Thread (2 messages) flat view 2 messages, 2 authors, 2007-08-06

RE: [2/2] 2.6.23-rc2: known regressions with patches

From: "Luck, Tony" <tony.luck@intel.com>
Date: 2007-08-06 18:45:48
Also in: linux-scsi, lkml

quoted
Caused-By       : Thomas Renninger [off-list ref]
                  commit 29b71a1ca74491fab9fed09e9d835d840d042690
A patch was sent to Tony. AFAIK it got accepted, not sure whether it
already is in any and which git tree...
The suggested patch adds manual padding to the acpi_device_id structure
definition in include/linux/mod_devicetable.h  I didn't take it, and it
appears that nobody else did either.  It is not in Linus' tree (as of 2.6.23-rc2).

I expressed doubts about whether this is the right fix.  The problem
is that when cross-compiling a locally compiled utility makes a size &
alignment check.  This is bogus.  We shouldn't care whether this structure
compiles to the same size/alignment as the kernel that will use on the
target platform.

Is fixing this the right way (make the scripts/mod/file2alias.c understand
target alignment rules in a cross environment) just too hideous to
contemplate ... and we should just sacrifice 7 bytes of padding in
order to keep life simple?

-Tony
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help