On 06/23/2011 05:33 PM, Jeff Garzik wrote:

On 06/23/2011 01:22 PM, Prarit Bhargava wrote:

quoted

drivers/ata changes for SMBIOS and System Firmware

As part of the new SMBIOS and System Firmware code:

- Replace old dmi* structures and functions with new sysfw* and smbios*
structures and functions in individual drivers
- cleanup sysfw_id lookup tables
- cleanup of includes for dmi.h and mod_devicetable.h which were
included in
some files that did not need them

Signed-off-by: Prarit Bhargava<redacted>
---
  drivers/ata/acard-ahci.c  |    1 -
  drivers/ata/ahci.c        |  177
+++++++++++++++++++++++----------------------
  drivers/ata/ata_piix.c    |  156
+++++++++++++++++++++++-----------------
  drivers/ata/pata_ali.c    |   18 +++---
  drivers/ata/pata_cs5530.c |   12 ++--
  drivers/ata/pata_rdc.c    |    1 -
  drivers/ata/pata_sch.c    |    1 -
  drivers/ata/pata_via.c    |   14 ++--
  drivers/ata/sata_sil.c    |   19 +++---
  9 files changed, 210 insertions(+), 189 deletions(-)

quoted

  #ifdef CONFIG_PM
+bool sysfw_match(int field, const char *str)
+{
+    const char *info = sysfw_lookup(field);
+
+    if (info == NULL || str == NULL)
+        return info == str;
+
+    return !strcmp(info, str);
+}

It is surprising that nobody else needed this.  I would rather that
this sort of thing go into the generic code, where it lived prior to
this change.

Jeff,

I took a closer look at this and, in fact, this isn't needed globally
AFAICT.

The issue is that the code really should do sysfw_callback with a
sysfw_id (which currently is a dmi_system_id).  The dmi_system_id struct
only has 4 matches/struct.  I *could* expand that to 8 but it seems like
a waste of space to do so.

But I see your point -- sysfw_match() implies a global function.  I'll
rework this to include a manual strcmp() instead of sysfw_match().  It
seems more appropriate IMO.

I'll post a [v2] shortly.

P.