Implemented pipelined hash calculation in LoadFirmware()

Implemented pipelined hash calculation in LoadFirmware()

[Bill Richardson, 2010-06-15 21:00:41]

one question

http://codereview.chromium.org/2802002/diff/2001/3004
File vboot_firmware/lib/vboot_firmware.c (right):

http://codereview.chromium.org/2802002/diff/2001/3004#newcode17
vboot_firmware/lib/vboot_firmware.c:17: static DigestContext ctx;
Do we really want static/global variables inside a library? It doesn't seem like
a great idea. I thought that's what the various *Params structs were for.

[Randall Spangler, 2010-06-15 21:45:21]

http://codereview.chromium.org/2802002/diff/2001/3004
File vboot_firmware/lib/vboot_firmware.c (right):

http://codereview.chromium.org/2802002/diff/2001/3004#newcode17
vboot_firmware/lib/vboot_firmware.c:17: static DigestContext ctx;
On 2010/06/15 21:00:41, Bill Richardson wrote:
> Do we really want static/global variables inside a library? It doesn't seem
like
> a great idea. I thought that's what the various *Params structs were for.
> 

I agree it's not really pretty.

UpdateFirmwareBodyHash2() will be called from deep inside the UEFI code which
does read-and-decompress-fv, where there's no good way for the UEFI code to get
a pointer to the params struct to pass it back to us.

I added a flag so at least we won't attempt to update the hash if we're called
outside LoadFirmware().

[Bill Richardson, 2010-06-15 21:48:09]

Okay then. LGTM.