region size computed incorrectly

region size computed incorrectly

[Albert Cahalan]

Consider the following chunk of code, produced by
the Intel disassembler:

                Device (HPET)
                {
                    Name (_HID, EisaId ("PNP0103"))
                    Name (_CRS, ResourceTemplate ()
                    {
                        IRQNoFlags () {0}
                        IRQNoFlags () {8}
                        Memory32Fixed (ReadOnly, 0xFEC01000, 0x00000400)
                    })
                    OperationRegion (^LPCR, PCI_Config, 0xA0, 0x04)
                    Field (^LPCR, AnyAcc, NoLock, Preserve)
                    {
                        HPTE,   32
                    }

                    Method (_STA, 0, NotSerialized)
                    {
                        And (HPTE, 0x01, Local0)
                        If (Local0)
                        {
                            Return (0x0F)
                        }

                        Return (0x00)
                    }
                }


There we have a 4-byte OperationRegion with a 32-bit
field called HPTE. That fits. Then, in _STA, HPTE is
read, and linux-2.6.0-test3 spits out errors:


ACPI: Subsystem revision 20030714
 tbxface-0117 [03] acpi_load_tables      : ACPI Tables successfully acquired
Parsing all Control Methods:.........................................................................................................................................
Table [DSDT](id F004) - 446 Objects with 42 Devices 137 Methods 13 Regions
ACPI Namespace successfully loaded at root ffffffff80590780
evxfevnt-0093 [04] acpi_enable           : Transition to ACPI mode successful
evgpeblk-0748 [06] ev_create_gpe_block   : GPE 00 to 15 [_GPE] 2 regs at 0000000000005020 on int 9
evgpeblk-0748 [06] ev_create_gpe_block   : GPE 16 to 47 [_GPE] 4 regs at 00000000000050B0 on int 9
Completing Region/Field/Buffer/Package initialization:............................................................................................
Initialized 12/14 Regions 26/26 Fields 38/38 Buffers 16/16 Packages (455 nodes)
Executing all Device _STA and_INI methods:...................... exfldio-0129 [23] ex_setup_region       : Field [HPTE] access width (1 bytes) too large for region [LPCR] (length 0)
 exfldio-0140 [23] ex_setup_region       : Field [HPTE] Base+Offset+Width 0+0+1 is beyond end of region [LPCR] (length 0)
 dswexec-0422 [15] ds_exec_end_op        : [And]: Could not resolve operands, AE_AML_REGION_LIMIT
 psparse-1121: *** Error: Method execution failed [\_SB_.PCI0.SBRG.HPET._STA] (Node 00000100081d36f0), AE_AML_REGION_LIMIT
  uteval-0098: *** Error: Method execution failed [\_SB_.PCI0.SBRG.HPET._STA] (Node 00000100081d36f0), AE_AML_REGION_LIMIT
.....................
43 Devices found containing: 42 _STA, 0 _INI methods
ACPI: Interpreter enabled
ACPI: Using IOAPIC for interrupt routing
ACPI: PCI Root Bridge [PCI0] (00:00)
PCI: Probing PCI hardware (bus 00)
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCI1._PRT]
 exfldio-0129 [21] ex_setup_region       : Field [HPTE] access width (1 bytes) too large for region [LPCR] (length 0)
 exfldio-0140 [21] ex_setup_region       : Field [HPTE] Base+Offset+Width 0+0+1 is beyond end of region [LPCR] (length 0)
 dswexec-0422 [13] ds_exec_end_op        : [And]: Could not resolve operands, AE_AML_REGION_LIMIT
 psparse-1121: *** Error: Method execution failed [\_SB_.PCI0.SBRG.HPET._STA] (Node 00000100081d36f0), AE_AML_REGION_LIMIT


Somehow, the LPCR region is being seen as length 0.
This is not correct; it is 4 bytes wide. Until the
boundry checking is fixed, it should be disabled.
Even then, a minor overflow should just printk() a
warning and keep on going.

BTW, what is a PNP0103 device?




-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01

RE: region size computed incorrectly

[Grover, Andrew]

I was able to reproduce this. Putting LPCR in the HPET scope instead of
the parent scope (by removing the ^s) seems to be a temporary
workaround. We appear to be creating 2 LPCR objects, 1 of which is
initialized correctly and 1 which is zeroed out. We look at the
not-initialized one, and throw errors.

BTW PNP0103 is the HPET, High Precision Event Timer. I don't know how
people are supposed to request PNP IDs these days, but the HPET spec
says 0103, so I guess they got it. :)

Regards -- Andy

> -----Original Message-----
> From: Albert Cahalan [mailto:albert-Rn4VEauK+AKRv+LV9MX5uipxlwaOVQ5f@public.gmane.org] 
> Sent: Saturday, August 16, 2003 8:57 AM
> To: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> Subject: [ACPI] region size computed incorrectly
> 
> 
> Consider the following chunk of code, produced by
> the Intel disassembler:
> 
>                 Device (HPET)
>                 {
>                     Name (_HID, EisaId ("PNP0103"))
>                     Name (_CRS, ResourceTemplate ()
>                     {
>                         IRQNoFlags () {0}
>                         IRQNoFlags () {8}
>                         Memory32Fixed (ReadOnly, 0xFEC01000, 
> 0x00000400)
>                     })
>                     OperationRegion (^LPCR, PCI_Config, 0xA0, 0x04)
>                     Field (^LPCR, AnyAcc, NoLock, Preserve)
>                     {
>                         HPTE,   32
>                     }
> 
>                     Method (_STA, 0, NotSerialized)
>                     {
>                         And (HPTE, 0x01, Local0)
>                         If (Local0)
>                         {
>                             Return (0x0F)
>                         }
> 
>                         Return (0x00)
>                     }
>                 }
> 
> 
> There we have a 4-byte OperationRegion with a 32-bit
> field called HPTE. That fits. Then, in _STA, HPTE is
> read, and linux-2.6.0-test3 spits out errors:
> 
> 
> ACPI: Subsystem revision 20030714
>  tbxface-0117 [03] acpi_load_tables      : ACPI Tables 
> successfully acquired
> Parsing all Control 
> Methods:......................................................
> ..............................................................
> .....................
> Table [DSDT](id F004) - 446 Objects with 42 Devices 137 
> Methods 13 Regions
> ACPI Namespace successfully loaded at root ffffffff80590780
> evxfevnt-0093 [04] acpi_enable           : Transition to ACPI 
> mode successful
> evgpeblk-0748 [06] ev_create_gpe_block   : GPE 00 to 15 
> [_GPE] 2 regs at 0000000000005020 on int 9
> evgpeblk-0748 [06] ev_create_gpe_block   : GPE 16 to 47 
> [_GPE] 4 regs at 00000000000050B0 on int 9
> Completing Region/Field/Buffer/Package 
> initialization:...............................................
> .............................................
> Initialized 12/14 Regions 26/26 Fields 38/38 Buffers 16/16 
> Packages (455 nodes)
> Executing all Device _STA and_INI 
> methods:...................... exfldio-0129 [23] 
> ex_setup_region       : Field [HPTE] access width (1 bytes) 
> too large for region [LPCR] (length 0)
>  exfldio-0140 [23] ex_setup_region       : Field [HPTE] 
> Base+Offset+Width 0+0+1 is beyond end of region [LPCR] (length 0)
>  dswexec-0422 [15] ds_exec_end_op        : [And]: Could not 
> resolve operands, AE_AML_REGION_LIMIT
>  psparse-1121: *** Error: Method execution failed 
> [\_SB_.PCI0.SBRG.HPET._STA] (Node 00000100081d36f0), 
> AE_AML_REGION_LIMIT
>   uteval-0098: *** Error: Method execution failed 
> [\_SB_.PCI0.SBRG.HPET._STA] (Node 00000100081d36f0), 
> AE_AML_REGION_LIMIT
> .....................
> 43 Devices found containing: 42 _STA, 0 _INI methods
> ACPI: Interpreter enabled
> ACPI: Using IOAPIC for interrupt routing
> ACPI: PCI Root Bridge [PCI0] (00:00)
> PCI: Probing PCI hardware (bus 00)
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCI1._PRT]
>  exfldio-0129 [21] ex_setup_region       : Field [HPTE] 
> access width (1 bytes) too large for region [LPCR] (length 0)
>  exfldio-0140 [21] ex_setup_region       : Field [HPTE] 
> Base+Offset+Width 0+0+1 is beyond end of region [LPCR] (length 0)
>  dswexec-0422 [13] ds_exec_end_op        : [And]: Could not 
> resolve operands, AE_AML_REGION_LIMIT
>  psparse-1121: *** Error: Method execution failed 
> [\_SB_.PCI0.SBRG.HPET._STA] (Node 00000100081d36f0), 
> AE_AML_REGION_LIMIT
> 
> 
> Somehow, the LPCR region is being seen as length 0.
> This is not correct; it is 4 bytes wide. Until the
> boundry checking is fixed, it should be disabled.
> Even then, a minor overflow should just printk() a
> warning and keep on going.
> 
> BTW, what is a PNP0103 device?
> 
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by: Free pre-built ASP.NET sites including
> Data Reports, E-commerce, Portals, and Forums are available now.
> Download today and enter to win an XBOX or Visual Studio .NET.
> http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet
> _072303_01/01
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 


-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01

RE: region size computed incorrectly

[Albert Cahalan]

On Mon, 2003-08-18 at 15:49, Grover, Andrew wrote:
> I was able to reproduce this. Putting LPCR in the HPET scope instead of
> the parent scope (by removing the ^s) seems to be a temporary
> workaround. We appear to be creating 2 LPCR objects, 1 of which is
> initialized correctly and 1 which is zeroed out. We look at the
> not-initialized one, and throw errors.
> 
> BTW PNP0103 is the HPET, High Precision Event Timer. I don't know how
> people are supposed to request PNP IDs these days, but the HPET spec
> says 0103, so I guess they got it. :)

Any hope for a fix? The 2.6.0-test5 kernel will be
using the HPET for timekeeping.




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf