gfx_passthru: warning when vgabios rom has invalid checksum
RFC PATCH 1/1 gfx_passthru: warning when vgabios rom has invalid checksum
Hi,
According to the "PCI Firmware Spec Rev 3.0",
the system firmware have to write the new checksum after resizing the expansion ROM area.
So, when re-using the shadowed VGABIOS, the checksum must be valid, and if it is invalid, we should consider the BIOS has a bug or memory area is corrupted.
This patch just add the warning message when checksum requires recalculation.
I tried this following environments.
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?
Regards,
Noboru.
hw/pass-through.c +3 -0
@@ -4262,7 +4262,10 @@ static int setup_vga_pt(void)for ( c = (char *)bios; c < ((char *)bios + bios_size); c++ ) checksum += *c;if ( checksum ) + { bios[bios_size -1 ] -= checksum; +PT_LOG ("vga bios checksum is adjusted!\n" ); + }cpu_physical_memory_rw (0xc0000 , bios, bios_size,1 );
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Quoted 10 lines · Noboru Iwamatsu
Hi,
According to the "PCI Firmware Spec Rev 3.0",
the system firmware have to write the new checksum after resizing the expansion ROM area.So, when re-using the shadowed VGABIOS, the checksum must be valid, and if it is invalid, we should consider the BIOS has a bug or memory area is corrupted.
Quoted 9 lines
This patch just add the warning message when checksum requires recalculation.
I tried this following environments.
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.
What does "checksum is bad" mean here?
Quoted 4 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?
Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
Regards,
Weidong
Quoted 3 lines
Regards,
Noboru.
Hi,
Quoted 5 lines
Quoted 4 lines
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
It means shadowed VGABIOS has invalid checksum
and requires recalculating.
Quoted 6 lines
Quoted 4 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
When the checksum is bad,
if it is not recalculated, guest has some trouble about handling the VGA (unable to show BIOS screen, unable to load the gfx driver properly, ...).
If it is recalculated, it seems to work fine.
Regards,
Noboru.
Quoted 34 lines
Quoted 10 lines · Noboru Iwamatsu
Hi,
According to the "PCI Firmware Spec Rev 3.0",
the system firmware have to write the new checksum after resizing the expansion ROM area.So, when re-using the shadowed VGABIOS, the checksum must be valid, and if it is invalid, we should consider the BIOS has a bug or memory area is corrupted.
Quoted 9 lines
This patch just add the warning message when checksum requires recalculation.
I tried this following environments.
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
Quoted 4 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
Regards,
WeidongQuoted 3 lines
Regards,
Noboru.
Quoted 15 lines · Noboru Iwamatsu
Hi,
Quoted 5 lines
Quoted 4 lines
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
It means shadowed VGABIOS has invalid checksum
and requires recalculating.Quoted 4 lines
Quoted 4 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?
Quoted 7 lines
Quoted 2 lines
Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
When the checksum is bad,
if it is not recalculated, guest has some trouble about handling the VGA (unable to show BIOS screen, unable to load the gfx driver properly, ...).
For this case, it's reasonable to warning and stop loading rom. We didn't encounter this issue.
Quoted 2 lines
If it is recalculated, it seems to work fine.
How did you recalculate it? is it possible to recalculate it when detect invalid checksum?
Regards,
Weidong
Quoted 46 lines
Regards,
Noboru.Quoted 39 lines
Quoted 10 lines · Noboru Iwamatsu
Hi,
According to the "PCI Firmware Spec Rev 3.0",
the system firmware have to write the new checksum after resizing the expansion ROM area.So, when re-using the shadowed VGABIOS, the checksum must be valid, and if it is invalid, we should consider the BIOS has a bug or memory area is corrupted.
Quoted 10 lines
This patch just add the warning message when checksum requires recalculation.
I tried this following environments.
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
Quoted 5 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
Regards,
WeidongQuoted 4 lines
Regards,
Noboru.
Hi Weidong,
Quoted 2 lines
How did you recalculate it? is it possible to recalculate it when detect invalid checksum?
Checking and recalculating the VGABIOS checksum is being processed in setup_vga_pt() in hw/pass-through.c.
I thought you have written that part.
If the checksum is invalid, current setup_vga_pt() always overwrites it with the recalculated value.
What I mention is: the VGABIOS checksum must have been already recalculated by system BIOS. So, when setup_vga_pt() detects the invalid one, something might be wrong.
But, I'm not sure all BIOS should work as described in PCIFW Spec. Actually, my Q35 had invalid checksum.
I attach another patch:
This stops loading vgabios if the checksum is wrong.
Quoted 72 lines
Quoted 15 lines · Noboru Iwamatsu
Hi,
Quoted 5 lines
Quoted 4 lines
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
It means shadowed VGABIOS has invalid checksum
and requires recalculating.Quoted 4 lines
Quoted 4 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Quoted 7 lines
Quoted 2 lines
Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
When the checksum is bad,
if it is not recalculated, guest has some trouble about handling the VGA (unable to show BIOS screen, unable to load the gfx driver properly, ...).For this case, it's reasonable to warning and stop loading rom. We didn't encounter this issue.
Quoted 1 line
If it is recalculated, it seems to work fine.
How did you recalculate it? is it possible to recalculate it when detect invalid checksum?
Regards,
WeidongQuoted 38 lines
Regards,
Noboru.Quoted 33 lines
Quoted 10 lines · Noboru Iwamatsu
Hi,
According to the "PCI Firmware Spec Rev 3.0",
the system firmware have to write the new checksum after resizing the expansion ROM area.So, when re-using the shadowed VGABIOS, the checksum must be valid, and if it is invalid, we should consider the BIOS has a bug or memory area is corrupted.
Quoted 9 lines
This patch just add the warning message when checksum requires recalculation.
I tried this following environments.
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
Quoted 4 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
Regards,
WeidongQuoted 3 lines
Regards,
Noboru.
Quoted 4 lines
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
hw/pass-through.c +6 -2
@@ -4258,11 +4258,15 @@ static int setup_vga_pt(void) goto out; } -/* Adjust the bios checksum */ +/* Validate the bios checksum */ for ( c = (char *)bios; c < ((char *)bios + bios_size); c++ ) checksum += *c;if ( checksum ) - bios[bios_size -1 ] -= checksum; + { +PT_LOG ("vga bios checksum is invalid!\n" ); + rc =-1 ; + goto out; + }cpu_physical_memory_rw (0xc0000 , bios, bios_size,1 );
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Quoted 12 lines · Noboru Iwamatsu
Hi Weidong,
Quoted 2 lines
How did you recalculate it? is it possible to recalculate it when detect invalid checksum?
Checking and recalculating the VGABIOS checksum is being processed in
setup_vga_pt()inhw/pass-through.c.
I thought you have written that part.If the checksum is invalid, current
setup_vga_pt()always overwrites it with the recalculated value.
Quoted 7 lines
What I mention is: the VGABIOS checksum must have been already recalculated by system BIOS. So, when
setup_vga_pt()detects the invalid one, something might be wrong.But, I'm not sure all BIOS should work as described in PCIFW Spec. Actually, my Q35 had invalid checksum.
Now I understand. Because your Q35 works with recalculated checksum, I prefer to only add a warning message, and continue to load rom for gfx passthru.
Regards,
Weidong
Quoted 99 lines
I attach another patch:
This stops loading vgabios if the checksum is wrong.Signed-off-by Noboru Iwamatsu <[email protected]>Quoted 90 lines
Quoted 19 lines · Noboru Iwamatsu
Hi,
Quoted 7 lines
Quoted 5 lines
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
It means shadowed VGABIOS has invalid checksum
and requires recalculating.Quoted 5 lines
Quoted 5 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Quoted 8 lines
Quoted 3 lines
Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
When the checksum is bad,
if it is not recalculated, guest has some trouble about handling the VGA (unable to show BIOS screen, unable to load the gfx driver properly, ...).For this case, it's reasonable to warning and stop loading rom. We didn't encounter this issue.
Quoted 2 lines
If it is recalculated, it seems to work fine.
How did you recalculate it? is it possible to recalculate it when detect invalid checksum?
Regards,
WeidongQuoted 44 lines
Regards,
Noboru.Quoted 39 lines
Quoted 10 lines · Noboru Iwamatsu
Hi,
According to the "PCI Firmware Spec Rev 3.0",
the system firmware have to write the new checksum after resizing the expansion ROM area.So, when re-using the shadowed VGABIOS, the checksum must be valid, and if it is invalid, we should consider the BIOS has a bug or memory area is corrupted.
Quoted 10 lines
This patch just add the warning message when checksum requires recalculation.
I tried this following environments.
Intel DX58SO + GeForce GTS250: checksum is OK.
Intel DX58SO + GeForce 9600GT: checksum is OK.
Fujitsu Q35 M/B + IGD: checksum is bad.What does "checksum is bad" mean here?
Quoted 5 lines
What do you think, Weidong?
Is just warning enough? Or, should we stop loading the rom?Does it still work or not when checksum is bad? If it works, we should not stop loading the rom.
Regards,
WeidongQuoted 4 lines
Regards,
Noboru._______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Weidong Han writes ("Re: [Xen-devel] Re: [PATCH][RFC] gfx_passthru: warning when vgabios rom has invalid checksum"):
Quoted 3 lines
Now I understand. Because your Q35 works with recalculated checksum, I prefer to only add a warning message, and continue to load rom for gfx passthru.
Having read this thread I'm still a bit confused. The problem is that the VGA BIOS on the graphics card is broken and has a broken checksum, and the proposed workaround is to recalculate the checksum for the benefit of the guest ?
Does this incorrectly checksummed BIOS work natively (ie without passthrough) and if so why is passthrough different ? Alternatively if it doesn't work native why are we trying to make it work with passthrough ?
On another level, Weidong, are you suggesting you'd like to see Noboru produce a different patch which just produces a warning ?
Thanks,
Ian.
Hi,
Quoted 9 lines
Weidong Han writes ("Re: [Xen-devel] Re: [PATCH][RFC]
gfx_passthru: warning when vgabios rom has invalid checksum"):Quoted 3 lines
Now I understand. Because your Q35 works with recalculated checksum, I prefer to only add a warning message, and continue to load rom for gfx passthru.
Having read this thread I'm still a bit confused. The problem is that the VGA BIOS on the graphics card is broken and has a broken checksum, and the proposed workaround is to recalculate the checksum for the benefit of the guest ?
In the native environment, the VGABIOS, the expansion ROM on the graphics card, is placed into the 0C0000h address space, and then executed. Of course, the checksum of the ROM must be valid.
After this initialization, the system BIOS, the actual BIOS of the M/B,
can resize the expansion ROM code to reduce the amount of occupied space. If the system BIOS resizes it, a new checksum must be calculated and stored in the ROM image that is on the RAM.
So, normally, shadowed VGABIOS, that is placed in 0C0000h, is already modified and its checksum must be recalculated.
Qemu-dm copies 0C0000h's contents of the dom0 to guest's 0C0000h. Guest re-uses dom0's used-up VGABIOS.
The problem that I mentioned is about this recalculated checksum.
System BIOS must guarantee the checksum after the resizing, but, some M/B does not.
However, after adjusting the checksum, guest seems to work, and current qemu-dm does so. The buggy system BIOS might just forgets to recalculate.
Should we check strictly here?
Quoted 7 lines
Does this incorrectly checksummed BIOS work natively (ie without passthrough) and if so why is passthrough different ? Alternatively if it doesn't work native why are we trying to make it work with passthrough ?
On another level, Weidong, are you suggesting you'd like to see Noboru produce a different patch which just produces a warning ?
I sent "just warning" patch on the first of this thread. I resend it.
Noboru.
hw/pass-through.c +3 -0
@@ -4262,7 +4262,10 @@ static int setup_vga_pt(void)for ( c = (char *)bios; c < ((char *)bios + bios_size); c++ ) checksum += *c;if ( checksum ) + { bios[bios_size -1 ] -= checksum; +PT_LOG ("vga bios checksum is adjusted!\n" ); + }cpu_physical_memory_rw (0xc0000 , bios, bios_size,1 );
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Quoted 45 lines · Noboru Iwamatsu
Hi,
Quoted 11 lines
Weidong Han writes ("Re: [Xen-devel] Re: [PATCH][RFC]
gfx_passthru: warning when vgabios rom has invalid checksum"):Quoted 4 lines
Now I understand. Because your Q35 works with recalculated checksum, I prefer to only add a warning message, and continue to load rom for gfx passthru.
Having read this thread I'm still a bit confused. The problem is that the VGA BIOS on the graphics card is broken and has a broken checksum, and the proposed workaround is to recalculate the checksum for the benefit of the guest ?
In the native environment, the VGABIOS, the expansion ROM on the graphics card, is placed into the 0C0000h address space, and then executed. Of course, the checksum of the ROM must be valid.
After this initialization, the system BIOS, the actual BIOS of the M/B,
can resize the expansion ROM code to reduce the amount of occupied space. If the system BIOS resizes it, a new checksum must be calculated and stored in the ROM image that is on the RAM.So, normally, shadowed VGABIOS, that is placed in 0C0000h, is already modified and its checksum must be recalculated.
Qemu-dm copies 0C0000h's contents of the dom0 to guest's 0C0000h. Guest re-uses dom0's used-up VGABIOS.
The problem that I mentioned is about this recalculated checksum.
System BIOS must guarantee the checksum after the resizing, but, some M/B does not.
However, after adjusting the checksum, guest seems to work, and current qemu-dm does so. The buggy system BIOS might just forgets to recalculate.Should we check strictly here?
Quoted 5 lines
Does this incorrectly checksummed BIOS work natively (ie without passthrough) and if so why is passthrough different ? Alternatively if it doesn't work native why are we trying to make it work with passthrough ?
Quoted 7 lines
Quoted 3 lines
On another level, Weidong, are you suggesting you'd like to see Noboru produce a different patch which just produces a warning ?
I sent "just warning" patch on the first of this thread. I resend it.
Yes, I ack this one. Acked-by: Weidong Han <[email protected]>
Regards,
Weidong
Quoted 6 lines
Noboru.
Signed-off-by Noboru Iwamatsu <[email protected]>
Hi,
Generally, we should validate the checksum right.
Weidong, have you seen the checksum adjusting?
I only saw it in Q35+IGD.
You know my Q35 is buggy. If this is the only issue of my Q35, 2nd patch (validate and disable) is the best way, I think.
Or, is this a IGD specific issue?
If so, checksum adjusting should be a workaround for IGD pass-through. I assume IGD's VGABIOS is included in the system BIOS and might be expanded to the RAM in a different way from the PCI gfx card.
I attach 3rd patch:
When invalid checksum is detected, adjust it if igd_passthru is enabled. In other case, stop loading.
I tried this on my Q35+IGD, the checksum has been adjusted and worked.
Do you have any idea?
Regards,
Noboru.
Quoted 46 lines
Quoted 45 lines · Noboru Iwamatsu
Hi,
Quoted 9 lines
Weidong Han writes ("Re: [Xen-devel] Re: [PATCH][RFC]
gfx_passthru:
warning when vgabios rom has invalid checksum"):Quoted 3 lines
Now I understand. Because your Q35 works with recalculated checksum, I prefer to only add a warning message, and continue to load rom for gfx passthru.
Having read this thread I'm still a bit confused. The problem is that the VGA BIOS on the graphics card is broken and has a broken checksum, and the proposed workaround is to recalculate the checksum for the benefit of the guest ?
In the native environment, the VGABIOS, the expansion ROM on the graphics card, is placed into the 0C0000h address space, and then executed. Of course, the checksum of the ROM must be valid.
After this initialization, the system BIOS, the actual BIOS of the M/B, can resize the expansion ROM code to reduce the amount of occupied space. If the system BIOS resizes it, a new checksum must be calculated and stored in the ROM image that is on the RAM.
So, normally, shadowed VGABIOS, that is placed in 0C0000h, is already modified and its checksum must be recalculated.
Qemu-dm copies 0C0000h's contents of the dom0 to guest's 0C0000h. Guest re-uses dom0's used-up VGABIOS.
The problem that I mentioned is about this recalculated checksum.
System BIOS must guarantee the checksum after the resizing, but, some M/B does not.
However, after adjusting the checksum, guest seems to work, and current qemu-dm does so. The buggy system BIOS might just forgets to recalculate.Should we check strictly here?
Quoted 7 lines
Does this incorrectly checksummed BIOS work natively (ie without passthrough) and if so why is passthrough different ? Alternatively if it doesn't work native why are we trying to make it work with passthrough ?
On another level, Weidong, are you suggesting you'd like to see Noboru produce a different patch which just produces a warning ?
Quoted 12 lines
Quoted 2 lines
I sent "just warning" patch on the first of this thread. I resend it.
Yes, I ack this one. Acked-by: Weidong Han <[email protected]>
Regards,
WeidongQuoted 5 lines
Noboru.
Signed-off-by Noboru Iwamatsu <[email protected]>
hw/pass-through.c +14 -2
@@ -4258,11 +4258,23 @@ static int setup_vga_pt(void) goto out; } -/* Adjust the bios checksum */ +/* Validate the bios checksum */ for ( c = (char *)bios; c < ((char *)bios + bios_size); c++ ) checksum += *c;if ( checksum ) - bios[bios_size -1 ] -= checksum; + { +if ( igd_passthru ) + { + bios[bios_size -1 ] -= checksum; +PT_LOG ("vga bios checksum is adjusted!\n" ); + } +else + { +PT_LOG ("vga bios checksum is invalid!\n" ); + rc =-1 ; + goto out; + } + }cpu_physical_memory_rw (0xc0000 , bios, bios_size,1 );
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel
Quoted 6 lines · Noboru Iwamatsu
Hi,
Generally, we should validate the checksum right.
Weidong, have you seen the checksum adjusting?
Quoted 9 lines
I only saw it in Q35+IGD.
You know my Q35 is buggy. If this is the only issue of my Q35, 2nd patch (validate and disable) is the best way, I think.Or, is this a IGD specific issue?
If so, checksum adjusting should be a workaround for IGD pass-through. I assume IGD's VGABIOS is included in the system BIOS and might be expanded to the RAM in a different way from the PCI gfx card.
I also found the checksum is invalid on my Q45 platform, of course it works fine with recalculated value.
Quoted 7 lines
I attach 3rd patch:
When invalid checksum is detected, adjust it ifigd_passthruis enabled. In other case, stop loading.I tried this on my Q35+IGD, the checksum has been adjusted and worked.
Do you have any idea?
I still prefer to only add a warning message about it, and replace it with recalculated checksum for both IGD and discrete gfx. There might be some discrete gfx cards which also have invalid checksum, and it's likely that it still works with recalculated checksum in guest.
Regards,
Weidong
Quoted 74 lines
Regards,
Noboru.Signed-off-by Noboru Iwamatsu <[email protected]>Quoted 65 lines
Quoted 46 lines · Noboru Iwamatsu
Hi,
Quoted 12 lines
Weidong Han writes ("Re: [Xen-devel] Re: [PATCH][RFC]
gfx_passthru:
warning when vgabios rom has invalid checksum"):Quoted 4 lines
Now I understand. Because your Q35 works with recalculated checksum, I prefer to only add a warning message, and continue to load rom for gfx passthru.
Having read this thread I'm still a bit confused. The problem is that the VGA BIOS on the graphics card is broken and has a broken checksum, and the proposed workaround is to recalculate the checksum for the benefit of the guest ?
In the native environment, the VGABIOS, the expansion ROM on the graphics card, is placed into the 0C0000h address space, and then executed. Of course, the checksum of the ROM must be valid.
After this initialization, the system BIOS, the actual BIOS of the M/B, can resize the expansion ROM code to reduce the amount of occupied space. If the system BIOS resizes it, a new checksum must be calculated and stored in the ROM image that is on the RAM.
So, normally, shadowed VGABIOS, that is placed in 0C0000h, is already modified and its checksum must be recalculated.
Qemu-dm copies 0C0000h's contents of the dom0 to guest's 0C0000h. Guest re-uses dom0's used-up VGABIOS.
The problem that I mentioned is about this recalculated checksum.
System BIOS must guarantee the checksum after the resizing, but, some M/B does not.
However, after adjusting the checksum, guest seems to work, and current qemu-dm does so. The buggy system BIOS might just forgets to recalculate.Should we check strictly here?
Quoted 5 lines
Does this incorrectly checksummed BIOS work natively (ie without passthrough) and if so why is passthrough different ? Alternatively if it doesn't work native why are we trying to make it work with passthrough ?
Quoted 6 lines
Quoted 3 lines
On another level, Weidong, are you suggesting you'd like to see Noboru produce a different patch which just produces a warning ?
I sent "just warning" patch on the first of this thread. I resend it.
Yes, I ack this one. Acked-by: Weidong Han <[email protected]>
Regards,
WeidongQuoted 6 lines
Noboru.
Signed-off-by Noboru Iwamatsu <[email protected]>