Hello,
I have a memory trouble with an old DOS OS from Digital Research called Concurrent DOS. It worked fine till Qemu v0.11.1. I use Qemu to make a legacy application run without any modification.
Starting with Qemu 0.12, my legacy application has not enough memory to run. Actually it seems that the SeaBIOS doesn't use the same memory mapping than PC-BIOS used to, causing trouble to my application.
Is there any way to get SeaBIOS use (more or less) the same memory amount and memory mapping than the older PC-BIOS ?
I tried to recomplie SeaBIOS 1.7.0 with less options but the result is the same. Here is my .config file :
CONFIG_RELOCATE_INIT=y CONFIG_ATA=y CONFIG_FLOPPY=y CONFIG_PS2PORT=y CONFIG_SERIAL=y CONFIG_LPT=y CONFIG_DRIVES=y CONFIG_OPTIONROMS=y CONFIG_PMM=y CONFIG_BOOT=y CONFIG_KEYBOARD=y CONFIG_KBD_CALL_INT15_4F=y CONFIG_MOUSE=y CONFIG_NO_VGABIOS=y CONFIG_DEBUG_LEVEL=0
I also join two print screen of the same C-DOS utility. One with Qemu 1.0 and SeaBIOS 1.7.0 and the other with Qemu 0.10.0. The amount of memory is quite different between the versions. If you want to try, get this floppy : https://docs.google.com/open?id=0B7mz0vq6Rpb7Q0tYbmxMVzVqTms and then type : STOP
If you have any idea...
Thanks, Alain
Alain Ribière wrote:
I also join two print screen of the same C-DOS utility. One with Qemu 1.0 and SeaBIOS 1.7.0 and the other with Qemu 0.10.0. The amount of memory is quite different between the versions.
It might be interesting to see even more verbose output, like from the mem tool in MS-DOS. Maybe you have some DOS images around and can compare?
//Peter
On Mon, May 07, 2012 at 08:58:45AM -0700, Alain Ribière wrote:
Hello,
I have a memory trouble with an old DOS OS from Digital Research called Concurrent DOS. It worked fine till Qemu v0.11.1. I use Qemu to make a legacy application run without any modification.
Starting with Qemu 0.12, my legacy application has not enough memory to run. Actually it seems that the SeaBIOS doesn't use the same memory mapping than PC-BIOS used to, causing trouble to my application.
Is there any way to get SeaBIOS use (more or less) the same memory amount and memory mapping than the older PC-BIOS ?
I tried to recomplie SeaBIOS 1.7.0 with less options but the result is the same.
When reporting an issue with SeaBIOS, please include the log obtained when running qemu with "-chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios".
I took a quick look at this, and I don't think this is a memory size issue - the screenshots actually show that running under seabios makes more memory available (15664 vs 15648). Instead, I'd guess concurrent dos is not making a "banked window" that is large enough. I don't know why the memory segments between 0x0b00 and 0x3400 aren't being used. There is no legacy bios interface to reserve memory under 1meg, so I'm sure SeaBIOS isn't doing this wrong. Indeed, SeaBIOS is much more careful about touching ram under 1meg than Bochs bios is.
So, I'd guess something unrelated is confusing concurrent dos and the above is just a symptom of the problem. Unfortunately, I don't know of any way to track that down without source code for the image and spending a large amount of debugging time.
-Kevin
Thanks for your answer.
Here is the debug log in attachement. I disabled the debug on the screen shot I sent because I noticed I could get a bit more memory without it. With the debug on, I got : Banked Window 416K at 3800:0 segment and a free memory of 15680K
I'll try to find what confuses C-DOS. It reacts a bit strangely... When I add or remove options from the SeaBIOS, the memory avalaible change and the size of banked window changes too. So I hoped there was a way to have enough memory to run my application... Otherwise, I will have to stuck with Qemu 0.11 and PC-BIOS.
Do you know if there is any DOS utility to get the memory map ?
Thanks,
Alain
________________________________ On Mon, May 07, 2012 at 08:58:45AM -0700, Alain Ribière wrote:
Hello,
I have a memory trouble with an old DOS OS from Digital Research
called Concurrent DOS.
It worked fine till Qemu v0.11.1. I
use Qemu to make a legacy application run without any modification.
Starting with Qemu 0.12, my legacy application has not enough memory to run. Actually it seems that the SeaBIOS doesn't use the same memory mapping than PC-BIOS used to, causing trouble to my application.
Is there any way to get SeaBIOS use (more or less) the same memory amount and memory mapping than the older PC-BIOS ?
I tried to recomplie SeaBIOS 1.7.0 with less options but the result is the same.
When reporting an issue with SeaBIOS, please include the log obtained when running qemu with "-chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios".
I took a quick look at this, and I don't think this is a memory size issue - the screenshots actually show that running under seabios makes more memory available (15664 vs 15648). Instead, I'd guess concurrent dos is not making a "banked window" that is large enough. I don't know why the memory segments between 0x0b00 and 0x3400 aren't being used. There is no legacy bios interface to reserve memory under 1meg, so I'm sure SeaBIOS isn't doing this wrong. Indeed, SeaBIOS is much more careful about touching ram under 1meg than Bochs bios is.
So, I'd guess something unrelated is confusing concurrent dos and the above is just a symptom of the problem. Unfortunately, I don't know of any way to track that down without source code for the image and spending a large amount of debugging time.
-Kevin
On Tue, May 08, 2012 at 09:47:56AM -0700, Alain Ribière wrote:
Thanks for your answer.
Here is the debug log in attachement. I disabled the debug on the screen shot I sent because I noticed I could get a bit more memory without it. With the debug on, I got : Banked Window 416K at 3800:0 segment and a free memory of 15680K
I'll try to find what confuses C-DOS. It reacts a bit strangely... When I add or remove options from the SeaBIOS, the memory avalaible change and the size of banked window changes too. So I hoped there was a way to have enough memory to run my application...
The options don't impact SeaBIOS' memory usage, so this reafirms my guess that what you are seeing is a secondary effect. As a guess, this could be stack usage - some old program call the bios with very small stacks and if the stack overflows silent corruption can occur. The fact that you're seeing different memory layouts with random changes to config options could be due to different options causing the compiler to layout the stack slightly differently and thus causing slightly different corruption - just a guess.
The other thing I noticed with your C-DOS image is that it doesn't know about the EBDA and uses that memory for its own purposes. Thus, the seabios runtime usage of the ebda along with c-dos could cause conflicts. I don't think this is the cause of your current issue though as some moving the ebda in some quick tests I ran doesn't seem to impact the output of "stop".
Otherwise, I will have to stuck with Qemu 0.11 and PC-BIOS.
Do you know if there is any DOS utility to get the memory map ?
I'm not familiar with DOS utilities. You could ask on the freedos mailing list.
-Kevin
Tanks for your answer.
I have to apologize, I made a mistake. I rechecked my application and used git bisect. So I found that this is not a SeaBIOS issue (actually it's more a C-DOS issues...). My problem comes from this Qemu commit :
49669fc551e0ccd2310a9584a9b7343a0bcae477 is the first bad commit commit 49669fc551e0ccd2310a9584a9b7343a0bcae477 Author: Glauber Costa glommer@redhat.com Date: Thu Apr 23 15:20:55 2009 -0400
suport device driver initialization model According to PnP specification, Appendix B, Option ROMs that support DDIM (device driver initialization model) should have their memory space writeable. KVM deviates from us here, by removing the IO_MEM_ROM flag, to allow for PCI option ROMs (they require DDIM). However, there's absolutely no reason we can't do the same. Signed-off-by: Glauber Costa glommer@redhat.com Signed-off-by: Anthony Liguori aliguori@us.ibm.com
:040000 040000 5186361da34236c2680af6ed568b5f5bdb2915ac c968308168e23ffa2678011ed2ff010389b528bd M hw
diff --git a/hw/pc.c b/hw/pc.c index 61f6e7b..340b313 100644 --- a/hw/pc.c +++ b/hw/pc.c @@ -920,7 +920,8 @@ static void pc_init1(ram_addr_t ram_size, int vga_ram_size, option_rom_offset = qemu_ram_alloc(0x20000); oprom_area_size = 0; - cpu_register_physical_memory(0xc0000, 0x20000, option_rom_offset); + cpu_register_physical_memory(0xc0000, 0x20000, + option_rom_offset | IO_MEM_ROM); if (using_vga) { /* VGA BIOS load */
Also I found the theorical memory map of the C-DOS in the documentation of my legacy application : start end
000000 9FFFF Program area
0A0000 0BFFFF Video memory
0C00000 0C7FFF BIOS graphic card 0C8000 0CBFFF BIOS cache controller 0CC000 0CFFFF BIOS removable disk 0D0000 0D3FFF BIOS network adapter 0D4000 0EFFFF CDOS system 0F0000 0FFFFF ROM BIOS
I wonder what is the CDOS system at 0D4000... It seems a bit strange.
I do not understand why this commit affects the C-DOS memory allocation. But you're probably right, C-DOS is using memory addresses he shouldn't for it's own purpose causing conflicts...
If you have any idea.
Thanks, and sorry again for the wrong issue (I will remove the SeaBIOS mailing list in the future mails)...
Alain
________________________________
On Tue, May 08, 2012 at 09:47:56AM -0700, Alain Ribière wrote:
Thanks for your answer.
Here is the debug log in attachement. I disabled the debug on the screen shot I sent because I noticed I could get a bit more memory without it. With the debug on, I got : Banked Window 416K at 3800:0 segment and a free memory of 15680K
I'll try to find what confuses C-DOS. It reacts a bit strangely... When I add or remove options from the SeaBIOS, the memory avalaible change and the size of banked window changes too. So I hoped there was a way to have enough memory to run my application...
The options don't impact SeaBIOS' memory usage, so this reafirms my guess that what you are seeing is a secondary effect. As a guess, this could be stack usage - some old program call the bios with very small stacks and if the stack overflows silent corruption can occur. The fact that you're seeing different memory layouts with random changes to config options could be due to different options causing the compiler to layout the stack slightly differently and thus causing slightly different corruption - just a guess.
The other thing I noticed with your C-DOS image is that it doesn't know about the EBDA and uses that memory for its own purposes. Thus, the seabios runtime usage of the ebda along with c-dos could cause conflicts. I don't think this is the cause of your current issue though as some moving the ebda in some quick tests I ran doesn't seem to impact the output of "stop".
Otherwise, I will have to stuck with Qemu 0.11 and PC-BIOS.
Do you know if there is any DOS utility to get the memory map ?
I'm not familiar with DOS utilities. You could ask on the freedos mailing list.
-Kevin
On Wed, May 09, 2012 at 08:57:53AM -0700, Alain Ribière wrote:
Tanks for your answer.
I have to apologize, I made a mistake. I rechecked my application and used git bisect. So I found that this is not a SeaBIOS issue (actually it's more a C-DOS issues...). My problem comes from this Qemu commit :
49669fc551e0ccd2310a9584a9b7343a0bcae477 is the first bad commit commit 49669fc551e0ccd2310a9584a9b7343a0bcae477
[...]
Also I found the theorical memory map of the C-DOS in the documentation of my legacy application : start end
000000 9FFFF Program area
0A0000 0BFFFF Video memory
0C00000 0C7FFF BIOS graphic card 0C8000 0CBFFF BIOS cache controller 0CC000 0CFFFF BIOS removable disk 0D0000 0D3FFF BIOS network adapter 0D4000 0EFFFF CDOS system 0F0000 0FFFFF ROM BIOS
I wonder what is the CDOS system at 0D4000... It seems a bit strange.
Very odd. This is still controlled by seabios. Indeed, if I apply this (clearly incorrect) patch to seabios:
diff --git a/src/shadow.c b/src/shadow.c index c0c8cc2..73c0670 100644 --- a/src/shadow.c +++ b/src/shadow.c @@ -83,7 +83,7 @@ make_bios_readonly_intel(u16 bdf, u32 pam0) for (i=0; i<6; i++) { u32 mem = BUILD_ROM_START + i * 32*1024; u32 pam = pam0 + 1 + i; - if (RomEnd <= mem + 16*1024) { + if (0 && RomEnd <= mem + 16*1024) { if (RomEnd > mem) pci_config_writeb(bdf, pam, 0x31); break;
Then the C-DOS image you posted seems happier. Basically, if C-DOS is allowed to write to the e-segment then it seems to do weird things.
I do not understand why this commit affects the C-DOS memory allocation. But you're probably right, C-DOS is using memory addresses he shouldn't for it's own purpose causing conflicts...
If you have any idea.
Thanks, and sorry again for the wrong issue (I will remove the SeaBIOS mailing list in the future mails)...
This is still in the SeaBIOS realm. (Though I'm thinking it's really a C-DOS bug.)
-Kevin
On Wed, May 09, 2012 at 08:44:45PM -0400, Kevin O'Connor wrote:
On Wed, May 09, 2012 at 08:57:53AM -0700, Alain Ribière wrote:
I wonder what is the CDOS system at 0D4000... It seems a bit strange.
Very odd. This is still controlled by seabios. Indeed, if I apply this (clearly incorrect) patch to seabios:
diff --git a/src/shadow.c b/src/shadow.c index c0c8cc2..73c0670 100644 --- a/src/shadow.c +++ b/src/shadow.c
[...]
Just for kicks, I tried a simple hack to try and reserve more memory in the c/d/e-segments, and it seems to work for the image you posted. Run:
python -c 'import sys; sys.stdout.write("\x55\xaa\xc8\xcb\x6e" + "\x00"*(100*1024 - 5))' > dummy.rom
and then add "-option-rom dummy.rom" to your qemu command line. (The idea is to build a 100K dummy option rom that seabios will load and mark as read-only memory.)
-Kevin
Both solutions work for my application ! Thanks a lot. I think I will use one of them, even if I know it's not really "clean".
Thanks again,
Alain
________________________________
On Wed, May 09, 2012 at 08:44:45PM -0400, Kevin O'Connor wrote:
On Wed, May 09, 2012 at 08:57:53AM -0700, Alain Ribière wrote:
I wonder what is the CDOS system at 0D4000... It seems a bit strange.
Very odd. This is still controlled by seabios. Indeed, if I apply this (clearly incorrect) patch to seabios:
diff --git a/src/shadow.c b/src/shadow.c index c0c8cc2..73c0670 100644 --- a/src/shadow.c +++ b/src/shadow.c
[...]
Just for kicks, I tried a simple hack to try and reserve more memory in the c/d/e-segments, and it seems to work for the image you posted. Run:
python -c 'import sys; sys.stdout.write("\x55\xaa\xc8\xcb\x6e" + "\x00"*(100*1024 - 5))' > dummy.rom
and then add "-option-rom dummy.rom" to your qemu command line. (The idea is to build a 100K dummy option rom that seabios will load and mark as read-only memory.)
-Kevin
-
Alain Ribière -
Kevin O'Connor -
Peter Stuge