On Wednesday 19 April 2006 11:48, Therien, Guy wrote: > My view has always been that _OSI strings should be run time loadable. > However, it is clear that drivers that may load the strings may not > always be loaded before their positive evaluation is required. I had > suggested a registration interface that means that the support exists on > the system i.e. it will be available later but certainly that has some > caveats also. I mentioned run-time _OSI strings as an example of the flexibility of an OSL interface, not because I think it's a good idea. In fact, it strikes me as a *bad* idea, because firmware writers might assume that _OSI strings are constant for a given boot. The spec isn't explicit either way, though it does recommend that _OSI be evaluated by \_SB.INI, i.e., early in OS initialization. Changing things later on sounds like a good way to introduce misunderstandings. What value do you see in run-time _OSI strings? > -----Original Message----- > From: Moore, Robert > Sent: Wednesday, April 19, 2006 8:37 AM > To: Bjorn Helgaas > Cc: Brown, Len; linux-acpi@xxxxxxxxxxxxxxx; Andrew Morton; Therien, Guy > Subject: RE: [1/2] ACPI: make _OSI strings static > > I'm thinking about defining a new OSL interface to let the host match > the optional _OSI strings like "Module Device", "Processor Device", "3.0 > Thermal Model", etc. > > This seems to me to be the easiest way to configure these strings > (rather than a bunch of #ifdefs), and is also much easier to expand as > the host implements new features and new strings are added in future > ACPI specifications. > > Bob > > > > -----Original Message----- > > From: linux-acpi-owner@xxxxxxxxxxxxxxx [mailto:linux-acpi- > > owner@xxxxxxxxxxxxxxx] On Behalf Of Bjorn Helgaas > > Sent: Tuesday, April 18, 2006 3:38 PM > > To: Moore, Robert > > Cc: Brown, Len; linux-acpi@xxxxxxxxxxxxxxx; Andrew Morton > > Subject: Re: [1/2] ACPI: make _OSI strings static > > > > On Tuesday 18 April 2006 16:29, Moore, Robert wrote: > > > While I'm certainly all for letting the compiler do such things, if > > > we are going to do something like this, I would like to convert all > > > similar instances all at once, not just onesey-twosey -- just to > > > keep everything consistent. > > > > Sure, that sounds fine. I don't want to get involved in that effort > > though, because any patches I would come up with would be against the > > Linux-ized source and hence less useful to you. > > > > > As far as the "Module Device" string, would this support not be > > > appropriate for any OS that implements this feature? If so, it > > > should be an OS-specific header configuration option. > > > > Yup, that sounds reasonable. I confess I don't really know where > > options like that live. include/acpi/platform/aclinux.h? I can dream > > > up a patch for that if you want, but it would probably depend on > > something like my first patch and it'd probably be more work for you > > to integrate than for you to just do it yourself. > > > > > Actually, I find this even more readable: > > > #define ACPI_NUM_OSI_STRINGS ACPI_ARRAY_SIZE > (acpi_valid_osi_strings) > > > > > > and leave the existing code as: > > > for (i = 0; i < ACPI_NUM_OSI_STRINGS; i++) { > > > > OK. > > > > > > -----Original Message----- > > > > From: Bjorn Helgaas [mailto:bjorn.helgaas@xxxxxx] > > > > Sent: Tuesday, April 18, 2006 3:18 PM > > > > To: Moore, Robert > > > > Cc: Brown, Len; linux-acpi@xxxxxxxxxxxxxxx; Andrew Morton > > > > Subject: Re: [1/2] ACPI: make _OSI strings static > > > > > > > > On Tuesday 18 April 2006 16:02, Moore, Robert wrote: > > > > > This seems like an incredible long way around just to get rid of > > > > > > one configuration constant. Is it really that much hassle to > > > > > update NUM_OSI_STRINGS? There is a big comment about this in > > > > > front of the > > > osi > > > > > string declarations. > > > > > > > > No, it's not *that* much hassle. But what's the point of > > > > mentioning acpi_gbl_valid_osi_strings[] in three files when it's > > > > only used in one? And why define NUM_OSI_STRINGS when you can > > > > easily have the compiler compute it and remove the possibility for > error? > > > > > > > > It might make more sense if you look at the second part of the > > > > patch, which adds "Module Device" to _OSI. That's trivial, but > > > > without this first path, we'd have to do something like: > > > > > > > > #ifdef CONFIG_ACPI_CONTAINER > > > > #define ACPI_NUM_OSI_STRINGS 11 > > > > #else > > > > #define ACPI_NUM_OSI_STRINGS 10 > > > > #endif > > > > > > > > which is ugly but maybe OK with only one CONFIG option involved. > > > > But when we add another one or two options, and we want to handle > > > > all combinations of them being enabled, it would be unmanageable. > > > > > > > > > > -----Original Message----- > > > > > > From: Bjorn Helgaas [mailto:bjorn.helgaas@xxxxxx] > > > > > > Sent: Tuesday, April 18, 2006 2:57 PM > > > > > > To: Brown, Len > > > > > > Cc: linux-acpi@xxxxxxxxxxxxxxx; Andrew Morton; Moore, Robert > > > > > > Subject: [1/2] ACPI: make _OSI strings static > > > > > > > > > > > > Move acpi_gbl_valid_osi_strings[] from utglobal.c to uteval.c > > > > > > and make it static. Then we can remove ACPI_NUM_OSI_STRINGS, > > > > > > which is a maintenance problem for OSPMs that want to change > > > > > > the features reported by _OSI. > > > > > > > > > > > > This patch is against the Linux-ized ACPI CA, since I don't > > > > > > have a good way to modify or test the original Intel version. > > > > > > > I hope it's small enough that you can easily make the > > > > > > corresponding > > > change > > > > > > to the Intel version. > > > > > > > > > > > > This modification may be used under either the GPL or the > > > BSD-style > > > > > > license used for the Intel ACPI CA. > > > > > > > > > > > > Signed-off-by: Bjorn Helgaas <bjorn.helgaas@xxxxxx> > > > > > > > > > > > > Index: work-mm5/drivers/acpi/utilities/uteval.c > > > > > > > > > =================================================================== > > > > > > --- work-mm5.orig/drivers/acpi/utilities/uteval.c > 2006-04-03 > > > > > > 15:04:31.000000000 -0600 > > > > > > +++ work-mm5/drivers/acpi/utilities/uteval.c 2006-04-18 > > > > > > 15:31:22.000000000 -0600 > > > > > > @@ -48,6 +48,29 @@ > > > > > > #define _COMPONENT ACPI_UTILITIES > > > > > > ACPI_MODULE_NAME("uteval") > > > > > > > > > > > > +#define ACPI_ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) > > > > > > + > > > > > > +/* > > > > > > + * Strings supported by the _OSI predefined (internal) > method. > > > > > > + */ > > > > > > +static const char *acpi_valid_osi_strings[] = { > > > > > > + /* Operating System Vendor Strings */ > > > > > > + > > > > > > + "Linux", > > > > > > + "Windows 2000", > > > > > > + "Windows 2001", > > > > > > + "Windows 2001.1", > > > > > > + "Windows 2001 SP0", > > > > > > + "Windows 2001 SP1", > > > > > > + "Windows 2001 SP2", > > > > > > + "Windows 2001 SP3", > > > > > > + "Windows 2001 SP4", > > > > > > + > > > > > > + /* Feature Group Strings */ > > > > > > + > > > > > > + "Extended Address Space Descriptor", }; > > > > > > + > > > > > > /* Local prototypes */ > > > > > > static void > > > > > > acpi_ut_copy_id_string(char *destination, char *source, > > > > > > acpi_size max_length); @@ -93,10 +116,9 @@ > > > > > > > > > > > > /* Compare input string to table of supported strings */ > > > > > > > > > > > > - for (i = 0; i < ACPI_NUM_OSI_STRINGS; i++) { > > > > > > + for (i = 0; i < ACPI_ARRAY_SIZE(acpi_valid_osi_strings); > > > > > > > +i++) > > { > > > > > > if (!ACPI_STRCMP(string_desc->string.pointer, > > > > > > - ACPI_CAST_PTR(char, > > > > > > - > > > > > acpi_gbl_valid_osi_strings[i]))) > > > > > > + > > > > > acpi_valid_osi_strings[i])) > > > > > > { > > > > > > > > > > > > /* This string is supported */ > > > > > > Index: work-mm5/drivers/acpi/utilities/utglobal.c > > > > > > > > > =================================================================== > > > > > > --- work-mm5.orig/drivers/acpi/utilities/utglobal.c > > > 2006-04-18 > > > > > > 12:42:24.000000000 -0600 > > > > > > +++ work-mm5/drivers/acpi/utilities/utglobal.c > 2006-04-18 > > > > > > 12:55:24.000000000 -0600 > > > > > > @@ -184,28 +184,6 @@ > > > > > > "_S4D" > > > > > > }; > > > > > > > > > > > > -/* > > > > > > - * Strings supported by the _OSI predefined (internal) > method. > > > > > > - * When adding strings, be sure to update > ACPI_NUM_OSI_STRINGS. > > > > > > - */ > > > > > > -const char *acpi_gbl_valid_osi_strings[ACPI_NUM_OSI_STRINGS] > = { > > > > > > - /* Operating System Vendor Strings */ > > > > > > - > > > > > > - "Linux", > > > > > > - "Windows 2000", > > > > > > - "Windows 2001", > > > > > > - "Windows 2001.1", > > > > > > - "Windows 2001 SP0", > > > > > > - "Windows 2001 SP1", > > > > > > - "Windows 2001 SP2", > > > > > > - "Windows 2001 SP3", > > > > > > - "Windows 2001 SP4", > > > > > > - > > > > > > - /* Feature Group Strings */ > > > > > > - > > > > > > - "Extended Address Space Descriptor" > > > > > > -}; > > > > > > - > > > > > > > > > > > > > > > > > > > > /******************************************************************* > > > **** > > > > > ** > > > > > > ****** > > > > > > * > > > > > > * Namespace globals > > > > > > Index: work-mm5/include/acpi/acconfig.h > > > > > > > > > =================================================================== > > > > > > --- work-mm5.orig/include/acpi/acconfig.h 2006-04-18 > > > > > 12:42:24.000000000 - > > > > > > 0600 > > > > > > +++ work-mm5/include/acpi/acconfig.h 2006-04-18 > > > 12:42:58.000000000 - > > > > > > 0600 > > > > > > @@ -187,10 +187,6 @@ > > > > > > > > > > > > #define ACPI_SMBUS_BUFFER_SIZE 34 > > > > > > > > > > > > -/* Number of strings associated with the _OSI reserved method > > > > > > > */ > > > > > > - > > > > > > -#define ACPI_NUM_OSI_STRINGS 10 > > > > > > - > > > > > > > > > > > > > > > > > > > > /******************************************************************* > > > **** > > > > > ** > > > > > > ***** > > > > > > * > > > > > > * ACPI AML Debugger > > > > > > Index: work-mm5/include/acpi/acglobal.h > > > > > > > > > =================================================================== > > > > > > --- work-mm5.orig/include/acpi/acglobal.h 2006-03-23 > > > > > 10:22:17.000000000 - > > > > > > 0700 > > > > > > +++ work-mm5/include/acpi/acglobal.h 2006-04-18 > > > 15:24:33.000000000 - > > > > > > 0600 > > > > > > @@ -244,7 +244,6 @@ > > > > > > extern const char *acpi_gbl_highest_dstate_names[4]; > > > > > > extern const struct acpi_opcode_info > > > > > > acpi_gbl_aml_op_info[AML_NUM_OPCODES]; > > > > > > extern const char > > > > > *acpi_gbl_region_types[ACPI_NUM_PREDEFINED_REGIONS]; > > > > > > -extern const char > > > *acpi_gbl_valid_osi_strings[ACPI_NUM_OSI_STRINGS]; > > > > > > > > > > > > > > > > > > > > > > > > > > /******************************************************************* > > > **** > > > > > ** > > > > > > **** > > > > > > * > > > > > > > > > > - > > To unsubscribe from this list: send the line "unsubscribe linux-acpi" > > in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo > > info at http://vger.kernel.org/majordomo-info.html > - > To unsubscribe from this list: send the line "unsubscribe linux-acpi" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html > - To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
