
Hui

porting reversible on arm/mips

[paawan oza]

Hi,

there are pending tasks as mentioned in
http://sourceware.org/gdb/wiki/ReverseDebug

I would like take one of the archs and port reversible on it.

If I correctly understand the task.

->  mips/arm architecture involves, complete writing to arm/mips instruction set and parsing of the same.

-> I assume that there is no basic framework done for mips/arm, such as function call recording, instruction parsing etc, everything has to be written from beginning.

Let me know If my understanding is correct of the current state of code.

-> can we merge the i386-fp patch ?

 regards,
 Oza.


      

Re: About the link to wiki in reversible.html

[Joel Brobecker]

> Could someone help me change it or tell me how to change it?
> Thanks a lot.

Why don't you send us the text that you want to change, with the new text,
and I will fix the WWW. It's a pain to change it, because there are two
copies that need to be kept in sync.

-- 
Joel

Re: porting reversible on arm/mips

[Michael Snyder]

paawan oza wrote:
> Hi,
> 
> there are pending tasks as mentioned in
> http://sourceware.org/gdb/wiki/ReverseDebug
> 
> I would like take one of the archs and port reversible on it.
> 
> If I correctly understand the task.
> 
> ->  mips/arm architecture involves, complete writing to arm/mips instruction set and parsing of the same.
> 
> -> I assume that there is no basic framework done for mips/arm, such as function call recording, instruction parsing etc, everything has to be written from beginning.
> 
> Let me know If my understanding is correct of the current state of code.

I think Hui has already done some work for Mips record/replay.
Maybe you and he could work together on that?

If you want to work on Arm, as far as I know, you would have
that field all to yourself.

Re: porting reversible on arm/mips

[paawan oza]

My target may be arm/mips.
but some of the common hurdles are following.

1) getting virtual machine which has capability gives me arm processor

2) getting the arm linux ISO. 

Can somebody give me some pointers regarding above ?

Regards,.
Oza.


----- Original Message ----
From: Michael Snyder <msnyder@vmware.com>
To: paawan oza <paawan1982@yahoo.com>
Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Fri, December 11, 2009 5:28:40 AM
Subject: Re: porting reversible on arm/mips

paawan oza wrote:
> Hi,
> 
> there are pending tasks as mentioned in
> http://sourceware.org/gdb/wiki/ReverseDebug
> 
> I would like take one of the archs and port reversible on it.
> 
> If I correctly understand the task.
> 
> ->  mips/arm architecture involves, complete writing to arm/mips instruction set and parsing of the same.
> 
> -> I assume that there is no basic framework done for mips/arm, such as function call recording, instruction parsing etc, everything has to be written from beginning.
> 
> Let me know If my understanding is correct of the current state of code.

I think Hui has already done some work for Mips record/replay.
Maybe you and he could work together on that?

If you want to work on Arm, as far as I know, you would have
that field all to yourself.


      

Re: porting reversible on arm/mips

[Hui Zhu]

For the sim, maybe you can try the qemu or skyeye.

Thanks,
Hui

On Fri, Dec 11, 2009 at 10:10, paawan oza <paawan1982@yahoo.com> wrote:
> My target may be arm/mips.
> but some of the common hurdles are following.
>
> 1) getting virtual machine which has capability gives me arm processor
>
> 2) getting the arm linux ISO.
>
> Can somebody give me some pointers regarding above ?
>
> Regards,.
> Oza.
>
>
> ----- Original Message ----
> From: Michael Snyder <msnyder@vmware.com>
> To: paawan oza <paawan1982@yahoo.com>
> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
> Sent: Fri, December 11, 2009 5:28:40 AM
> Subject: Re: porting reversible on arm/mips
>
> paawan oza wrote:
>> Hi,
>>
>> there are pending tasks as mentioned in
>> http://sourceware.org/gdb/wiki/ReverseDebug
>>
>> I would like take one of the archs and port reversible on it.
>>
>> If I correctly understand the task.
>>
>> ->  mips/arm architecture involves, complete writing to arm/mips instruction set and parsing of the same.
>>
>> -> I assume that there is no basic framework done for mips/arm, such as function call recording, instruction parsing etc, everything has to be written from beginning.
>>
>> Let me know If my understanding is correct of the current state of code.
>
> I think Hui has already done some work for Mips record/replay.
> Maybe you and he could work together on that?
>
> If you want to work on Arm, as far as I know, you would have
> that field all to yourself.
>
>
>
>

Re: porting reversible on arm/mips

[Michael Snyder]

paawan oza wrote:
> My target may be arm/mips.
> but some of the common hurdles are following.
> 
> 1) getting virtual machine which has capability gives me arm processor
> 
> 2) getting the arm linux ISO. 
> 
> Can somebody give me some pointers regarding above ?

There's a nice separation in prec between architecture and
OS/ABI.  You could begin with the arm simulator that comes
built-in to gdb, and do the architecture part before
tackling the Linux ABI part.

Re: porting reversible on arm/mips

[Sean Chen]

On Fri, Dec 11, 2009 at 11:11 AM, Michael Snyder <msnyder@vmware.com> wrote:
> paawan oza wrote:
>>
>> My target may be arm/mips.
>> but some of the common hurdles are following.
>>
>> 1) getting virtual machine which has capability gives me arm processor
>>
>> 2) getting the arm linux ISO.
>> Can somebody give me some pointers regarding above ?
>
> There's a nice separation in prec between architecture and
> OS/ABI.  You could begin with the arm simulator that comes
> built-in to gdb, and do the architecture part before
> tackling the Linux ABI part.
>
>

I was interested in the porting on ARM. But later I found that the
performance impact on ARM might damage the usage of process record. In
my experiment, reversible debugging is about 20000x slower, which
might be endurable on the modern computer. However, ARM target is tens
of times (or even more if we consider the memory) slower than PC. So
recording instructions will be very slow, about thousands of
instructions per second.

Please correct me if I made a mistake.

-- 
Best Regards,
Sean Chen

RE: porting reversible on arm/mips

[Jakob Engblom]

> I was interested in the porting on ARM. But later I found that the
> performance impact on ARM might damage the usage of process record. In
> my experiment, reversible debugging is about 20000x slower, which
> might be endurable on the modern computer. However, ARM target is tens
> of times (or even more if we consider the memory) slower than PC. So
> recording instructions will be very slow, about thousands of
> instructions per second.

I just must pitch in and say that it depends on the simulator.

An advantage to using a full simulator is that you simplify the system and no
longer have to care about OS calls: the OS is just part of the context you save
and reverse.  So the overhead actually goes down compared to native prec.  I
think a reversible ARM simulator can be made to run within a factor of ten of
native speed, easily. 


Best regards,

/jakob

_______________________________________________________

Jakob Engblom, PhD, Technical Marketing Manager

Virtutech                   Direct: +46 8 690 07 47   
Drottningholmsvägen 22      Mobile: +46 709 242 646  
11243 Stockholm             Web:    www.virtutech.com 
Sweden
________________________________________________________
  

Re: porting reversible on arm/mips

[paawan oza]

Do I need to cross compile the kernel for arm?
or I can get readily arm linux to be deployed directly on qemu or skyeye ?


----- Original Message ----
From: Hui Zhu <teawater@gmail.com>
To: paawan oza <paawan1982@yahoo.com>
Cc: Michael Snyder <msnyder@vmware.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Fri, December 11, 2009 7:45:11 AM
Subject: Re: porting reversible on arm/mips

For the sim, maybe you can try the qemu or skyeye.

Thanks,
Hui

On Fri, Dec 11, 2009 at 10:10, paawan oza <paawan1982@yahoo.com> wrote:
> My target may be arm/mips.
> but some of the common hurdles are following.
>
> 1) getting virtual machine which has capability gives me arm processor
>
> 2) getting the arm linux ISO.
>
> Can somebody give me some pointers regarding above ?
>
> Regards,.
> Oza.
>
>
> ----- Original Message ----
> From: Michael Snyder <msnyder@vmware.com>
> To: paawan oza <paawan1982@yahoo.com>
> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
> Sent: Fri, December 11, 2009 5:28:40 AM
> Subject: Re: porting reversible on arm/mips
>
> paawan oza wrote:
>> Hi,
>>
>> there are pending tasks as mentioned in
>> http://sourceware.org/gdb/wiki/ReverseDebug
>>
>> I would like take one of the archs and port reversible on it.
>>
>> If I correctly understand the task.
>>
>> ->  mips/arm architecture involves, complete writing to arm/mips instruction set and parsing of the same.
>>
>> -> I assume that there is no basic framework done for mips/arm, such as function call recording, instruction parsing etc, everything has to be written from beginning.
>>
>> Let me know If my understanding is correct of the current state of code.
>
> I think Hui has already done some work for Mips record/replay.
> Maybe you and he could work together on that?
>
> If you want to work on Arm, as far as I know, you would have
> that field all to yourself.
>
>
>
>



      

Re: porting reversible on arm/mips

[paawan oza]

I am not sure whether I got the term 'arm simulator built in gdb"
can you give some more pointers ?

If I understand correctly,

there is no need for simulators, 
compile gdb with some options, on x86, that will simulate arm gdb with some basic arm environment!
and I can try to finish architecture part first.

Can you please elaborate ?

Regards,
Oza.

 


----- Original Message ----
From: Michael Snyder <msnyder@vmware.com>
To: paawan oza <paawan1982@yahoo.com>
Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Fri, December 11, 2009 8:41:27 AM
Subject: Re: porting reversible on arm/mips

paawan oza wrote:
> My target may be arm/mips.
> but some of the common hurdles are following.
> 
> 1) getting virtual machine which has capability gives me arm processor
> 
> 2) getting the arm linux ISO. 
> Can somebody give me some pointers regarding above ?

There's a nice separation in prec between architecture and
OS/ABI.  You could begin with the arm simulator that comes
built-in to gdb, and do the architecture part before
tackling the Linux ABI part.


      

Re: porting reversible on arm/mips

[paawan oza]

I am not sure how arm can be drastically slower than x86 ! considering arm 32 bit.
but at the first point, if in some way if prec arch level stuff and abis related framework are in place,
then optimization on the same may lead us to get faster recording like cache implementation and so on.
But I am not sure of any specific reason why on arm it could be very slower, having the same conf as x86. 

Regards,
Oza.


----- Original Message ----
From: Jakob Engblom <jakob@virtutech.com>
To: Sean Chen <sean.chen1234@gmail.com>; Michael Snyder <msnyder@vmware.com>
Cc: paawan oza <paawan1982@yahoo.com>; Hui Zhu <teawater@gmail.com>; gdb@sourceware.org
Sent: Fri, December 11, 2009 1:45:48 PM
Subject: RE: porting reversible on arm/mips

> I was interested in the porting on ARM. But later I found that the
> performance impact on ARM might damage the usage of process record. In
> my experiment, reversible debugging is about 20000x slower, which
> might be endurable on the modern computer. However, ARM target is tens
> of times (or even more if we consider the memory) slower than PC. So
> recording instructions will be very slow, about thousands of
> instructions per second.

I just must pitch in and say that it depends on the simulator.

An advantage to using a full simulator is that you simplify the system and no
longer have to care about OS calls: the OS is just part of the context you save
and reverse.  So the overhead actually goes down compared to native prec.  I
think a reversible ARM simulator can be made to run within a factor of ten of
native speed, easily. 


Best regards,

/jakob

_______________________________________________________

Jakob Engblom, PhD, Technical Marketing Manager

Virtutech                   Direct: +46 8 690 07 47   
Drottningholmsvägen 22      Mobile: +46 709 242 646  
11243 Stockholm             Web:    www.virtutech.com 
Sweden
________________________________________________________

Re: porting reversible on arm/mips

[Sean Chen]

On Fri, Dec 11, 2009 at 4:15 PM, Jakob Engblom <jakob@virtutech.com> wrote:
>> I was interested in the porting on ARM. But later I found that the
>> performance impact on ARM might damage the usage of process record. In
>> my experiment, reversible debugging is about 20000x slower, which
>> might be endurable on the modern computer. However, ARM target is tens
>> of times (or even more if we consider the memory) slower than PC. So
>> recording instructions will be very slow, about thousands of
>> instructions per second.
>
> I just must pitch in and say that it depends on the simulator.
>
> An advantage to using a full simulator is that you simplify the system and no
> longer have to care about OS calls: the OS is just part of the context you save
> and reverse.  So the overhead actually goes down compared to native prec.  I
> think a reversible ARM simulator can be made to run within a factor of ten of
> native speed, easily.
>
>
> Best regards,
>
> /jakob
>
> _______________________________________________________
>
> Jakob Engblom, PhD, Technical Marketing Manager
>
> Virtutech                   Direct: +46 8 690 07 47
> Drottningholmsvägen 22      Mobile: +46 709 242 646
> 11243 Stockholm             Web:    www.virtutech.com
> Sweden
> ________________________________________________________
>
>
>
>
>
>

The performance of reversible debugging with simulator build-in
support should not be so low. But if Oza means reversible debugging
with process record, performance will be the truth we have to face.

-- 
Best Regards,
Sean Chen

Re: porting reversible on arm/mips

[Sean Chen]

On Fri, Dec 11, 2009 at 9:03 PM, paawan oza <paawan1982@yahoo.com> wrote:
> Do I need to cross compile the kernel for arm?
> or I can get readily arm linux to be deployed directly on qemu or skyeye ?
>
>
> ----- Original Message ----
> From: Hui Zhu <teawater@gmail.com>
> To: paawan oza <paawan1982@yahoo.com>
> Cc: Michael Snyder <msnyder@vmware.com>; "gdb@sourceware.org" <gdb@sourceware.org>
> Sent: Fri, December 11, 2009 7:45:11 AM
> Subject: Re: porting reversible on arm/mips
>
> For the sim, maybe you can try the qemu or skyeye.
>
> Thanks,
> Hui
>
> On Fri, Dec 11, 2009 at 10:10, paawan oza <paawan1982@yahoo.com> wrote:
>> My target may be arm/mips.
>> but some of the common hurdles are following.
>>
>> 1) getting virtual machine which has capability gives me arm processor
>>
>> 2) getting the arm linux ISO.
>>
>> Can somebody give me some pointers regarding above ?
>>
>> Regards,.
>> Oza.
>>
>>
>> ----- Original Message ----
>> From: Michael Snyder <msnyder@vmware.com>
>> To: paawan oza <paawan1982@yahoo.com>
>> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
>> Sent: Fri, December 11, 2009 5:28:40 AM
>> Subject: Re: porting reversible on arm/mips
>>
>> paawan oza wrote:
>>> Hi,
>>>
>>> there are pending tasks as mentioned in
>>> http://sourceware.org/gdb/wiki/ReverseDebug
>>>
>>> I would like take one of the archs and port reversible on it.
>>>
>>> If I correctly understand the task.
>>>
>>> ->  mips/arm architecture involves, complete writing to arm/mips instruction set and parsing of the same.
>>>
>>> -> I assume that there is no basic framework done for mips/arm, such as function call recording, instruction parsing etc, everything has to be written from beginning.
>>>
>>> Let me know If my understanding is correct of the current state of code.
>>
>> I think Hui has already done some work for Mips record/replay.
>> Maybe you and he could work together on that?
>>
>> If you want to work on Arm, as far as I know, you would have
>> that field all to yourself.
>>
>>
>>
>>
>
>
>
>
>

Hui, let me answer these questions for you.

No, there are prebuilt Linux kernel images available. You can use them
with QEMU.

-- 
Best Regards,
Sean Chen

Re: porting reversible on arm/mips

[Sean Chen]

On Fri, Dec 11, 2009 at 9:24 PM, paawan oza <paawan1982@yahoo.com> wrote:
> I am not sure how arm can be drastically slower than x86 ! considering arm 32 bit.
> but at the first point, if in some way if prec arch level stuff and abis related framework are in place,
> then optimization on the same may lead us to get faster recording like cache implementation and so on.
> But I am not sure of any specific reason why on arm it could be very slower, having the same conf as x86.
>
> Regards,
> Oza.
>
>
> ----- Original Message ----
> From: Jakob Engblom <jakob@virtutech.com>
> To: Sean Chen <sean.chen1234@gmail.com>; Michael Snyder <msnyder@vmware.com>
> Cc: paawan oza <paawan1982@yahoo.com>; Hui Zhu <teawater@gmail.com>; gdb@sourceware.org
> Sent: Fri, December 11, 2009 1:45:48 PM
> Subject: RE: porting reversible on arm/mips
>
>> I was interested in the porting on ARM. But later I found that the
>> performance impact on ARM might damage the usage of process record. In
>> my experiment, reversible debugging is about 20000x slower, which
>> might be endurable on the modern computer. However, ARM target is tens
>> of times (or even more if we consider the memory) slower than PC. So
>> recording instructions will be very slow, about thousands of
>> instructions per second.
>
> I just must pitch in and say that it depends on the simulator.
>
> An advantage to using a full simulator is that you simplify the system and no
> longer have to care about OS calls: the OS is just part of the context you save
> and reverse.  So the overhead actually goes down compared to native prec.  I
> think a reversible ARM simulator can be made to run within a factor of ten of
> native speed, easily.
>
>
> Best regards,
>
> /jakob
>
> _______________________________________________________
>
> Jakob Engblom, PhD, Technical Marketing Manager
>
> Virtutech                   Direct: +46 8 690 07 47
> Drottningholmsvägen 22      Mobile: +46 709 242 646
> 11243 Stockholm             Web:    www.virtutech.com
> Sweden
> ________________________________________________________
>
>
>
>

I think ARM does be at least tens of times slower than x86. Image a PC
and a phone, ARM architecture has to sacrifice the performance to gain
the advantage of power and size saving.

Here I assume that process record is more than 20000x slower on both
x86 and ARM. I agree with you and believe this can be improved a lot
in the future.

-- 
Best Regards,
Sean Chen

Re: porting reversible on arm/mips

[Michael Snyder]

paawan oza wrote:
> I am not sure whether I got the term 'arm simulator built in gdb"
> can you give some more pointers ?

If you configure gdb as "arm-elf", gdb will build with
its own linked-in arm simulator which you can use by
means of the gdb command "target simulator".


> If I understand correctly,
> 
> there is no need for simulators, 
> compile gdb with some options, on x86, that will simulate arm gdb with some basic arm environment!
> and I can try to finish architecture part first.

That is correct.   You can build arm-elf-gdb on x86 host.
You will probably want to build arm-elf-gcc and asm and linker,
so that you can compile test programs to simulate.


> ----- Original Message ----
> From: Michael Snyder <msnyder@vmware.com>
> To: paawan oza <paawan1982@yahoo.com>
> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
> Sent: Fri, December 11, 2009 8:41:27 AM
> Subject: Re: porting reversible on arm/mips
> 
> paawan oza wrote:
>> My target may be arm/mips.
>> but some of the common hurdles are following.
>>
>> 1) getting virtual machine which has capability gives me arm processor
>>
>> 2) getting the arm linux ISO. 
>> Can somebody give me some pointers regarding above ?
> 
> There's a nice separation in prec between architecture and
> OS/ABI.  You could begin with the arm simulator that comes
> built-in to gdb, and do the architecture part before
> tackling the Linux ABI part.
> 
> 
>       

Re: porting reversible on arm/mips

[Ramana Radhakrishnan]

> That is correct.   You can build arm-elf-gdb on x86 host.
> You will probably want to build arm-elf-gcc and asm and linker,
> so that you can compile test programs to simulate.


I would strongly encourage anyone building tools for ARM and doing
active development to be using arm-eabi target rather than arm-elf
today . The arm-elf / arm-linux targets are more or less in
maintenance only mode at the minute while arm-eabi and
arm-linux-gnueabi enjoys all the attention today.

cheers
Ramana

>
>
>> ----- Original Message ----
>> From: Michael Snyder <msnyder@vmware.com>
>> To: paawan oza <paawan1982@yahoo.com>
>> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org"
>> <gdb@sourceware.org>
>> Sent: Fri, December 11, 2009 8:41:27 AM
>> Subject: Re: porting reversible on arm/mips
>>
>> paawan oza wrote:
>>>
>>> My target may be arm/mips.
>>> but some of the common hurdles are following.
>>>
>>> 1) getting virtual machine which has capability gives me arm processor
>>>
>>> 2) getting the arm linux ISO. Can somebody give me some pointers
>>> regarding above ?
>>
>> There's a nice separation in prec between architecture and
>> OS/ABI.  You could begin with the arm simulator that comes
>> built-in to gdb, and do the architecture part before
>> tackling the Linux ABI part.
>>
>>
>>
>
>

Re: porting reversible on arm/mips

[Michael Snyder]

Ramana Radhakrishnan wrote:
>> That is correct.   You can build arm-elf-gdb on x86 host.
>> You will probably want to build arm-elf-gcc and asm and linker,
>> so that you can compile test programs to simulate.
> 
> 
> I would strongly encourage anyone building tools for ARM and doing
> active development to be using arm-eabi target rather than arm-elf
> today . The arm-elf / arm-linux targets are more or less in
> maintenance only mode at the minute while arm-eabi and
> arm-linux-gnueabi enjoys all the attention today.

Fair enough.

Ramana, do you have a favorite online guide or "how-to"
for someone who wants to start building an arm cross tool chain?

Re: porting reversible on arm/mips

[Ramana Radhakrishnan]

On Fri, Dec 11, 2009 at 7:12 PM, Michael Snyder <msnyder@vmware.com> wrote:
> Fair enough.
>
> Ramana, do you have a favorite online guide or "how-to"
> for someone who wants to start building an arm cross tool chain?

Depends on what you want to do here -

If you are looking to build cross bare-metal toolchains - I prefer to
use the combined build method as here
http://gcc.gnu.org/wiki/Building_Cross_Toolchains_with_gcc with the
--target=arm-eabi option. I use that for most of my development
activity and that works just fine.

I haven't been following the gdb reverse work as closely as I should
but if your work doesn't involve tweaking the kernel and is only in
gdb userland, then there are some ARM devel boards running GNU/Linux
available on the GCC compile farm for use by open source developers.
You could contact Laurent Guerby for access to these boards.
http://gcc.gnu.org/wiki/CompileFarm

If you want to build ARM Linux cross-tools using eglibc, then using a
script based on instructions as in the eglibc repository are what I
use and have set up successfully. Dan Kegel's crosstool appears to be
very good for building with older configurations but I'm not sure how
it copes with the new ports infrastructure in (e)glibc.

cheers
Ramana






>
>

Re: porting reversible on arm/mips

[paawan oza]

Hi All,

In conclusion we have following approaches to go ahead with.

1) have qemu or skyeye on windows and get the arm kernel on to it.

2) Using arm-elf-gdb, and use 'target simulator' command, and cross compiling test programs for arm.

1st gives me complete environment.
2nd may be a part of it.

any thing is ok as long as that gives smooth start.
I will try to setup qemu first and see.

Thanks for your help.
Regards,
Oza.



----- Original Message ----
From: Michael Snyder <msnyder@vmware.com>
To: paawan oza <paawan1982@yahoo.com>
Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Sat, December 12, 2009 12:19:12 AM
Subject: Re: porting reversible on arm/mips

paawan oza wrote:
> I am not sure whether I got the term 'arm simulator built in gdb"
> can you give some more pointers ?

If you configure gdb as "arm-elf", gdb will build with
its own linked-in arm simulator which you can use by
means of the gdb command "target simulator".


> If I understand correctly,
> 
> there is no need for simulators, compile gdb with some options, on x86, that will simulate arm gdb with some basic arm environment!
> and I can try to finish architecture part first.

That is correct.   You can build arm-elf-gdb on x86 host.
You will probably want to build arm-elf-gcc and asm and linker,
so that you can compile test programs to simulate.


> ----- Original Message ----
> From: Michael Snyder <msnyder@vmware.com>
> To: paawan oza <paawan1982@yahoo.com>
> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
> Sent: Fri, December 11, 2009 8:41:27 AM
> Subject: Re: porting reversible on arm/mips
> 
> paawan oza wrote:
>> My target may be arm/mips.
>> but some of the common hurdles are following.
>> 
>> 1) getting virtual machine which has capability gives me arm processor
>> 
>> 2) getting the arm linux ISO. Can somebody give me some pointers regarding above ?
> 
> There's a nice separation in prec between architecture and
> OS/ABI.  You could begin with the arm simulator that comes
> built-in to gdb, and do the architecture part before
> tackling the Linux ABI part.
> 
> 
>      


      

Re: porting reversible on arm/mips

[paawan oza]

Hi Hui and Michael,

I have been trying to setup qemu, on suse10 sp1/2 host... I did not get correct source till now.
I got some binaries for suse10.2.    (kqemu.ko was not compatible)
And also arm kernel images, I did not get the one.
setting up qemu on 2.6.16 suse10.2 with arm kernel seems to be not so friendly.
if any pointer links, will be helpful to me.

another option what Micahel suggested...
"There's a nice separation in prec between architecture and OS/ABI.  You could begin with the arm simulator that comes
built-in to gdb, and do the architecture part before tackling the Linux ABI part."
above also seems to be good too. for the time being I just concentrate on arch part and forget Linux ABI for the time being..

please provide your inputs.

Regards,
Oza.



----- Original Message ----
From: paawan oza <paawan1982@yahoo.com>
To: Michael Snyder <msnyder@vmware.com>
Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Sat, December 12, 2009 8:58:11 AM
Subject: Re: porting reversible on arm/mips

Hi All,

In conclusion we have following approaches to go ahead with.

1) have qemu or skyeye on windows and get the arm kernel on to it.

2) Using arm-elf-gdb, and use 'target simulator' command, and cross compiling test programs for arm.

1st gives me complete environment.
2nd may be a part of it.

any thing is ok as long as that gives smooth start.
I will try to setup qemu first and see.

Thanks for your help.
Regards,
Oza.



----- Original Message ----
From: Michael Snyder <msnyder@vmware.com>
To: paawan oza <paawan1982@yahoo.com>
Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Sat, December 12, 2009 12:19:12 AM
Subject: Re: porting reversible on arm/mips

paawan oza wrote:
> I am not sure whether I got the term 'arm simulator built in gdb"
> can you give some more pointers ?

If you configure gdb as "arm-elf", gdb will build with
its own linked-in arm simulator which you can use by
means of the gdb command "target simulator".


> If I understand correctly,
> 
> there is no need for simulators, compile gdb with some options, on x86, that will simulate arm gdb with some basic arm environment!
> and I can try to finish architecture part first.

That is correct.   You can build arm-elf-gdb on x86 host.
You will probably want to build arm-elf-gcc and asm and linker,
so that you can compile test programs to simulate.


> ----- Original Message ----
> From: Michael Snyder <msnyder@vmware.com>
> To: paawan oza <paawan1982@yahoo.com>
> Cc: Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
> Sent: Fri, December 11, 2009 8:41:27 AM
> Subject: Re: porting reversible on arm/mips
> 
> paawan oza wrote:
>> My target may be arm/mips.
>> but some of the common hurdles are following.
>> 
>> 1) getting virtual machine which has capability gives me arm processor
>> 
>> 2) getting the arm linux ISO. Can somebody give me some pointers regarding above ?
> 
> There's a nice separation in prec between architecture and
> OS/ABI.  You could begin with the arm simulator that comes
> built-in to gdb, and do the architecture part before
> tackling the Linux ABI part.
> 
> 
>      


      

Re: porting reversible on arm/mips

[Michael Snyder]

paawan oza wrote:

> another option what Micahel suggested...
> "There's a nice separation in prec between architecture and OS/ABI.  You could begin with the arm simulator that comes
> built-in to gdb, and do the architecture part before tackling the Linux ABI part."
> above also seems to be good too. for the time being I just concentrate on arch part and forget Linux ABI for the time being..

Seems to me you should be able to get quite far using this approach.

Should be able to simulate and reverse-execute any piece of code that
does not make system calls (directly or indirectly).  Should be able
to cover most of the user instruction set, especially if you are
willing to write tests in assembler.

Re: porting reversible on arm/mips

[paawan oza]

Hi,

-> datasheet which requires registration to get from ARM 

-> need to contact qemu guys to get arm target specially.

I am not sure, how to go about above two things, assuming that there are no alternatives.

PS: both are very important points as having right spec and target, will end up doing right coding.

Regards,
Oza.



----- Original Message ----
From: Joseph S. Myers <joseph@codesourcery.com>
To: Sean Chen <sean.chen1234@gmail.com>
Cc: paawan oza <paawan1982@yahoo.com>; Michael Snyder <msnyder@vmware.com>; Hui Zhu <teawater@gmail.com>; "gdb@sourceware.org" <gdb@sourceware.org>
Sent: Mon, March 22, 2010 6:01:40 PM
Subject: Re: porting reversible on arm/mips

On Mon, 22 Mar 2010, Sean Chen wrote:

> > 1) which Arm family/data sheet need to be refereed ?
> DDI0100I_ARM_Architecture_Reference_Manual and
> DDI0308C_thumb2_supplement are the latest basic documents. To

Those are long obsolete.  You want DDI0406B (for v7-A/v7-R; DDI0403 for 
v7-M), which requires registration to get from ARM.

-- 
Joseph S. Myers
joseph@codesourcery.com



      

Re: porting reversible on arm/mips

[Joseph S. Myers]

On Mon, 22 Mar 2010, Sean Chen wrote:

> > 1) which Arm family/data sheet need to be refereed ?
> DDI0100I_ARM_Architecture_Reference_Manual and
> DDI0308C_thumb2_supplement are the latest basic documents. To

Those are long obsolete.  You want DDI0406B (for v7-A/v7-R; DDI0403 for 
v7-M), which requires registration to get from ARM.

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: porting reversible on arm/mips

[Sean Chen]

> 1) which Arm family/data sheet need to be refereed ?
DDI0100I_ARM_Architecture_Reference_Manual and
DDI0308C_thumb2_supplement are the latest basic documents. To
implement coprocessor instructions which is a plus such as VPF, NEON
and WMMX, you might need the coresponding specs.
> 2) I am planning to use 'target simulator' command, and cross compiling test programs for arm.
> so with that simulator I can finish everything except Linux ABI
Arm-elf is a choice. However, I am not sure whether it supports the
latest ARMv7 and Thumb2 instructions. I think a real ARM target or
QEMU is necessary, otherwise, it's very inconvenient for you to test
and implement the Linux function call handler.
> 3) and at the last step I will be trying to get arm kernel and finish linux ABI part.
You will need  real ARM target or QEMU absolutely.
> PS : some way it is getting difficult to use qemu or other virtual emulation and fit it on to x86 emulating ARM.
QEMU is a good choice if you don't have any ARM target. You might need
to contact QEMU guys.

-- 
Best Regards,
Sean Chen

porting reversible on arm/mips

[paawan oza]

Hi,

I will be trying to port reversible on to arm.
would you please help me with some of the things.

1) which Arm family/data sheet need to be refereed ?

2) I am planning to use 'target simulator' command, and cross compiling test programs for arm.
so with that simulator I can finish everything except Linux ABI

3) and at the last step I will be trying to get arm kernel and finish linux ABI part.

I hope what way I am thinking will be enough to achieve porting.
please provide your inputs.

PS : some way it is getting difficult to use qemu or other virtual emulation and fit it on to x86 emulating ARM. 

Regards,
Oza.