2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andy Isaacson]

The PCMCIA slot on my Thinkpad X40 stopped working sometime between
2.6.17-rc4-mm3 and 2.6.17-rc5-mm3, and is still not working as of
2.6.17-rc6-mm2.  There's one diff early in dmesg that looks interesting:

 ACPI: PCI Interrupt 0000:00:1f.6[B] -> GSI 17 (level, low) -> IRQ 177
 PCI: Setting latency timer of device 0000:00:1f.6 to 64
 MC'97 1 converters and GPIO not ready (0xff00)
 PM: Adding info for ac97:1-1:unknown codec
-ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 16 (level, low) -> IRQ 169
 Yenta: CardBus bridge found at 0000:02:00.0 [1014:0555]
-Yenta: ISA IRQ mask 0x04b8, PCI irq 169
+Yenta: ISA IRQ mask 0x0cb8, PCI irq 169
 Socket status: 30000006
 pcmcia: parent PCI bridge I/O window: 0x3000 - 0x7fff
 cs: IO port probe 0x3000-0x7fff: clean.
 pcmcia: parent PCI bridge Memory window: 0xd0200000 - 0xdfffffff
 pcmcia: parent PCI bridge Memory window: 0xf0000000 - 0xf7ffffff

Then when I insert the card, instead of

 pccard: PCMCIA card inserted into slot 0
 cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
 cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff
    0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff
    0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff
    0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff
    0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff
    0xdfa00000-0xe01fffff
 pcmcia: registering new device pcmcia0.0
 PM: Adding info for pcmcia:0.0
-Probing IDE interface ide2...
-hde: CF Card, CFA DISK drive
-PM: Adding info for No Bus:ide2
-hdf: probing with STATUS(0x50) instead of ALTSTATUS(0x0a)

I get

 pccard: PCMCIA card inserted into slot 0
 cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
 cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff
     0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff
     0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff
     0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff
     0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff
     0xdfa00000-0xe01fffff
 pcmcia: registering new device pcmcia0.0
 PM: Adding info for pcmcia:0.0
+ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
+ide2: ports already in use, skipping probe
+ide2: I/O resource 0xF8A8A01E-0xF8A8A01E not free.
+ide2: ports already in use, skipping probe
+ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
+ide2: ports already in use, skipping probe

With 2.6.17-rc6-mm2 I have

           CPU0       
  0:     188343   IO-APIC-edge     timer
  1:       2685   IO-APIC-edge     i8042
  9:        393   IO-APIC-fasteoi  acpi
 12:      17840   IO-APIC-edge     i8042
 14:       4616   IO-APIC-edge     ide0
169:          1   IO-APIC-fasteoi  uhci_hcd:usb1, yenta
177:          0   IO-APIC-fasteoi  Intel 82801DB-ICH4, Intel 82801DB-ICH4 Modem
185:          0   IO-APIC-fasteoi  uhci_hcd:usb3
193:          0   IO-APIC-fasteoi  uhci_hcd:usb2
201:          0   IO-APIC-fasteoi  ehci_hcd:usb4
209:       2524   IO-APIC-fasteoi  eth0
217:      43647   IO-APIC-fasteoi  ipw2200
NMI:          0 
LOC:     175985 
ERR:          0
MIS:          0

The card I'm trying to use is a "PQI Compact Flash Adapter", which
appears to just be some wires connecting the PCMCIA pins to the CF pins.
The CF card is a "TwinMOS Ultra-X 140X 1GB" card, part number
FCF1GBUPS-K W543F5TM0832.

dmesgs at

http://web.hexapodia.org/~adi/tmp/dmesg.2.6.17-rc4-mm3.txt
http://web.hexapodia.org/~adi/tmp/dmesg.2.6.17-rc6-mm2.txt

(although the rc4-mm3 one is truncated, sorry.)

-andy

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andrew Morton]

On Thu, 15 Jun 2006 09:28:59 -0700
Andy Isaacson <adi@hexapodia.org> wrote:

> The PCMCIA slot on my Thinkpad X40 stopped working sometime between
> 2.6.17-rc4-mm3 and 2.6.17-rc5-mm3, and is still not working as of
> 2.6.17-rc6-mm2.  There's one diff early in dmesg that looks interesting:
> 
>  ACPI: PCI Interrupt 0000:00:1f.6[B] -> GSI 17 (level, low) -> IRQ 177
>  PCI: Setting latency timer of device 0000:00:1f.6 to 64
>  MC'97 1 converters and GPIO not ready (0xff00)
>  PM: Adding info for ac97:1-1:unknown codec
> -ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 16 (level, low) -> IRQ 169
>  Yenta: CardBus bridge found at 0000:02:00.0 [1014:0555]
> -Yenta: ISA IRQ mask 0x04b8, PCI irq 169
> +Yenta: ISA IRQ mask 0x0cb8, PCI irq 169
>  Socket status: 30000006
>  pcmcia: parent PCI bridge I/O window: 0x3000 - 0x7fff
>  cs: IO port probe 0x3000-0x7fff: clean.
>  pcmcia: parent PCI bridge Memory window: 0xd0200000 - 0xdfffffff
>  pcmcia: parent PCI bridge Memory window: 0xf0000000 - 0xf7ffffff
> 
> Then when I insert the card, instead of
> 
>  pccard: PCMCIA card inserted into slot 0
>  cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
>  cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff
>     0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff
>     0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff
>     0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff
>     0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff
>     0xdfa00000-0xe01fffff
>  pcmcia: registering new device pcmcia0.0
>  PM: Adding info for pcmcia:0.0
> -Probing IDE interface ide2...
> -hde: CF Card, CFA DISK drive
> -PM: Adding info for No Bus:ide2
> -hdf: probing with STATUS(0x50) instead of ALTSTATUS(0x0a)
> 
> I get
> 
>  pccard: PCMCIA card inserted into slot 0
>  cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
>  cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff
>      0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff
>      0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff
>      0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff
>      0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff
>      0xdfa00000-0xe01fffff
>  pcmcia: registering new device pcmcia0.0
>  PM: Adding info for pcmcia:0.0
> +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> +ide2: ports already in use, skipping probe
> +ide2: I/O resource 0xF8A8A01E-0xF8A8A01E not free.
> +ide2: ports already in use, skipping probe
> +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> +ide2: ports already in use, skipping probe
> 
> With 2.6.17-rc6-mm2 I have
> 
>            CPU0       
>   0:     188343   IO-APIC-edge     timer
>   1:       2685   IO-APIC-edge     i8042
>   9:        393   IO-APIC-fasteoi  acpi
>  12:      17840   IO-APIC-edge     i8042
>  14:       4616   IO-APIC-edge     ide0
> 169:          1   IO-APIC-fasteoi  uhci_hcd:usb1, yenta
> 177:          0   IO-APIC-fasteoi  Intel 82801DB-ICH4, Intel 82801DB-ICH4 Modem
> 185:          0   IO-APIC-fasteoi  uhci_hcd:usb3
> 193:          0   IO-APIC-fasteoi  uhci_hcd:usb2
> 201:          0   IO-APIC-fasteoi  ehci_hcd:usb4
> 209:       2524   IO-APIC-fasteoi  eth0
> 217:      43647   IO-APIC-fasteoi  ipw2200
> NMI:          0 
> LOC:     175985 
> ERR:          0
> MIS:          0
> 
> The card I'm trying to use is a "PQI Compact Flash Adapter", which
> appears to just be some wires connecting the PCMCIA pins to the CF pins.
> The CF card is a "TwinMOS Ultra-X 140X 1GB" card, part number
> FCF1GBUPS-K W543F5TM0832.
> 
> dmesgs at
> 
> http://web.hexapodia.org/~adi/tmp/dmesg.2.6.17-rc4-mm3.txt
> http://web.hexapodia.org/~adi/tmp/dmesg.2.6.17-rc6-mm2.txt
> 

hm.   I don't know who to blame for this yet ;)

The contents of /proc/ioports on both kernels might be useful.  Let's see
which device+driver is already using those ports, and whether the older
kenrel uses the same addresses.


Thanks.

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andy Isaacson]

On Sat, Jun 17, 2006 at 10:03:27AM -0700, Andrew Morton wrote:
> > The PCMCIA slot on my Thinkpad X40 stopped working sometime between
> > 2.6.17-rc4-mm3 and 2.6.17-rc5-mm3, and is still not working as of
> > 2.6.17-rc6-mm2.
[snip]
> > -Probing IDE interface ide2...
> > -hde: CF Card, CFA DISK drive
> > -PM: Adding info for No Bus:ide2
> > -hdf: probing with STATUS(0x50) instead of ALTSTATUS(0x0a)
> > +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> > +ide2: ports already in use, skipping probe
> > +ide2: I/O resource 0xF8A8A01E-0xF8A8A01E not free.
> > +ide2: ports already in use, skipping probe
> > +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> > +ide2: ports already in use, skipping probe
> 
> hm.   I don't know who to blame for this yet ;)
> 
> The contents of /proc/ioports on both kernels might be useful.  Let's see
> which device+driver is already using those ports, and whether the older
> kenrel uses the same addresses.

In further testing, -rc6 is fine while -rc6-mm2 fails.

Under 2.6.17-rc6 (after having inserted and removed the card, but that
doesn't seem to make much difference) I have

0000-001f : dma1
0020-0021 : pic1
0040-0043 : timer0
0050-0053 : timer1
0060-006f : keyboard
0080-008f : dma page reg
00a0-00a1 : pic2
00c0-00df : dma2
00f0-00ff : fpu
01f0-01f7 : ide0
03c0-03df : vga+
03f6-03f6 : ide0
1000-107f : 0000:00:1f.0
  1000-107f : motherboard
    1000-1003 : PM1a_EVT_BLK
    1004-1005 : PM1a_CNT_BLK
    1008-100b : PM_TMR
    1010-1015 : ACPI CPU throttle
    1020-1020 : PM2_CNT_BLK
    1028-102f : GPE0_BLK
1180-11bf : 0000:00:1f.0
  1180-11bf : motherboard
15e0-15ef : motherboard
1600-167f : motherboard
1680-169f : motherboard
1800-1807 : 0000:00:02.0
1810-181f : 0000:00:1f.1
  1810-1817 : ide0
  1818-181f : ide1
1820-183f : 0000:00:1d.0
  1820-183f : uhci_hcd
1840-185f : 0000:00:1d.1
  1840-185f : uhci_hcd
1860-187f : 0000:00:1d.2
  1860-187f : uhci_hcd
1880-189f : 0000:00:1f.3
18c0-18ff : 0000:00:1f.5
  18c0-18ff : Intel 82801DB-ICH4
1c00-1cff : 0000:00:1f.5
  1c00-1cff : Intel 82801DB-ICH4
2000-207f : 0000:00:1f.6
  2000-207f : Intel 82801DB-ICH4 Modem
2400-24ff : 0000:00:1f.6
  2400-24ff : Intel 82801DB-ICH4 Modem
3000-7fff : PCI Bus #02
  3000-30ff : PCI CardBus #03
  3400-34ff : PCI CardBus #03
  7000-703f : 0000:02:01.0
    7000-703f : e1000

The diff between -rc6 and -rc6-mm2 shows that they have the same ioport
assignment (there's only a textual diff due to ACPI string changes).

I've put just about everything you could want to know about the two
kernels at
http://web.hexapodia.org/~adi/bobble/bobble_2.6.17-rc6_20060620093733/
and
http://web.hexapodia.org/~adi/bobble/bobble_2.6.17-rc6-mm2_20060620094254/

-andy

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andrew Morton]

Andy Isaacson <adi@hexapodia.org> wrote:
>
> On Sat, Jun 17, 2006 at 10:03:27AM -0700, Andrew Morton wrote:
> > > The PCMCIA slot on my Thinkpad X40 stopped working sometime between
> > > 2.6.17-rc4-mm3 and 2.6.17-rc5-mm3, and is still not working as of
> > > 2.6.17-rc6-mm2.
> [snip]
> > > -Probing IDE interface ide2...
> > > -hde: CF Card, CFA DISK drive
> > > -PM: Adding info for No Bus:ide2
> > > -hdf: probing with STATUS(0x50) instead of ALTSTATUS(0x0a)
> > > +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> > > +ide2: ports already in use, skipping probe
> > > +ide2: I/O resource 0xF8A8A01E-0xF8A8A01E not free.
> > > +ide2: ports already in use, skipping probe
> > > +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> > > +ide2: ports already in use, skipping probe
> > 
> > hm.   I don't know who to blame for this yet ;)
> > 
> > The contents of /proc/ioports on both kernels might be useful.  Let's see
> > which device+driver is already using those ports, and whether the older
> > kenrel uses the same addresses.
> 
> In further testing, -rc6 is fine while -rc6-mm2 fails.

OK.

> Under 2.6.17-rc6 (after having inserted and removed the card, but that
> doesn't seem to make much difference) I have
> 
> 0000-001f : dma1
> ...

I should have said iomem, not ioports.  But you have it there on those very
detailed web pages.  There's nothing at 0xF8A8Axxx in either kernel.

> The diff between -rc6 and -rc6-mm2 shows that they have the same ioport
> assignment (there's only a textual diff due to ACPI string changes).

yep.  The only reelvant diff in the rc6 versus rc6-mm2 /proc/iomem is:

@@ -29,7 +29,6 @@
   d0220000-d0220fff : 0000:02:02.0
     d0220000-d0220fff : ipw2200
   d0221000-d02210ff : 0000:02:00.1
-  d1200000-d1200fff : pcmcia_socket0
   d2000000-d3ffffff : PCI CardBus #03
 e0000000-e7ffffff : 0000:00:02.0
 e8000000-efffffff : 0000:00:02.1

whch is to be expected if you removed the card.


> I've put just about everything you could want to know about the two
> kernels at
> http://web.hexapodia.org/~adi/bobble/bobble_2.6.17-rc6_20060620093733/
> and
> http://web.hexapodia.org/~adi/bobble/bobble_2.6.17-rc6-mm2_20060620094254/
> 

It's strange (to me) that IDE is requesting a single-byte memory region. 
Possibly we've broken the resource.c code such that it has some off-by-one
and is now rejecting single-byte requests.

If you're keen , the place to poke around is in
kernel/resource.c:__request_region().   Here's a starting patch:


--- a/kernel/resource.c~a
+++ a/kernel/resource.c
@@ -501,6 +501,10 @@ struct resource * __request_region(struc
 			conflict = __request_resource(parent, res);
 			if (!conflict)
 				break;
+			printk("conflict: %s[%Lx->%Lx]\n",
+				conflict->name,
+				(unsigned long long)conflict->start,
+				(unsigned long long)conflict->end);
 			if (conflict != parent) {
 				parent = conflict;
 				if (!(conflict->flags & IORESOURCE_BUSY))
_

Which should tell us what we're allegedly conflicting with.




Then again, perhaps IDE is broken (as well?), because you don't have any
single-byte iomem regions in your 2.6.17-rc6 /proc/iomem.  It would be
interesting to run this patch in both 2.6.17-rc6 and in 2.6.17-rc6-mm2, see
what it says:

diff -puN drivers/ide/ide.c~a drivers/ide/ide.c
--- 25/drivers/ide/ide.c~a	Tue Jun 20 15:01:32 2006
+++ 25-akpm/drivers/ide/ide.c	Tue Jun 20 15:02:15 2006
@@ -364,6 +364,10 @@ static struct resource* hwif_request_reg
 {
 	struct resource *res = request_region(addr, num, hwif->name);
 
+	if (num == 1) {
+		printk("%s: single-byte request\n", __FUNCTION__);
+		dump_stack();
+	}
 	if (!res)
 		printk(KERN_ERR "%s: I/O resource 0x%lX-0x%lX not free.\n",
 				hwif->name, addr, addr+num-1);
_

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Randy.Dunlap]

On Tue, 20 Jun 2006 15:03:17 -0700 Andrew Morton wrote:

> Andy Isaacson <adi@hexapodia.org> wrote:
> >
> > On Sat, Jun 17, 2006 at 10:03:27AM -0700, Andrew Morton wrote:
> > > > The PCMCIA slot on my Thinkpad X40 stopped working sometime between
> > > > 2.6.17-rc4-mm3 and 2.6.17-rc5-mm3, and is still not working as of
> > > > 2.6.17-rc6-mm2.
> > [snip]
> > > > -Probing IDE interface ide2...
> > > > -hde: CF Card, CFA DISK drive
> > > > -PM: Adding info for No Bus:ide2
> > > > -hdf: probing with STATUS(0x50) instead of ALTSTATUS(0x0a)
> > > > +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> > > > +ide2: ports already in use, skipping probe
> > > > +ide2: I/O resource 0xF8A8A01E-0xF8A8A01E not free.
> > > > +ide2: ports already in use, skipping probe
> > > > +ide2: I/O resource 0xF8A8A00E-0xF8A8A00E not free.
> > > > +ide2: ports already in use, skipping probe
> > > 
> > > hm.   I don't know who to blame for this yet ;)
> > > 
> > > The contents of /proc/ioports on both kernels might be useful.  Let's see
> > > which device+driver is already using those ports, and whether the older
> > > kenrel uses the same addresses.
> > 
> > In further testing, -rc6 is fine while -rc6-mm2 fails.
> 
> OK.
> 
> > Under 2.6.17-rc6 (after having inserted and removed the card, but that
> > doesn't seem to make much difference) I have
> > 
> > 0000-001f : dma1
> > ...
> 
> I should have said iomem, not ioports.  But you have it there on those very
> detailed web pages.  There's nothing at 0xF8A8Axxx in either kernel.
> 
> > The diff between -rc6 and -rc6-mm2 shows that they have the same ioport
> > assignment (there's only a textual diff due to ACPI string changes).
> 
> yep.  The only reelvant diff in the rc6 versus rc6-mm2 /proc/iomem is:
> 
> @@ -29,7 +29,6 @@
>    d0220000-d0220fff : 0000:02:02.0
>      d0220000-d0220fff : ipw2200
>    d0221000-d02210ff : 0000:02:00.1
> -  d1200000-d1200fff : pcmcia_socket0
>    d2000000-d3ffffff : PCI CardBus #03
>  e0000000-e7ffffff : 0000:00:02.0
>  e8000000-efffffff : 0000:00:02.1
> 
> whch is to be expected if you removed the card.
> 
> 
> > I've put just about everything you could want to know about the two
> > kernels at
> > http://web.hexapodia.org/~adi/bobble/bobble_2.6.17-rc6_20060620093733/
> > and
> > http://web.hexapodia.org/~adi/bobble/bobble_2.6.17-rc6-mm2_20060620094254/
> > 
> 
> It's strange (to me) that IDE is requesting a single-byte memory region. 
> Possibly we've broken the resource.c code such that it has some off-by-one
> and is now rejecting single-byte requests.

Looks like ide.c uses single-byte requests just for checking
region availability, but I would think that it should use
full-sized requests for that.


> If you're keen , the place to poke around is in
> kernel/resource.c:__request_region().   Here's a starting patch:
> 
> 
> --- a/kernel/resource.c~a
> +++ a/kernel/resource.c
> @@ -501,6 +501,10 @@ struct resource * __request_region(struc
>  			conflict = __request_resource(parent, res);
>  			if (!conflict)
>  				break;
> +			printk("conflict: %s[%Lx->%Lx]\n",
> +				conflict->name,
> +				(unsigned long long)conflict->start,
> +				(unsigned long long)conflict->end);
>  			if (conflict != parent) {
>  				parent = conflict;
>  				if (!(conflict->flags & IORESOURCE_BUSY))
> _
> 
> Which should tell us what we're allegedly conflicting with.
> 
> 
> 
> 
> Then again, perhaps IDE is broken (as well?), because you don't have any
> single-byte iomem regions in your 2.6.17-rc6 /proc/iomem.  It would be
> interesting to run this patch in both 2.6.17-rc6 and in 2.6.17-rc6-mm2, see
> what it says:
> 
> diff -puN drivers/ide/ide.c~a drivers/ide/ide.c
> --- 25/drivers/ide/ide.c~a	Tue Jun 20 15:01:32 2006
> +++ 25-akpm/drivers/ide/ide.c	Tue Jun 20 15:02:15 2006
> @@ -364,6 +364,10 @@ static struct resource* hwif_request_reg
>  {
>  	struct resource *res = request_region(addr, num, hwif->name);
>  
> +	if (num == 1) {
> +		printk("%s: single-byte request\n", __FUNCTION__);
> +		dump_stack();
> +	}
>  	if (!res)
>  		printk(KERN_ERR "%s: I/O resource 0x%lX-0x%lX not free.\n",
>  				hwif->name, addr, addr+num-1);
> _


linux-2617-pv/drivers/ide> grep hwif_request_reg *.c
ide.c:static struct resource* hwif_request_region(ide_hwif_t *hwif,
ide.c: *        ide_hwif_request_regions - request resources for IDE
ide.c:int ide_hwif_request_regions(ide_hwif_t *hwif)
ide.c:  if (addr && !hwif_request_region(hwif, addr, 1))
ide.c:          if (!hwif_request_region(hwif, addr, 8))
ide.c:          if (!hwif_request_region(hwif, addr, 1)) {

---
~Randy

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andy Isaacson]

On Tue, Jun 20, 2006 at 03:03:17PM -0700, Andrew Morton wrote:
> It's strange (to me) that IDE is requesting a single-byte memory region. 
> Possibly we've broken the resource.c code such that it has some off-by-one
> and is now rejecting single-byte requests.
[snip]
> +++ a/kernel/resource.c
> +			printk("conflict: %s[%Lx->%Lx]\n",
[snip]
> Then again, perhaps IDE is broken (as well?), because you don't have any
> single-byte iomem regions in your 2.6.17-rc6 /proc/iomem.  It would be
> interesting to run this patch in both 2.6.17-rc6 and in 2.6.17-rc6-mm2, see
> what it says:
[snip]
> +++ 25-akpm/drivers/ide/ide.c	Tue Jun 20 15:02:15 2006
> +		printk("%s: single-byte request\n", __FUNCTION__);
> +		dump_stack();

With both patches, rc6-mm2 dumps the following.

[ 2034.052000] pccard: PCMCIA card inserted into slot 0
[ 2034.052000] cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
[ 2034.052000] cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff 0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff 0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff 0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff 0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff 0xdfa00000-0xe01fffff
[ 2034.060000] pcmcia: registering new device pcmcia0.0
[ 2034.060000] PM: Adding info for pcmcia:0.0
[ 2035.976000] conflict: PCI IO[0->ffff]
[ 2035.976000] hwif_request_region: single-byte request for ide2
[ 2035.976000]  [<c0257386>] hwif_request_region+0xa6/0xb0
[ 2035.976000]  [<c02574bf>] ide_hwif_request_regions+0x12f/0x150
[ 2035.976000]  [<c0257177>] init_hwif_default+0x37/0x120
[ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
[ 2035.976000]  [<c025e8a9>] probe_hwif+0x29/0x4f0
[ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
[ 2035.976000]  [<c025ed85>] probe_hwif_init_with_fixup+0x15/0xa0
[ 2035.976000]  [<c0257ea5>] ide_register_hw_with_fixup+0x165/0x1b0
[ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
[ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
[ 2035.976000]  [<f8b07163>] idecs_register+0x83/0xd0 [ide_cs]
[ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
[ 2035.976000]  [<f8b071c0>] outb_mem+0x0/0x10 [ide_cs]
[ 2035.976000]  [<f8b075e4>] ide_config+0x414/0x790 [ide_cs]
[ 2035.976000]  [<f8b071c0>] outb_mem+0x0/0x10 [ide_cs]
[ 2035.976000]  [<f8b0f4c8>] pcmcia_device_probe+0xa8/0x160 [pcmcia]
[ 2035.976000]  [<c025104c>] driver_probe_device+0xbc/0xe0
[ 2035.976000]  [<c02510f0>] __driver_attach+0x0/0x80
[ 2035.976000]  [<c0251160>] __driver_attach+0x70/0x80
[ 2035.976000]  [<c0250369>] bus_for_each_dev+0x69/0x80
[ 2035.976000]  [<c0251195>] driver_attach+0x25/0x30
[ 2035.976000]  [<c02510f0>] __driver_attach+0x0/0x80
[ 2035.976000]  [<c0250968>] bus_add_driver+0x88/0xd0
[ 2035.976000]  [<f8b0b00f>] init_ide_cs+0xf/0x11 [ide_cs]
[ 2035.976000]  [<c0139f32>] sys_init_module+0xf2/0x190
[ 2035.976000]  [<c02e3105>] sysenter_past_esp+0x56/0x79
[ 2035.976000]  [<c02e007b>] inet6_lookup_listener+0x2b/0x100
[ 2035.976000] ide2: I/O resource 0xF8B0200E-0xF8B0200E not free.
[ 2035.976000] ide2: ports already in use, skipping probe
[ 2035.976000] conflict: PCI IO[0->ffff]
[ 2035.976000] hwif_request_region: single-byte request for ide2
[ 2035.976000]  [<c0257386>] hwif_request_region+0xa6/0xb0
[ 2035.976000]  [<c02574bf>] ide_hwif_request_regions+0x12f/0x150
[ 2035.976000]  [<c0257177>] init_hwif_default+0x37/0x120

[snip 1400 lines more]

-andy

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andrew Morton]

On Tue, 20 Jun 2006 23:50:36 -0700
Andy Isaacson <adi@hexapodia.org> wrote:

> On Tue, Jun 20, 2006 at 03:03:17PM -0700, Andrew Morton wrote:
> > It's strange (to me) that IDE is requesting a single-byte memory region. 
> > Possibly we've broken the resource.c code such that it has some off-by-one
> > and is now rejecting single-byte requests.
> [snip]
> > +++ a/kernel/resource.c
> > +			printk("conflict: %s[%Lx->%Lx]\n",
> [snip]
> > Then again, perhaps IDE is broken (as well?), because you don't have any
> > single-byte iomem regions in your 2.6.17-rc6 /proc/iomem.  It would be
> > interesting to run this patch in both 2.6.17-rc6 and in 2.6.17-rc6-mm2, see
> > what it says:
> [snip]
> > +++ 25-akpm/drivers/ide/ide.c	Tue Jun 20 15:02:15 2006
> > +		printk("%s: single-byte request\n", __FUNCTION__);
> > +		dump_stack();
> 
> With both patches, rc6-mm2 dumps the following.
> 
> [ 2034.052000] pccard: PCMCIA card inserted into slot 0
> [ 2034.052000] cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
> [ 2034.052000] cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff 0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff 0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff 0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff 0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff 0xdfa00000-0xe01fffff
> [ 2034.060000] pcmcia: registering new device pcmcia0.0
> [ 2034.060000] PM: Adding info for pcmcia:0.0
> [ 2035.976000] conflict: PCI IO[0->ffff]
> [ 2035.976000] hwif_request_region: single-byte request for ide2
> [ 2035.976000]  [<c0257386>] hwif_request_region+0xa6/0xb0
> [ 2035.976000]  [<c02574bf>] ide_hwif_request_regions+0x12f/0x150
> [ 2035.976000]  [<c0257177>] init_hwif_default+0x37/0x120
> [ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
> [ 2035.976000]  [<c025e8a9>] probe_hwif+0x29/0x4f0
> [ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
> [ 2035.976000]  [<c025ed85>] probe_hwif_init_with_fixup+0x15/0xa0
> [ 2035.976000]  [<c0257ea5>] ide_register_hw_with_fixup+0x165/0x1b0
> [ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
> [ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
> [ 2035.976000]  [<f8b07163>] idecs_register+0x83/0xd0 [ide_cs]
> [ 2035.976000]  [<f8b070b0>] idecs_mmio_fixup+0x0/0x30 [ide_cs]
> [ 2035.976000]  [<f8b071c0>] outb_mem+0x0/0x10 [ide_cs]
> [ 2035.976000]  [<f8b075e4>] ide_config+0x414/0x790 [ide_cs]
> [ 2035.976000]  [<f8b071c0>] outb_mem+0x0/0x10 [ide_cs]
> [ 2035.976000]  [<f8b0f4c8>] pcmcia_device_probe+0xa8/0x160 [pcmcia]
> [ 2035.976000]  [<c025104c>] driver_probe_device+0xbc/0xe0
> [ 2035.976000]  [<c02510f0>] __driver_attach+0x0/0x80
> [ 2035.976000]  [<c0251160>] __driver_attach+0x70/0x80
> [ 2035.976000]  [<c0250369>] bus_for_each_dev+0x69/0x80
> [ 2035.976000]  [<c0251195>] driver_attach+0x25/0x30
> [ 2035.976000]  [<c02510f0>] __driver_attach+0x0/0x80
> [ 2035.976000]  [<c0250968>] bus_add_driver+0x88/0xd0
> [ 2035.976000]  [<f8b0b00f>] init_ide_cs+0xf/0x11 [ide_cs]
> [ 2035.976000]  [<c0139f32>] sys_init_module+0xf2/0x190
> [ 2035.976000]  [<c02e3105>] sysenter_past_esp+0x56/0x79
> [ 2035.976000]  [<c02e007b>] inet6_lookup_listener+0x2b/0x100
> [ 2035.976000] ide2: I/O resource 0xF8B0200E-0xF8B0200E not free.
> [ 2035.976000] ide2: ports already in use, skipping probe

hm.  It appears to have decided that 0 < 0xF8B0200E < 0xffff, which is
clever of it.

Does it help if you set CONFIG_RESOURCES_32BIT?

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andy Isaacson]

On Wed, Jun 21, 2006 at 12:46:30AM -0700, Andrew Morton wrote:
> > [ 2034.060000] pcmcia: registering new device pcmcia0.0
> > [ 2034.060000] PM: Adding info for pcmcia:0.0
> > [ 2035.976000] conflict: PCI IO[0->ffff]
> > [ 2035.976000] hwif_request_region: single-byte request for ide2
> > [ 2035.976000]  [<c0257386>] hwif_request_region+0xa6/0xb0
[snip]
> > [ 2035.976000] ide2: I/O resource 0xF8B0200E-0xF8B0200E not free.
> > [ 2035.976000] ide2: ports already in use, skipping probe
> 
> hm.  It appears to have decided that 0 < 0xF8B0200E < 0xffff, which is
> clever of it.
> 
> Does it help if you set CONFIG_RESOURCES_32BIT?

Nope, same conflict with CONFIG_RESOURCES_32BIT set.  You're right, it
is deciding that 0xF8B0200E conflicts with that range:

conflict: PCI IO[0->ffff] conflicts with ide2[f8b3c00e->f8b3c00e]

Looking at the code, I don't understand how this could have worked in
-rc6; __request_resource hasn't changed, and it says

    167         if (end < start)
    168                 return root;
    169         if (start < root->start)
    170                 return root;
    171         if (end > root->end)
    172                 return root;

If root-> start == 0 and root->end == 0xffff, we should always hit line
172, unless sign extension is in effect... and all the variables are
unsigned long in -rc6, so that doesn't make sense.

Rebooting into -rc6 with some debugging...

-andy

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Andy Isaacson]

On Wed, Jun 21, 2006 at 10:10:48AM -0700, Andy Isaacson wrote:
> > > [ 2035.976000] ide2: I/O resource 0xF8B0200E-0xF8B0200E not free.
> > > [ 2035.976000] ide2: ports already in use, skipping probe
> > 
> > hm.  It appears to have decided that 0 < 0xF8B0200E < 0xffff, which is
> > clever of it.
> > 
> > Does it help if you set CONFIG_RESOURCES_32BIT?
> 
> Nope, same conflict with CONFIG_RESOURCES_32BIT set.  You're right, it
> is deciding that 0xF8B0200E conflicts with that range:
> 
> conflict: PCI IO[0->ffff] conflicts with ide2[f8b3c00e->f8b3c00e]
> 
> Looking at the code, I don't understand how this could have worked in
> -rc6; __request_resource hasn't changed, and it says
> 
>     167         if (end < start)
>     168                 return root;
>     169         if (start < root->start)
>     170                 return root;
>     171         if (end > root->end)
>     172                 return root;
> 
> If root-> start == 0 and root->end == 0xffff, we should always hit line
> 172, unless sign extension is in effect... and all the variables are
> unsigned long in -rc6, so that doesn't make sense.
> 
> Rebooting into -rc6 with some debugging...

So 2.6.17-rc6 plus the conflict-printer patch and the
ide-single-byte-request patch does the following (booted with the
PCMCIA/CF adapter in by accident, and I'm using a different CF card now,
but it seems to have done the right thing):

cs: memory probe 0xf0000000-0xf7ffffff: excluding 0xf0000000-0xf7ffffff
cs: memory probe 0xd0200000-0xdfffffff: excluding 0xd0200000-0xd11fffff 0xd1a00000-0xd41fffff 0xd4a00000-0xd51fffff 0xd5a00000-0xd61fffff 0xd6a00000-0xd71fffff 0xd7a00000-0xd81fffff 0xd8a00000-0xd91fffff 0xd9a00000-0xda1fffff 0xdaa00000-0xdb1fffff 0xdba00000-0xdc1fffff 0xdca00000-0xdd1fffff 0xdda00000-0xde1fffff 0xdea00000-0xdf1fffff 0xdfa00000-0xe01fffff
pcmcia: registering new device pcmcia0.0
pcmcia: Detected deprecated PCMCIA ioctl usage from process: cardmgr.
pcmcia: This interface will soon be removed from the kernel; please expect breakage unless you upgrade to new tools.
pcmcia: see http://www.kernel.org/pub/linux/utils/kernel/pcmcia/pcmcia.html for details.
cs: IO port probe 0x100-0x4ff: excluding 0x170-0x177 0x370-0x377 0x4d0-0x4d7
cs: IO port probe 0x800-0x8ff: clean.
cs: IO port probe 0xc00-0xcff: excluding 0xcf8-0xcff
cs: IO port probe 0xa00-0xaff: clean.
conflict: PCI Bus #02[3000->7fff] conflicts with ide2[310e->310e]
conflict: pcmcia_socket0[3100->310f] conflicts with ide2[310e->310e]
hwif_request_region: single-byte request for ide2
 <c0242abe> hwif_request_region+0x9e/0xa0  <c0242bbf> ide_hwif_request_regions+0xff/0x120
 <c0249c90> ide_undecoded_slave+0x0/0xc0  <c0249d79> probe_hwif+0x29/0x4f0
 <c0249c90> ide_undecoded_slave+0x0/0xc0  <c024a255> probe_hwif_init_with_fixup+0x15/0xa0
 <c0243556> ide_register_hw_with_fixup+0x166/0x1b0  <c0249c90> ide_undecoded_slave+0x0/0xc0
 <f8a8612b> idecs_register+0x7b/0xb0 [ide_cs]  <c0249c90> ide_undecoded_slave+0x0/0xc0
 <f8a865a2> ide_config+0x442/0x6c0 [ide_cs]  <f8ab348e> pcmcia_device_probe+0x6e/0x100 [pcmcia]
 <c023ca1c> driver_probe_device+0xbc/0xe0  <c023cac0> __driver_attach+0x0/0x80
 <c023cb30> __driver_attach+0x70/0x80  <c023bd69> bus_for_each_dev+0x69/0x80
 <c01e5723> kobject_set_name+0x43/0xe0  <c023cb65> driver_attach+0x25/0x30
 <c023cac0> __driver_attach+0x0/0x80  <c023c338> bus_add_driver+0x88/0xd0
 <c023d077> driver_register+0x97/0xa0  <c023cfc0> klist_devices_get+0x0/0x10
 <c023cfd0> klist_devices_put+0x0/0x10  <f8ab32f7> pcmcia_register_driver+0x17/0x50 [pcmcia]
 <f894200f> init_ide_cs+0xf/0x11 [ide_cs]  <c0136da2> sys_init_module+0xf2/0x180
 <c0102f2f> sysenter_past_esp+0x54/0x75 
conflict: PCI Bus #02[3000->7fff] conflicts with ide2[3100->3107]
conflict: pcmcia_socket0[3100->310f] conflicts with ide2[3100->3107]
Probing IDE interface ide2...
hde: LEXAR ATA FLASH, CFA DISK drive
ide2 at 0x3100-0x3107,0x310e on irq 3
hde: max request size: 128KiB
hde: 64256 sectors (32 MB) w/1KiB Cache, CHS=1004/2/32
hde: cache flushes not supported
 hde: hde1
ide-cs: hde: Vpp = 0.0

I hope somebody who understands resource.c better than me is reading
this, because I'm pretty much out of time to spend reading code.  (Happy
to test patches, of course.)

-andy

Re: 2.6.17-rc[56]-mm*: pcmcia "I/O resource not free"

[Vivek Goyal]

On Wed, Jun 21, 2006 at 10:10:48AM -0700, Andy Isaacson wrote:
> On Wed, Jun 21, 2006 at 12:46:30AM -0700, Andrew Morton wrote:
> > > [ 2034.060000] pcmcia: registering new device pcmcia0.0
> > > [ 2034.060000] PM: Adding info for pcmcia:0.0
> > > [ 2035.976000] conflict: PCI IO[0->ffff]
> > > [ 2035.976000] hwif_request_region: single-byte request for ide2
> > > [ 2035.976000]  [<c0257386>] hwif_request_region+0xa6/0xb0
> [snip]
> > > [ 2035.976000] ide2: I/O resource 0xF8B0200E-0xF8B0200E not free.
> > > [ 2035.976000] ide2: ports already in use, skipping probe
> > 
> > hm.  It appears to have decided that 0 < 0xF8B0200E < 0xffff, which is
> > clever of it.
> > 
> > Does it help if you set CONFIG_RESOURCES_32BIT?
> 
> Nope, same conflict with CONFIG_RESOURCES_32BIT set.  You're right, it
> is deciding that 0xF8B0200E conflicts with that range:
> 
> conflict: PCI IO[0->ffff] conflicts with ide2[f8b3c00e->f8b3c00e]
> 
> Looking at the code, I don't understand how this could have worked in
> -rc6; __request_resource hasn't changed, and it says
> 
>     167         if (end < start)
>     168                 return root;
>     169         if (start < root->start)
>     170                 return root;
>     171         if (end > root->end)
>     172                 return root;
> 
> If root-> start == 0 and root->end == 0xffff, we should always hit line
> 172, unless sign extension is in effect... and all the variables are
> unsigned long in -rc6, so that doesn't make sense.
> 

I think this makes sense. We are hitting line 172 in case of -mm because
f8b3c00e is not a valid io port at all. Maximum valid value can be 0xffff.
So _request_region considers this to be a conflict and returns.
 
It succeeds in -rc6 because ide code is requesting a valid ioport region
ide2[310e->310e].

So __request_region() code seems to be fine. Problem seems to be that
why do we get an invalid ioport range in following call.

addr = hwif->io_ports[IDE_CONTROL_OFFSET];

Either hwif pointer is bad or somehow the location it is pointing to
is corrupt or something else. Can you do some more tracing on hwif.

Thanks
Vivek