2.6.17-rc1: collie -- oopsen in pccardd?

2.6.17-rc1: collie -- oopsen in pccardd?

[Pavel Machek]

Hi!

I'm getting some oopses when inserting/removing pccard (on collie,
oopses in pccardd). It does not break boot, so it is not immediate
problem, but I wonder if it also happens on non-collie machines?
								Pavel
-- 
Picture of sleeping (Linux) penguin wanted...

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Russell King]

On Tue, Apr 04, 2006 at 02:22:12PM +0200, Pavel Machek wrote:
> I'm getting some oopses when inserting/removing pccard (on collie,
> oopses in pccardd). It does not break boot, so it is not immediate
> problem, but I wonder if it also happens on non-collie machines?

No idea what so ever.  Not even any clues as to what might be going wrong
due to the lack of oops dump.  (Not that I even look after PCMCIA anymore.)

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Pavel Machek]

Hi!

On Tue 04-04-06 13:43:50, Russell King wrote:
> On Tue, Apr 04, 2006 at 02:22:12PM +0200, Pavel Machek wrote:
> > I'm getting some oopses when inserting/removing pccard (on collie,
> > oopses in pccardd). It does not break boot, so it is not immediate
> > problem, but I wonder if it also happens on non-collie machines?
> 
> No idea what so ever.  Not even any clues as to what might be going wrong
> due to the lack of oops dump.  (Not that I even look after PCMCIA anymore.)

Sorry for lack of oops. I was not expecting you to debug it, I
expected some voices telling me it is broken for them, too :-).

							Pavel
-- 
Thanks, Sharp!

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Richard Purdie]

On Tue, 2006-04-04 at 00:01 +0000, Pavel Machek wrote:
> Hi!
> 
> On Tue 04-04-06 13:43:50, Russell King wrote:
> > On Tue, Apr 04, 2006 at 02:22:12PM +0200, Pavel Machek wrote:
> > > I'm getting some oopses when inserting/removing pccard (on collie,
> > > oopses in pccardd). It does not break boot, so it is not immediate
> > > problem, but I wonder if it also happens on non-collie machines?
> > 
> > No idea what so ever.  Not even any clues as to what might be going wrong
> > due to the lack of oops dump.  (Not that I even look after PCMCIA anymore.)
> 
> Sorry for lack of oops. I was not expecting you to debug it, I
> expected some voices telling me it is broken for them, too :-).

With a recent git kernel (907d91d708d9999bec0185d630062576ac4181a7) I
see the oops below when booting spitz (SL-C3000 - ARM pxa270 based). Was
this the same oops you saw Pavel?

Spitz boots off a CompactFlash microdrive so this is somewhat fatal for
the device. 2.6.16 is the last kernel version I tested and that works.

Unable to handle kernel NULL pointer dereference at virtual address 0000000c
pgd = c0004000
[0000000c] *pgd=00000000
Internal error: Oops: f5 [#1]
Modules linked in:
CPU: 0
PC is at alloc_io_space+0x188/0x294
LR is at pcmcia_request_io+0x90/0x124
pc : [<c0161ea8>]    lr : [<c0162044>]    Not tainted
sp : c032bdd0  ip : 00000000  fp : c032be04
r10: 00000000  r9 : c3cec220  r8 : 00000010
r7 : c037e6e0  r6 : c032e2f8  r5 : 00000010  r4 : c3cec220
r3 : c4820000  r2 : 00000000  r1 : 00000010  r0 : 00000000
Flags: NzCv  IRQs on  FIQs on  Mode SVC_32  Segment kernel
Control: 397F  Table: A0004000  DAC: 00000017
Process swapper (pid: 1, stack limit = 0xc032a1a8)
Stack: (0xc032bdd0 to 0xc032c000)
bdc0:                                     c015d1dc 00000010 0000004b c3cec220
bde0: c3cec200 c032e2f8 c037e6e0 c035de60 d1b71759 00000000 c032be2c c032be08
be00: c0162044 c0161d2c 00000004 c3cec200 c3caa900 c032be38 c3cec200 c3caa800
be20: c032be84 c032be30 c0150974 c0161fc0 c032be50 00000000 00000000 0000001b
be40: 00000110 00000000 0000004b ff00061b c032be06 c3caa800 c0106294 c022c820
be60: c3cec200 c3cec290 c022c838 00000000 c032bfa0 c0019f58 c032bea4 c032be88
be80: c0161354 c015068c c3cec354 c3cec290 c0161298 c022c838 c032bec4 c032bea8
bea0: c013c048 c01612a4 c3cec354 c3cec290 c022c838 c013c164 c032bedc c032bec8
bec0: c013c248 c013c000 00000000 c032bee4 c032bf0c c032bee0 c013ba0c c013c170
bee0: 00000000 c022e290 c022e290 c3cec2d8 c022c84c c022c838 c022e1a8 c022d07c
bf00: c032bf1c c032bf10 c013bf80 c013b9c4 c032bf44 c032bf20 c013b618 c013bf6c
bf20: c022c838 c022d058 00000004 c022d07c c022c820 c0019f58 c032bf5c c032bf48
bf40: c013c4e4 c013b5b4 c01ff480 00000000 c032bf8c c032bf60 c0161100 c013c488
bf60: c032bf8c c032bf70 c011f960 c0019f4c 00000000 c032a000 00000001 c001a404
bf80: c032bf9c c032bf90 c00166e8 c0161038 c032bff4 c032bfa0 c001c130 c00166e0
bfa0: 00000000 c032bfb0 c001cea4 c0036a9c 00000000 00000000 c001c02c c003db78
bfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
bfe0: 00000000 00000000 00000000 c032bff8 c003db78 c001c038 f818f800 18181818
Backtrace:
[<c0161d20>] (alloc_io_space+0x0/0x294) from [<c0162044>] (pcmcia_request_io+0x90/0x124)
[<c0161fb4>] (pcmcia_request_io+0x0/0x124) from [<c0150974>] (ide_probe+0x2f4/0x570)
 r7 = C3CAA800  r6 = C3CEC200  r5 = C032BE38  r4 = C3CAA900
[<c0150680>] (ide_probe+0x0/0x570) from [<c0161354>] (pcmcia_device_probe+0xbc/0x1b0)
[<c0161298>] (pcmcia_device_probe+0x0/0x1b0) from [<c013c048>] (driver_probe_device+0x54/0xd0)
 r7 = C022C838  r6 = C0161298  r5 = C3CEC290  r4 = C3CEC354
[<c013bff4>] (driver_probe_device+0x0/0xd0) from [<c013c248>] (__driver_attach+0xe4/0xe8)
 r7 = C013C164  r6 = C022C838  r5 = C3CEC290  r4 = C3CEC354
[<c013c164>] (__driver_attach+0x0/0xe8) from [<c013ba0c>] (bus_for_each_dev+0x54/0x80)
 r5 = C032BEE4  r4 = 00000000
[<c013b9b8>] (bus_for_each_dev+0x0/0x80) from [<c013bf80>] (driver_attach+0x20/0x28)
 r7 = C022D07C  r6 = C022E1A8  r5 = C022C838  r4 = C022C84C
[<c013bf60>] (driver_attach+0x0/0x28) from [<c013b618>] (bus_add_driver+0x70/0x140)
[<c013b5a8>] (bus_add_driver+0x0/0x140) from [<c013c4e4>] (driver_register+0x68/0xac)
[<c013c47c>] (driver_register+0x0/0xac) from [<c0161100>] (pcmcia_register_driver+0xd4/0xfc)
 r4 = 00000000
[<c016102c>] (pcmcia_register_driver+0x0/0xfc) from [<c00166e8>] (init_ide_cs+0x14/0x1c)
 r8 = C001A404  r7 = 00000001  r6 = C032A000  r5 = 00000000
 r4 = C0019F4C
[<c00166d4>] (init_ide_cs+0x0/0x1c) from [<c001c130>] (init+0x104/0x2ac)
[<c001c02c>] (init+0x0/0x2ac) from [<c003db78>] (do_exit+0x0/0xa78)
Code: e5893000 e596203c e20010ff e3a00000 (e592300c)
 <0>Kernel panic - not syncing: Attempted to kill init!

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Pavel Machek]

Hi!

> > > > I'm getting some oopses when inserting/removing pccard (on collie,
> > > > oopses in pccardd). It does not break boot, so it is not immediate
> > > > problem, but I wonder if it also happens on non-collie machines?
> > > 
> > > No idea what so ever.  Not even any clues as to what might be going wrong
> > > due to the lack of oops dump.  (Not that I even look after PCMCIA anymore.)
> > 
> > Sorry for lack of oops. I was not expecting you to debug it, I
> > expected some voices telling me it is broken for them, too :-).
> 
> With a recent git kernel (907d91d708d9999bec0185d630062576ac4181a7) I
> see the oops below when booting spitz (SL-C3000 - ARM pxa270 based). Was
> this the same oops you saw Pavel?

I think so.
								Pavel

-- 
Thanks, Sharp!

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Russell King]

On Thu, Apr 13, 2006 at 06:47:06PM +0200, Pavel Machek wrote:
> Hi!
> 
> > > > > I'm getting some oopses when inserting/removing pccard (on collie,
> > > > > oopses in pccardd). It does not break boot, so it is not immediate
> > > > > problem, but I wonder if it also happens on non-collie machines?
> > > > 
> > > > No idea what so ever.  Not even any clues as to what might be going wrong
> > > > due to the lack of oops dump.  (Not that I even look after PCMCIA anymore.)
> > > 
> > > Sorry for lack of oops. I was not expecting you to debug it, I
> > > expected some voices telling me it is broken for them, too :-).
> > 
> > With a recent git kernel (907d91d708d9999bec0185d630062576ac4181a7) I
> > see the oops below when booting spitz (SL-C3000 - ARM pxa270 based). Was
> > this the same oops you saw Pavel?
> 
> I think so.

diff --git a/drivers/pcmcia/pcmcia_resource.c b/drivers/pcmcia/pcmcia_resource.c
--- a/drivers/pcmcia/pcmcia_resource.c
+++ b/drivers/pcmcia/pcmcia_resource.c
@@ -89,7 +88,7 @@ static int alloc_io_space(struct pcmcia_
        }
        if ((s->features & SS_CAP_STATIC_MAP) && s->io_offset) {
                *base = s->io_offset | (*base & 0x0fff);
-               s->io[0].Attributes = attr;
+               s->io[0].res->flags = (s->io[0].res->flags & ~IORESOURCE_BITS) | (attr & IORESOURCE_BITS);
                return 0;
        }
        /* Check for an already-allocated window that must conflict with

will probably be the culpret - which is from commit
c7d006935dfda9174187aa557e94a137ced10c30.

Static maps do not have IO resources, so s->io[].Attributes was not a
"duplicated" field in this case.  This part of this change needs
reverting.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Dominik Brodowski]

On Thu, Apr 13, 2006 at 05:54:52PM +0100, Russell King wrote:
> diff --git a/drivers/pcmcia/pcmcia_resource.c b/drivers/pcmcia/pcmcia_resource.c
> --- a/drivers/pcmcia/pcmcia_resource.c
> +++ b/drivers/pcmcia/pcmcia_resource.c
> @@ -89,7 +88,7 @@ static int alloc_io_space(struct pcmcia_
>         }
>         if ((s->features & SS_CAP_STATIC_MAP) && s->io_offset) {
>                 *base = s->io_offset | (*base & 0x0fff);
> -               s->io[0].Attributes = attr;
> +               s->io[0].res->flags = (s->io[0].res->flags & ~IORESOURCE_BITS) | (attr & IORESOURCE_BITS);
>                 return 0;
>         }
>         /* Check for an already-allocated window that must conflict with
> 
> will probably be the culpret - which is from commit
> c7d006935dfda9174187aa557e94a137ced10c30.
> 
> Static maps do not have IO resources, so s->io[].Attributes was not a
> "duplicated" field in this case.  This part of this change needs
> reverting.

Oh yes, mea culpa. However, we can simply remove setting res->flags here, as
we never read it in this case anyways.

Thanks,
	Dominik


From: Dominik Brodowski <linux@dominikbrodowski.net>
Date: Thu Apr 13 19:06:49 2006 +0200
Subject: [PATCH] pcmcia: fix oops in static mapping case

As static maps do not have IO resources, this setting oopses. However, as
we do not ever read this value, we can safely remove it.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

---

 drivers/pcmcia/pcmcia_resource.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

618d4702f894d73df4f6e631e83699ea5cc87abc
diff --git a/drivers/pcmcia/pcmcia_resource.c b/drivers/pcmcia/pcmcia_resource.c
index 2539c0b..cc3402c 100644
--- a/drivers/pcmcia/pcmcia_resource.c
+++ b/drivers/pcmcia/pcmcia_resource.c
@@ -88,7 +88,6 @@ static int alloc_io_space(struct pcmcia_
 	}
 	if ((s->features & SS_CAP_STATIC_MAP) && s->io_offset) {
 		*base = s->io_offset | (*base & 0x0fff);
-		s->io[0].res->flags = (s->io[0].res->flags & ~IORESOURCE_BITS) | (attr & IORESOURCE_BITS);
 		return 0;
 	}
 	/* Check for an already-allocated window that must conflict with
-- 
1.2.4

 

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Pavel Machek]

> On Thu, Apr 13, 2006 at 05:54:52PM +0100, Russell King wrote:
> > diff --git a/drivers/pcmcia/pcmcia_resource.c b/drivers/pcmcia/pcmcia_resource.c
> > --- a/drivers/pcmcia/pcmcia_resource.c
> > +++ b/drivers/pcmcia/pcmcia_resource.c
> > @@ -89,7 +88,7 @@ static int alloc_io_space(struct pcmcia_
> >         }
> >         if ((s->features & SS_CAP_STATIC_MAP) && s->io_offset) {
> >                 *base = s->io_offset | (*base & 0x0fff);
> > -               s->io[0].Attributes = attr;
> > +               s->io[0].res->flags = (s->io[0].res->flags & ~IORESOURCE_BITS) | (attr & IORESOURCE_BITS);
> >                 return 0;
> >         }
> >         /* Check for an already-allocated window that must conflict with
> > 
> > will probably be the culpret - which is from commit
> > c7d006935dfda9174187aa557e94a137ced10c30.
> > 
> > Static maps do not have IO resources, so s->io[].Attributes was not a
> > "duplicated" field in this case.  This part of this change needs
> > reverting.
> 
> Oh yes, mea culpa. However, we can simply remove setting res->flags here, as
> we never read it in this case anyways.

It fixes the problem for me.
									Pavel

-- 
Thanks, Sharp!

Re: 2.6.17-rc1: collie -- oopsen in pccardd?

[Richard Purdie]

On Thu, 2006-04-13 at 19:14 +0200, Dominik Brodowski wrote:
> Oh yes, mea culpa. However, we can simply remove setting res->flags here, as
> we never read it in this case anyways.
> 
> Thanks,
> 	Dominik
> 
> 
> From: Dominik Brodowski <linux@dominikbrodowski.net>
> Date: Thu Apr 13 19:06:49 2006 +0200
> Subject: [PATCH] pcmcia: fix oops in static mapping case
> 
> As static maps do not have IO resources, this setting oopses. However, as
> we do not ever read this value, we can safely remove it.
> 
> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
>
> diff --git a/drivers/pcmcia/pcmcia_resource.c
> b/drivers/pcmcia/pcmcia_resource.c
> index 2539c0b..cc3402c 100644
> --- a/drivers/pcmcia/pcmcia_resource.c
> +++ b/drivers/pcmcia/pcmcia_resource.c
> @@ -88,7 +88,6 @@ static int alloc_io_space(struct pcmcia_
>         }
>         if ((s->features & SS_CAP_STATIC_MAP) && s->io_offset) {
>                 *base = s->io_offset | (*base & 0x0fff);
> -               s->io[0].res->flags = (s->io[0].res->flags &
> ~IORESOURCE_BITS) | (attr & IORESOURCE_BITS);
>                 return 0;
>         }
>         /* Check for an already-allocated window that must conflict 


I can confirm this fixes the problem I was seeing as well.

Thanks,

Richard