* Re: fd0 not found
2005-07-26 10:58 fd0 not found Kent Tong
@ 2005-07-26 23:33 ` Kumba
2005-07-26 23:46 ` David S. Miller
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Kumba @ 2005-07-26 23:33 UTC (permalink / raw)
To: sparclinux
Kent Tong wrote:
> Hi,
>
> I'm running Debian Sarge (kernel 2.6) on a Sun Blade 100.
> I tried to to format my floppy using "fdformat /dev/fd0h1440"
> but it says "no such device or address". In addition, mesg
> shows there is an error:
>
> SABRE0: Uncorrectable Error, primary error type[DMA Write:Translation
> Error]
> SABRE0: bytemask[00ff] dword_offset[0] was_block(0)
> SABRE0: UE AFAR [000000005fa7c000]
> SABRE0: UE Secondary errors [(none)]
> SABRE0: IOMMU Error, type[Invalid Error]
> SABRE0: IOMMU TAG(4)[RAW(0000000000c6f800)error(Invalid
> Error)wr(0)sz(8K)vpg(df0
> 00000)]
> SABRE0: IOMMU
> DATA(4)[RAW(000000006fe00000)valid(1)used(1)cache(0)ppg(0000000000
> 000000)
>
> and it doesn't mention anything about floppy or fd at all.
> Any idea?
The Floppy drive on Sun Blade 100 has never worked. I might be off on this, but
I believe it hangs off of an ISA bus built into the system (or some bus), and no
one's ever invested the time to detect it.
--Kumba
--
Gentoo/MIPS Team Lead
Gentoo Foundation Board of Trustees
"Such is oft the course of deeds that move the wheels of the world: small hands
do them because they must, while the eyes of the great are elsewhere." --Elrond
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: fd0 not found
2005-07-26 10:58 fd0 not found Kent Tong
2005-07-26 23:33 ` Kumba
@ 2005-07-26 23:46 ` David S. Miller
2005-07-29 2:16 ` Kumba
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: David S. Miller @ 2005-07-26 23:46 UTC (permalink / raw)
To: sparclinux
From: Kumba <kumba@gentoo.org>
Date: Tue, 26 Jul 2005 19:33:52 -0400
> The Floppy drive on Sun Blade 100 has never worked. I might be off
> on this, but I believe it hangs off of an ISA bus built into the
> system (or some bus), and no one's ever invested the time to detect
> it.
It's not a matter of detection, it's working around the brain damaged
design of this thing.
The floppy device wants to do ISA DMA, but on the ISA bus it is
illegal for other agents to issue ISA I/O port operations while a DMA
is in progress. The ISA bus has a signal that the processor is
supposed to use to determine when DMA is in progress and thus I/O port
operations cannot be initiated, but there is no hookup for this to the
UltraSPARC chip in the SB100.
Thus the only way to handle this is to make the current thread of
execution sleep if it tries to do an I/O port access while a floppy
DMA is in progress, and since we can't sleep in interrupt handler
context there is no way to implement this awefully gross workaround
for hardware mis-design.
Just get a USB memory stick if you want small portable storage.
The floppy device on the SB100 is unlikely to ever be supported.
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: fd0 not found
2005-07-26 10:58 fd0 not found Kent Tong
2005-07-26 23:33 ` Kumba
2005-07-26 23:46 ` David S. Miller
@ 2005-07-29 2:16 ` Kumba
2005-07-29 3:01 ` David S. Miller
2005-07-29 5:11 ` Kumba
4 siblings, 0 replies; 6+ messages in thread
From: Kumba @ 2005-07-29 2:16 UTC (permalink / raw)
To: sparclinux
David S. Miller wrote:
>
> It's not a matter of detection, it's working around the brain damaged
> design of this thing.
>
> The floppy device wants to do ISA DMA, but on the ISA bus it is
> illegal for other agents to issue ISA I/O port operations while a DMA
> is in progress. The ISA bus has a signal that the processor is
> supposed to use to determine when DMA is in progress and thus I/O port
> operations cannot be initiated, but there is no hookup for this to the
> UltraSPARC chip in the SB100.
>
> Thus the only way to handle this is to make the current thread of
> execution sleep if it tries to do an I/O port access while a floppy
> DMA is in progress, and since we can't sleep in interrupt handler
> context there is no way to implement this awefully gross workaround
> for hardware mis-design.
>
> Just get a USB memory stick if you want small portable storage.
> The floppy device on the SB100 is unlikely to ever be supported.
I don't suppose any of the kludges found in OpenSolaris code are portable into
linux sources to address some of the design flaws/oddities? (Portable meaning
license friendly, etc..)
--Kumba
--
Gentoo/MIPS Team Lead
Gentoo Foundation Board of Trustees
"Such is oft the course of deeds that move the wheels of the world: small hands
do them because they must, while the eyes of the great are elsewhere." --Elrond
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: fd0 not found
2005-07-26 10:58 fd0 not found Kent Tong
` (2 preceding siblings ...)
2005-07-29 2:16 ` Kumba
@ 2005-07-29 3:01 ` David S. Miller
2005-07-29 5:11 ` Kumba
4 siblings, 0 replies; 6+ messages in thread
From: David S. Miller @ 2005-07-29 3:01 UTC (permalink / raw)
To: sparclinux
From: Kumba <kumba@gentoo.org>
Date: Thu, 28 Jul 2005 22:16:28 -0400
> I don't suppose any of the kludges found in OpenSolaris code are
> portable into linux sources to address some of the design
> flaws/oddities? (Portable meaning license friendly, etc..)
Solaris uses interrupt threads, which can thus block in interrupt
context, which is how they are able to kludge around this.
So they just make the interrupt thread block when an I/O port
access is attempted on any of the ISA bus devices while a DMA
transfer is running.
Linux doesn't have interrupt threads, so we can't deal with it that
way.
Furthermore, anyone planning to "learn something" from opensolaris and
then incorporate such things into Linux had better well explicitly
implement and properly document a closed room implementation where one
person reads the OpenSolaris code, explains what is happening to
another person, who in turn implements the Linux version.
Otherwise there are grave contamination and copyright infringement
issues possible. One would, in addition to the properly documented
closed room implementation, need to do a patent check on the ideas
being copied over as well.
As a result of all that, looking to OpenSolaris for information to
help out the Linux sparc port is a bad idea.
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: fd0 not found
2005-07-26 10:58 fd0 not found Kent Tong
` (3 preceding siblings ...)
2005-07-29 3:01 ` David S. Miller
@ 2005-07-29 5:11 ` Kumba
4 siblings, 0 replies; 6+ messages in thread
From: Kumba @ 2005-07-29 5:11 UTC (permalink / raw)
To: sparclinux
David S. Miller wrote:
>
> Furthermore, anyone planning to "learn something" from opensolaris and
> then incorporate such things into Linux had better well explicitly
> implement and properly document a closed room implementation where one
> person reads the OpenSolaris code, explains what is happening to
> another person, who in turn implements the Linux version.
>
> Otherwise there are grave contamination and copyright infringement
> issues possible. One would, in addition to the properly documented
> closed room implementation, need to do a patent check on the ideas
> being copied over as well.
>
> As a result of all that, looking to OpenSolaris for information to
> help out the Linux sparc port is a bad idea.
I assumed this was the case likely. Was more a mild curiosity on what the
feasbility of that was (And whether it's even remotely possible). I've already
peered at OS code out of boredem and in doing so, I noticed they had a
kludge/hack in place for that IDE DMA issue that's always plagued the b100
series, so it piqued my interest.
--Kumba
--
Gentoo/MIPS Team Lead
Gentoo Foundation Board of Trustees
"Such is oft the course of deeds that move the wheels of the world: small hands
do them because they must, while the eyes of the great are elsewhere." --Elrond
^ permalink raw reply [flat|nested] 6+ messages in thread