coremu-list-devel Mailing List for coremu-mailing-list (Page 3)
Brought to you by:
cyfdecyf
Menu
▾
▴
coremu-list-devel — coremu devel mailing list
You can subscribe to this list here.
| 2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(4) |
Sep
(1) |
Oct
(21) |
Nov
(2) |
Dec
(1) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2011 |
Jan
(4) |
Feb
(4) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(6) |
Sep
(2) |
Oct
(2) |
Nov
(1) |
Dec
|
| 2012 |
Jan
(8) |
Feb
|
Mar
|
Apr
(3) |
May
(7) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2013 |
Jan
|
Feb
(10) |
Mar
(7) |
Apr
(2) |
May
|
Jun
|
Jul
|
Aug
(2) |
Sep
(3) |
Oct
|
Nov
|
Dec
|
| 2014 |
Jan
|
Feb
(2) |
Mar
|
Apr
(1) |
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Tommy <th...@gm...> - 2011-09-19 16:58:58
|
Hi, Can I use the checkpointing feature in Coremu? I have tried to use my QEMU image and it booted on Coremu. But I couldn't use the "savevm" command to save the states. Do I have to create a new image? Any suggestion? Thanks. Tommy |
|
From: Chen Y. <cyf...@gm...> - 2011-08-30 13:53:37
|
I suggest you do a git rebase to the 0.15.0 version of QEMU. We did this a few times when we want to move COREMU onto the latest version of QEMU. Of course you could do a merge with the 0.15.0 QEMU instead, and try to solve the conflicts. But the lengthy rebase process will make you know more about what has changed in QEMU and get more chance to identify which might break COREMU. Identifying changes that will break COREMU during the rebase process is better than debugging after merging/rebasing the code. Note the rebase process can not make you identify all changes that may break COREMU, so the process require some care. We have several deadlines on other projects so we don't have time to move COREMU to the latest version QEMU now. On Mon, Aug 29, 2011 at 2:05 PM, Diviya Jain <div...@gm...> wrote: > Thanks for the info. Another question that I have regard porting is > that latest version of Qemu i.e v 0.15.0 has many powerpc related > changes. There has been significant development for ppc platform in > latest version as compared to the patched Qemu currently being used as > part of coremu. What is the best way to make latest qemu v 0.15.0 work > with coremu? > > > > On 8/26/11, zhaoguo wang <tig...@gm...> wrote: >> Diviya: >> I'm sorry there is no porting guide available. However we are >> very glad to help >> you for the porting work.To port QEMU for a new architecture, please >> consider >> the following aspects: >> >> 1. atomic instruction emulation. >> see target-arm/translate.c and target-arm/cm-atomic.c/h >> >> 2. devices emulation. >> You need to modify the interrupt controller emulation by using >> interfaces >> provided by coremu (e.g.coremu_send_intr). see >> target-arm/cm-target-intr.c >> and hw/arm_pic.c. >> >> You also need to make emulated devices can be concurrently accessed. >> (e.g. Two emulated cpu threads may access a device at the same time.) >> For arm architecture, we use a global lock to protect the emulated >> device, it's >> simple but no efficient. For x86_64 architecture, we parallel every >> device >> emulation which is very efficient but the development time is a little >> long. >> >> We use #if defined(CONFIG_COREMU) to encapsulate all the changes in >> QEMU. >> For the code shared with other architecture, we use >> #if defined(CONFIG_COREMU) && defined(TARGET_ARM) to encapsulate >> the changes specially for arm architecture. >> >> >> 2011/8/26 Diviya Jain <div...@gm...>: >>> Hi, >>> We need to port Qemu for powerpc to Coremu. Is there any porting guide >>> available? As per your paper porting Coremu for ARM was very easy and >>> required only 150 lines of code change to Qemu/Coremu. >>> Could you please help us understand what all steps are required to port >>> Coremu for powerpc. Is there any document for the same? Or perhaps if you >>> could tell us changes done for ARM, so that we can do similar changes for >>> powerpc to Qemu/Coremu source code. >>> Another way might be grepping through qemu/coremu source code based on >>> some >>> keyword. Is there some #ifdef etc in qemu/coremu which encapsulate all the >>> changes done to make it work for ARM platform? We can do similar for >>> powerpc. >>> Thanks, >>> diviya >>> ------------------------------------------------------------------------------ >>> EMC VNX: the world's simplest storage, starting under $10K >>> The only unified storage solution that offers unified management >>> Up to 160% more powerful than alternatives and 25% more efficient. >>> Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev >>> _______________________________________________ >>> Coremu-list-devel mailing list >>> Cor...@li... >>> https://lists.sourceforge.net/lists/listinfo/coremu-list-devel >>> >>> >> >> >> >> -- >> Zhaoguo Wang, Parallel Processing Institute, Fudan University >> >> Address: Room 320, Software Building, 825 Zhangheng Road, Shanghai, China >> >> tig...@gm... >> http://ppi.fudan.edu.cn/zhaoguo_wang >> > > ------------------------------------------------------------------------------ > EMC VNX: the world's simplest storage, starting under $10K > The only unified storage solution that offers unified management > Up to 160% more powerful than alternatives and 25% more efficient. > Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > -- Best regards, Chen Yufei |
|
From: Diviya J. <div...@gm...> - 2011-08-29 06:05:58
|
Thanks for the info. Another question that I have regard porting is that latest version of Qemu i.e v 0.15.0 has many powerpc related changes. There has been significant development for ppc platform in latest version as compared to the patched Qemu currently being used as part of coremu. What is the best way to make latest qemu v 0.15.0 work with coremu? On 8/26/11, zhaoguo wang <tig...@gm...> wrote: > Diviya: > I'm sorry there is no porting guide available. However we are > very glad to help > you for the porting work.To port QEMU for a new architecture, please > consider > the following aspects: > > 1. atomic instruction emulation. > see target-arm/translate.c and target-arm/cm-atomic.c/h > > 2. devices emulation. > You need to modify the interrupt controller emulation by using > interfaces > provided by coremu (e.g.coremu_send_intr). see > target-arm/cm-target-intr.c > and hw/arm_pic.c. > > You also need to make emulated devices can be concurrently accessed. > (e.g. Two emulated cpu threads may access a device at the same time.) > For arm architecture, we use a global lock to protect the emulated > device, it's > simple but no efficient. For x86_64 architecture, we parallel every > device > emulation which is very efficient but the development time is a little > long. > > We use #if defined(CONFIG_COREMU) to encapsulate all the changes in > QEMU. > For the code shared with other architecture, we use > #if defined(CONFIG_COREMU) && defined(TARGET_ARM) to encapsulate > the changes specially for arm architecture. > > > 2011/8/26 Diviya Jain <div...@gm...>: >> Hi, >> We need to port Qemu for powerpc to Coremu. Is there any porting guide >> available? As per your paper porting Coremu for ARM was very easy and >> required only 150 lines of code change to Qemu/Coremu. >> Could you please help us understand what all steps are required to port >> Coremu for powerpc. Is there any document for the same? Or perhaps if you >> could tell us changes done for ARM, so that we can do similar changes for >> powerpc to Qemu/Coremu source code. >> Another way might be grepping through qemu/coremu source code based on >> some >> keyword. Is there some #ifdef etc in qemu/coremu which encapsulate all the >> changes done to make it work for ARM platform? We can do similar for >> powerpc. >> Thanks, >> diviya >> ------------------------------------------------------------------------------ >> EMC VNX: the world's simplest storage, starting under $10K >> The only unified storage solution that offers unified management >> Up to 160% more powerful than alternatives and 25% more efficient. >> Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev >> _______________________________________________ >> Coremu-list-devel mailing list >> Cor...@li... >> https://lists.sourceforge.net/lists/listinfo/coremu-list-devel >> >> > > > > -- > Zhaoguo Wang, Parallel Processing Institute, Fudan University > > Address: Room 320, Software Building, 825 Zhangheng Road, Shanghai, China > > tig...@gm... > http://ppi.fudan.edu.cn/zhaoguo_wang > |
|
From: Ching-Tsun C. <chi...@gm...> - 2011-08-26 21:08:03
|
Do a diff between QEMU commit 56779034530944eb6171d843f652f3fba710ed30 and the QEMU part of COREMU and you'll see all the modifications and additions the COREMU team did to QEMU. Needless to say, a good diff utility would be very helpful here. - Ching Tsun On Fri, Aug 26, 2011 at 5:30 AM, zhaoguo wang <tig...@gm...>wrote: > Diviya: > I'm sorry there is no porting guide available. However we are > very glad to help > you for the porting work.To port QEMU for a new architecture, please > consider > the following aspects: > > 1. atomic instruction emulation. > see target-arm/translate.c and target-arm/cm-atomic.c/h > > 2. devices emulation. > You need to modify the interrupt controller emulation by using > interfaces > provided by coremu (e.g.coremu_send_intr). see > target-arm/cm-target-intr.c > and hw/arm_pic.c. > > You also need to make emulated devices can be concurrently accessed. > (e.g. Two emulated cpu threads may access a device at the same time.) > For arm architecture, we use a global lock to protect the emulated > device, it's > simple but no efficient. For x86_64 architecture, we parallel every > device > emulation which is very efficient but the development time is a little > long. > > We use #if defined(CONFIG_COREMU) to encapsulate all the changes in > QEMU. > For the code shared with other architecture, we use > #if defined(CONFIG_COREMU) && defined(TARGET_ARM) to encapsulate > the changes specially for arm architecture. > > > 2011/8/26 Diviya Jain <div...@gm...>: > > Hi, > > We need to port Qemu for powerpc to Coremu. Is there any porting guide > > available? As per your paper porting Coremu for ARM was very easy and > > required only 150 lines of code change to Qemu/Coremu. > > Could you please help us understand what all steps are required to port > > Coremu for powerpc. Is there any document for the same? Or perhaps if you > > could tell us changes done for ARM, so that we can do similar changes > for > > powerpc to Qemu/Coremu source code. > > Another way might be grepping through qemu/coremu source code based on > some > > keyword. Is there some #ifdef etc in qemu/coremu which encapsulate all > the > > changes done to make it work for ARM platform? We can do similar for > > powerpc. > > Thanks, > > diviya > > > ------------------------------------------------------------------------------ > > EMC VNX: the world's simplest storage, starting under $10K > > The only unified storage solution that offers unified management > > Up to 160% more powerful than alternatives and 25% more efficient. > > Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev > > _______________________________________________ > > Coremu-list-devel mailing list > > Cor...@li... > > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > > > > > > > > -- > Zhaoguo Wang, Parallel Processing Institute, Fudan University > > Address: Room 320, Software Building, 825 Zhangheng Road, Shanghai, China > > tig...@gm... > http://ppi.fudan.edu.cn/zhaoguo_wang > > > ------------------------------------------------------------------------------ > EMC VNX: the world's simplest storage, starting under $10K > The only unified storage solution that offers unified management > Up to 160% more powerful than alternatives and 25% more efficient. > Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > |
|
From: Yu-Wen H. <pla...@gm...> - 2011-08-26 14:58:04
|
Hello, You may first try to grep "CONFIG_COREMU" under the "qemu/target-arm" directory. Basically, you have to implement the emulation of atomic instructions of the powerpc ISA. Also, you have to enable the translation scheme for powerpc to be executed multi-threadedly. For example, you have to make those global variables in "translate.c" thread-private, etc. I'm not a guy from the COREMU group, but I hope maybe this could provide some directions for you. And hope maybe they will provide a more complete guide for you later. Best Regards, Yu-wen 2011/8/26 <cor...@li...> > Send Coremu-list-devel mailing list submissions to > cor...@li... > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > or, via email, send a message with subject or body 'help' to > cor...@li... > > You can reach the person managing the list at > cor...@li... > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Coremu-list-devel digest..." > > > Today's Topics: > > 1. Porting coremu for powerpc platform (Diviya Jain) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Fri, 26 Aug 2011 14:25:59 +0530 > From: Diviya Jain <div...@gm...> > Subject: [Coremu-list-devel] Porting coremu for powerpc platform > To: cor...@li... > Message-ID: > <CAMpxYCMFrPTb91X6o0phB6B68Zec13PJth5TYQ-n2=03SUt4=Q...@ma... > > > Content-Type: text/plain; charset="iso-8859-1" > > Hi, > > We need to port Qemu for powerpc to Coremu. Is there any porting guide > available? As per your paper porting Coremu for ARM was very easy and > required only 150 lines of code change to Qemu/Coremu. > > Could you please help us understand what all steps are required to port > Coremu for powerpc. Is there any document for the same? Or perhaps if you > could tell us changes done for ARM, so that we can do similar changes for > powerpc to Qemu/Coremu source code. > > Another way might be grepping through qemu/coremu source code based on some > keyword. Is there some #ifdef etc in qemu/coremu which encapsulate all the > changes done to make it work for ARM platform? We can do similar for > powerpc. > > Thanks, > diviya > -------------- next part -------------- > An HTML attachment was scrubbed... > > ------------------------------ > > > ------------------------------------------------------------------------------ > EMC VNX: the world's simplest storage, starting under $10K > The only unified storage solution that offers unified management > Up to 160% more powerful than alternatives and 25% more efficient. > Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev > > ------------------------------ > > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > > > End of Coremu-list-devel Digest, Vol 8, Issue 1 > *********************************************** > |
|
From: zhaoguo w. <tig...@gm...> - 2011-08-26 12:30:25
|
Diviya:
I'm sorry there is no porting guide available. However we are
very glad to help
you for the porting work.To port QEMU for a new architecture, please consider
the following aspects:
1. atomic instruction emulation.
see target-arm/translate.c and target-arm/cm-atomic.c/h
2. devices emulation.
You need to modify the interrupt controller emulation by using interfaces
provided by coremu (e.g.coremu_send_intr). see target-arm/cm-target-intr.c
and hw/arm_pic.c.
You also need to make emulated devices can be concurrently accessed.
(e.g. Two emulated cpu threads may access a device at the same time.)
For arm architecture, we use a global lock to protect the emulated
device, it's
simple but no efficient. For x86_64 architecture, we parallel every device
emulation which is very efficient but the development time is a little long.
We use #if defined(CONFIG_COREMU) to encapsulate all the changes in QEMU.
For the code shared with other architecture, we use
#if defined(CONFIG_COREMU) && defined(TARGET_ARM) to encapsulate
the changes specially for arm architecture.
2011/8/26 Diviya Jain <div...@gm...>:
> Hi,
> We need to port Qemu for powerpc to Coremu. Is there any porting guide
> available? As per your paper porting Coremu for ARM was very easy and
> required only 150 lines of code change to Qemu/Coremu.
> Could you please help us understand what all steps are required to port
> Coremu for powerpc. Is there any document for the same? Or perhaps if you
> could tell us changes done for ARM, so that we can do similar changes for
> powerpc to Qemu/Coremu source code.
> Another way might be grepping through qemu/coremu source code based on some
> keyword. Is there some #ifdef etc in qemu/coremu which encapsulate all the
> changes done to make it work for ARM platform? We can do similar for
> powerpc.
> Thanks,
> diviya
> ------------------------------------------------------------------------------
> EMC VNX: the world's simplest storage, starting under $10K
> The only unified storage solution that offers unified management
> Up to 160% more powerful than alternatives and 25% more efficient.
> Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
> _______________________________________________
> Coremu-list-devel mailing list
> Cor...@li...
> https://lists.sourceforge.net/lists/listinfo/coremu-list-devel
>
>
--
Zhaoguo Wang, Parallel Processing Institute, Fudan University
Address: Room 320, Software Building, 825 Zhangheng Road, Shanghai, China
tig...@gm...
http://ppi.fudan.edu.cn/zhaoguo_wang
|
|
From: Diviya J. <div...@gm...> - 2011-08-26 08:56:06
|
Hi, We need to port Qemu for powerpc to Coremu. Is there any porting guide available? As per your paper porting Coremu for ARM was very easy and required only 150 lines of code change to Qemu/Coremu. Could you please help us understand what all steps are required to port Coremu for powerpc. Is there any document for the same? Or perhaps if you could tell us changes done for ARM, so that we can do similar changes for powerpc to Qemu/Coremu source code. Another way might be grepping through qemu/coremu source code based on some keyword. Is there some #ifdef etc in qemu/coremu which encapsulate all the changes done to make it work for ARM platform? We can do similar for powerpc. Thanks, diviya |
|
From: Chen Y. <cyf...@gm...> - 2011-02-24 07:54:09
|
The error happened when linking seabios. We only modified seabios a little to make it support 255 cores and are not expert on it. So I suggest you find other machine with a different version of binutils and try again. We use debian 5 and binutils version 2.18. 2011/2/24 Jungho Park <dav...@gm...>: > Hi, > > Our system is x86_64 linux machine. > gcc version is 4.1.2 and the version of binutil is 2.21. > > Command for configure is following. > ./configure --target=x86_64 --prefix=/home/jj/coremu-0.1.1 > /home/jj/coremu/coremu-0.1.1/qemu > make > make install > > I don't touch any source code :) > What do you guess of this make install problem? > > Thanks. > > 2011-02-23 오후 5:21, Chen Yufei 쓴 글: > > Thanks for your interest in COREMU. > > I've never encountered these error message. What's the system you used > and how did you run configure? > > On Wed, Feb 23, 2011 at 1:58 PM, Jungho Park <dav...@gm...> wrote: > > > Hi, I read the PPoPP'11 paper about this project, so I want to execution it. > But, I cannot install because of following problem. > > make[1]: Entering directory > `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' > Linking out/rom.o > out/rombios.lds:25 cannot move location counter backwards (from > 00000000000f7594 to 00000000000f758c) > make[1]: *** [out/rom.o] Error 1 > make[1]: Leaving directory > `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' > make: *** > [/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios/out/bios.bin] Error 2 > > How can I solve that problem...? > Please help me. Thanks. > > ------------------------------------------------------------------------------ > Free Software Download: Index, Search & Analyze Logs and other IT data in > Real-Time with Splunk. Collect, index and harness all the fast moving IT > data > generated by your applications, servers and devices whether physical, > virtual > or in the cloud. Deliver compliance at lower cost and gain new business > insights. http://p.sf.net/sfu/splunk-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > > > > ------------------------------------------------------------------------------ > Free Software Download: Index, Search & Analyze Logs and other IT data in > Real-Time with Splunk. Collect, index and harness all the fast moving IT > data > generated by your applications, servers and devices whether physical, > virtual > or in the cloud. Deliver compliance at lower cost and gain new business > insights. http://p.sf.net/sfu/splunk-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > > -- Best regards, Chen Yufei |
|
From: Jungho P. <dav...@gm...> - 2011-02-24 07:17:40
|
Hi, Our system is x86_64 linux machine. gcc version is 4.1.2 and the version of binutil is 2.21. Command for configure is following. ./configure --target=x86_64 --prefix=/home/jj/coremu-0.1.1 /home/jj/coremu/coremu-0.1.1/qemu make make install I don't touch any source code :) What do you guess of this make install problem? Thanks. 2011-02-23 오후 5:21, Chen Yufei 쓴 글: Thanks for your interest in COREMU. I've never encountered these error message. What's the system you used and how did you run configure? On Wed, Feb 23, 2011 at 1:58 PM, Jungho Park <dav...@gm...> wrote: > Hi, I read the PPoPP'11 paper about this project, so I want to execution it. > But, I cannot install because of following problem. > > make[1]: Entering directory > `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' > Linking out/rom.o > out/rombios.lds:25 cannot move location counter backwards (from > 00000000000f7594 to 00000000000f758c) > make[1]: *** [out/rom.o] Error 1 > make[1]: Leaving directory > `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' > make: *** > [/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios/out/bios.bin] Error 2 > > How can I solve that problem...? > Please help me. Thanks. > > ------------------------------------------------------------------------------ > Free Software Download: Index, Search & Analyze Logs and other IT data in > Real-Time with Splunk. Collect, index and harness all the fast moving IT data > generated by your applications, servers and devices whether physical, virtual > or in the cloud. Deliver compliance at lower cost and gain new business > insights. http://p.sf.net/sfu/splunk-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > > |
|
From: Chen Y. <cyf...@gm...> - 2011-02-23 08:22:34
|
Thanks for your interest in COREMU. I've never encountered these error message. What's the system you used and how did you run configure? On Wed, Feb 23, 2011 at 1:58 PM, Jungho Park <dav...@gm...> wrote: > Hi, I read the PPoPP'11 paper about this project, so I want to execution it. > But, I cannot install because of following problem. > > make[1]: Entering directory > `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' > Linking out/rom.o > out/rombios.lds:25 cannot move location counter backwards (from > 00000000000f7594 to 00000000000f758c) > make[1]: *** [out/rom.o] Error 1 > make[1]: Leaving directory > `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' > make: *** > [/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios/out/bios.bin] Error 2 > > How can I solve that problem...? > Please help me. Thanks. > > ------------------------------------------------------------------------------ > Free Software Download: Index, Search & Analyze Logs and other IT data in > Real-Time with Splunk. Collect, index and harness all the fast moving IT data > generated by your applications, servers and devices whether physical, virtual > or in the cloud. Deliver compliance at lower cost and gain new business > insights. http://p.sf.net/sfu/splunk-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > -- Best regards, Chen Yufei |
|
From: Jungho P. <dav...@gm...> - 2011-02-23 05:59:01
|
Hi, I read the PPoPP'11 paper about this project, so I want to execution it. But, I cannot install because of following problem. make[1]: Entering directory `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' Linking out/rom.o out/rombios.lds:25 cannot move location counter backwards (from 00000000000f7594 to 00000000000f758c) make[1]: *** [out/rom.o] Error 1 make[1]: Leaving directory `/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios' make: *** [/home/jj/coremu/coremu-0.1.1/coremu/bios/seabios/out/bios.bin] Error 2 How can I solve that problem...? Please help me. Thanks. |
|
From: Chen Y. <cyf...@gm...> - 2011-01-24 02:26:40
|
Thanks for your interest. COREMU can only run on x86_64 Linux system. You can try it on a x86_64 Linux system. We've only tested COREMU on Debian Lenny, Ubuntu should also be OK. But in case there's any problem, feel free to send us emails. 2011/1/23 郭旭峰 <gu...@16...>: > hi, i'm very interested in this emulator and want to try it. > But there are some compliing errors, seeing the attachment. > I use Ubuntu 9.10, gcc 4.4.1. Should I install some other lib packages? > > > thanks~ > > > -- > Xufeng Guo > > > > > ------------------------------------------------------------------------------ > Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)! > Finally, a world-class log management solution at an even better price-free! > Download using promo code Free_Logger_4_Dev2Dev. Offer expires > February 28th, so secure your free ArcSight Logger TODAY! > http://p.sf.net/sfu/arcsight-sfd2d > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > > -- Best regards, Chen Yufei |
|
From: Chen Y. <cyf...@gm...> - 2011-01-21 07:35:15
|
Download available on sourceforge.
Changelog:
0.1.2:
- Support i386 target. Upstream qemu (i386 target) can only start Linux with 8
cores, we have the same problem
- Support qcow2 image format
- Merge upstream qemu 56779034530944eb6171d843f652f3fba710ed30
--
Best regards,
Chen Yufei
|
|
From: Chen Y. <cyf...@gm...> - 2011-01-02 14:59:05
|
The fix is in the "merge-upstream" branch. It simply makes "bdrv_read_em" and "bdrv_write_em" really synchronous, to avoid cpu and hardware thread from calling AIO callbacks and bottom halves at the same time. -- Best regards, Chen Yufei |
|
From: Yufei C. <cyf...@gm...> - 2010-12-12 09:35:21
|
We've created a new branch which will merge the latest upstream QEMU into to the latest COREMU master branch. Both the COREMU and patched QEMU git repositories have this branch. PS. We tried git rebase first, but as we have multiple developers and different branches for different features, rebase causes some troubles. After reading the git manual, we found that using "git log --first-parent" can get a clean history even using merge. -- Best regards, Chen Yufei |
|
From: wang T. <tig...@gm...> - 2010-11-08 12:53:43
|
2010/11/8 Antti P Miettinen <an...@ik...> > From: wang Tiger <tig...@gm...> > > I'm trying to make COREMU support e1000 net card emulation for x86 > platform > > now. But the process is not very fast. :-) > > I haven't been able to put much effort into COREMU lately but one > thing that occured to me is that the current implementation of > ldrex/strex for ARM has at least in principle ABA-problem since it > relies on the stored value being the same as before. However, I do not > see any way to fix this without a large performance penalty. And for > Linux spinlocks the ABA-problem should not be an issue, right? Yes, but it not be an issue for atomic instruction emulation. PS: the hang problem is caused by the serial port emulation but i haven't identify the reason exactly. :-) > > -- > http://www.iki.fi/~ananaza/ > -- Zhaoguo Wang, Parallel Processing Institute, Fudan University Address: Room 320, Software Building, 825 Zhangheng Road, Shanghai, China tig...@gm... http://ppi.fudan.edu.cn/zhaoguo_wang |
|
From: Antti P M. <an...@ik...> - 2010-11-08 08:43:11
|
From: wang Tiger <tig...@gm...> > I'm trying to make COREMU support e1000 net card emulation for x86 platform > now. But the process is not very fast. :-) I haven't been able to put much effort into COREMU lately but one thing that occured to me is that the current implementation of ldrex/strex for ARM has at least in principle ABA-problem since it relies on the stored value being the same as before. However, I do not see any way to fix this without a large performance penalty. And for Linux spinlocks the ABA-problem should not be an issue, right? -- http://www.iki.fi/~ananaza/ |
|
From: wang T. <tig...@gm...> - 2010-10-25 03:08:29
|
2010/10/24 Antti P Miettinen <an...@ik...> > I haven't been able to make COREMU ARM guest with NFS root completely > stable but I thought I'd share my current thoughts. Maybe someone > finds/knows the critical places needing fixing. > > One thing is that COREMU calls the event notifier directly when it is > sending an interrupt to the calling core. This means that the notifier > can get pre-empted by a signal handler (signal coming from e.g. main > thread handling IO). So should signals be blocked while calling event > notifier directly? > > Another issue is timer handling. For Realview PBX, the relevant code > is in hw/mpcore.c. For example mpcore_timer_tick() is called from main > thread context, whereas the IO read/write callbacks are called from > VCPU thread context. These call the timer functions in > qemu-timer.c. Then there is the host_alarm_handler() signal handler > that can pre-empt the main thread. > > host_alarm_handler is the timer handler for emulation of timer interrupt. It can only pre-empt the vcpu thread. > I've tried adding signal blocking, spinlocks, mutex locking but > haven't been able to find a well working configuration. I can make the > boot from NFS root functionally quite robust but then the boot process > has weird delays indicating that timer handling is still somehow > broken. > I'm trying to make COREMU support e1000 net card emulation for x86 platform now. But the process is not very fast. :-) > > -- > http://www.iki.fi/~ananaza/ > > > ------------------------------------------------------------------------------ > Nokia and AT&T present the 2010 Calling All Innovators-North America > contest > Create new apps & games for the Nokia N8 for consumers in U.S. and Canada > $10 million total in prizes - $4M cash, 500 devices, nearly $6M in > marketing > Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store > http://p.sf.net/sfu/nokia-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel > -- Zhaoguo Wang, Parallel Processing Institute, Fudan University Address: Room 320, Software Building, 825 Zhangheng Road, Shanghai, China tig...@gm... http://ppi.fudan.edu.cn/zhaoguo_wang |
|
From: Antti P M. <an...@ik...> - 2010-10-24 08:32:28
|
I haven't been able to make COREMU ARM guest with NFS root completely stable but I thought I'd share my current thoughts. Maybe someone finds/knows the critical places needing fixing. One thing is that COREMU calls the event notifier directly when it is sending an interrupt to the calling core. This means that the notifier can get pre-empted by a signal handler (signal coming from e.g. main thread handling IO). So should signals be blocked while calling event notifier directly? Another issue is timer handling. For Realview PBX, the relevant code is in hw/mpcore.c. For example mpcore_timer_tick() is called from main thread context, whereas the IO read/write callbacks are called from VCPU thread context. These call the timer functions in qemu-timer.c. Then there is the host_alarm_handler() signal handler that can pre-empt the main thread. I've tried adding signal blocking, spinlocks, mutex locking but haven't been able to find a well working configuration. I can make the boot from NFS root functionally quite robust but then the boot process has weird delays indicating that timer handling is still somehow broken. -- http://www.iki.fi/~ananaza/ |
|
From: Chen Y. <cyf...@gm...> - 2010-10-21 03:25:27
|
On 2010-10-20, at 下午12:35, Antti P Miettinen wrote: > Seems that --use-lockfree does not really enable lockfree queue. This > is for the coremu git tree. Thanks. I'll update the COREMU makefile. > > diff --git a/Makefile b/Makefile > index 3e45d58..a5dbbf1 100644 > --- a/Makefile > +++ b/Makefile > @@ -13,6 +13,7 @@ > all: > > include config.mk > +include coremu.mk > include $(addsuffix /module.mk, $(modules)) > > objects += $(patsubst %.c, $(OBJDIR)/%.o, $(foreach dir, $(modules), $(wildcard $(dir)/*.c))) > @@ -30,7 +31,7 @@ LIBOBJS += $(OBJDIR)/main/sched.o > LIBOBJS += $(OBJDIR)/main/timer.o > LIBOBJS += $(OBJDIR)/main/intr.o > LIBOBJS += $(OBJDIR)/main/malloc-helper.o > -ifdef $(LOCKFREE) > +ifdef LOCKFREE > LIBOBJS += $(OBJDIR)/main/ms-queue.o > else > LIBOBJS += $(OBJDIR)/main/lock-queue.o > @@ -39,7 +40,7 @@ LIBOBJS += $(OBJDIR)/main/utils.o > > $(OBJDIR)/libcoremu.a: $(LIBOBJS) > $(call quiet-command,rm -f $@ && $(AR) rcs $@ $^," AR $(TARGET_DIR)$@") > -qemu: > +qemu: $(OBJDIR)/libcoremu.a > @echo "---- [coremu]: build qemu ----" > $(MAKE) -C $(COREMUDIR)/obj/qemu > > > -- > http://www.iki.fi/~ananaza/ > > ------------------------------------------------------------------------------ > Download new Adobe(R) Flash(R) Builder(TM) 4 > The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly > Flex(R) Builder(TM)) enable the development of rich applications that run > across multiple browsers and platforms. Download your free trials today! > http://p.sf.net/sfu/adobe-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel |
|
From: Antti P M. <an...@ik...> - 2010-10-20 04:36:00
|
Seems that --use-lockfree does not really enable lockfree queue. This is for the coremu git tree. diff --git a/Makefile b/Makefile index 3e45d58..a5dbbf1 100644 --- a/Makefile +++ b/Makefile @@ -13,6 +13,7 @@ all: include config.mk +include coremu.mk include $(addsuffix /module.mk, $(modules)) objects += $(patsubst %.c, $(OBJDIR)/%.o, $(foreach dir, $(modules), $(wildcard $(dir)/*.c))) @@ -30,7 +31,7 @@ LIBOBJS += $(OBJDIR)/main/sched.o LIBOBJS += $(OBJDIR)/main/timer.o LIBOBJS += $(OBJDIR)/main/intr.o LIBOBJS += $(OBJDIR)/main/malloc-helper.o -ifdef $(LOCKFREE) +ifdef LOCKFREE LIBOBJS += $(OBJDIR)/main/ms-queue.o else LIBOBJS += $(OBJDIR)/main/lock-queue.o @@ -39,7 +40,7 @@ LIBOBJS += $(OBJDIR)/main/utils.o $(OBJDIR)/libcoremu.a: $(LIBOBJS) $(call quiet-command,rm -f $@ && $(AR) rcs $@ $^," AR $(TARGET_DIR)$@") -qemu: +qemu: $(OBJDIR)/libcoremu.a @echo "---- [coremu]: build qemu ----" $(MAKE) -C $(COREMUDIR)/obj/qemu -- http://www.iki.fi/~ananaza/ |
|
From: Yufei C. <cyf...@gm...> - 2010-10-19 03:38:48
|
On Tue, Oct 19, 2010 at 12:20 AM, Antti P Miettinen <an...@ik...> wrote:
> From: Chen Yufei <cyf...@gm...>
>> I rebased COREMU to the latest qemu a couple of weeks before and
>> fixed several problems to make x86_64 working.
>>
>> To my knowledge, rebased branch are not to be published, since other
>> people can't pull it again if I do a rebase later.
>>
>> Use merge can avoid the problem of rebasing. But it would be
>> difficult if we want to submit our change to upstream later.
>>
>> Do you have some suggestions on this subject?
>
> As far as I understand, rebase changes history so I can see how
> rebased branches in public trees are a problem. I have not run into a
> situation where I would want to do a rebase. I've been happy with
> fetch and merge. Diffing against master in a branch to which master is
> merged should look pretty similar to diffing a rebased branch. So why
> not just preserve the history? :-)
Rebase would give us a history with our modifications to qemu
consecutively. But the history results from merge would clutter our
modification with upstream's. So it's difficult to find our
modifications using merge. ('git diff --topo-order' would help a bit
on this problem.)
I've seen people maintaining two branch to keep up with upstream. One
uses rebase for own development and submitting patches to upstream;
the other uses merge to publish his own modifications. But this
approach requires extra work.
>
> Submitting the changes upstream would probably require splitting up
> the current work to small pieces. I have not thought how that would be
> organized in git - several small branches?
>
> --
> http://www.iki.fi/~ananaza/
>
--
Best regards,
Chen Yufei
|
|
From: Antti P M. <an...@ik...> - 2010-10-18 16:20:36
|
From: Chen Yufei <cyf...@gm...> > I rebased COREMU to the latest qemu a couple of weeks before and > fixed several problems to make x86_64 working. > > To my knowledge, rebased branch are not to be published, since other > people can't pull it again if I do a rebase later. > > Use merge can avoid the problem of rebasing. But it would be > difficult if we want to submit our change to upstream later. > > Do you have some suggestions on this subject? As far as I understand, rebase changes history so I can see how rebased branches in public trees are a problem. I have not run into a situation where I would want to do a rebase. I've been happy with fetch and merge. Diffing against master in a branch to which master is merged should look pretty similar to diffing a rebased branch. So why not just preserve the history? :-) Submitting the changes upstream would probably require splitting up the current work to small pieces. I have not thought how that would be organized in git - several small branches? -- http://www.iki.fi/~ananaza/ |
|
From: Chen Y. <cyf...@gm...> - 2010-10-18 13:11:23
|
On 2010-10-18, at 下午8:34, Antti P Miettinen wrote: > I tried bringing the COREMU changes up to date with the latest QEMU > master branch. The work is in coremu_forward branch in > git://epm0.nrln.net/git/coremu-apm.git repo. It compiles and ARM quest > boots (as unrealiably as before) but I have no idea whether PC quest > is working. I rebased COREMU to the latest qemu a couple of weeks before and fixed several problems to make x86_64 working. To my knowledge, rebased branch are not to be published, since other people can't pull it again if I do a rebase later. Use merge can avoid the problem of rebasing. But it would be difficult if we want to submit our change to upstream later. Do you have some suggestions on this subject? > > -- > http://www.iki.fi/~ananaza/ > > ------------------------------------------------------------------------------ > Download new Adobe(R) Flash(R) Builder(TM) 4 > The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly > Flex(R) Builder(TM)) enable the development of rich applications that run > across multiple browsers and platforms. Download your free trials today! > http://p.sf.net/sfu/adobe-dev2dev > _______________________________________________ > Coremu-list-devel mailing list > Cor...@li... > https://lists.sourceforge.net/lists/listinfo/coremu-list-devel |
|
From: Antti P M. <an...@ik...> - 2010-10-18 12:35:09
|
I tried bringing the COREMU changes up to date with the latest QEMU master branch. The work is in coremu_forward branch in git://epm0.nrln.net/git/coremu-apm.git repo. It compiles and ARM quest boots (as unrealiably as before) but I have no idea whether PC quest is working. -- http://www.iki.fi/~ananaza/ |
1 message has been excluded from this view by a project administrator.
