[Bug build/20463] New: Compile error when configure using --param arg=value in CFLAGS

[snorkellingcactus at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20463

            Bug ID: 20463
           Summary: Compile error when configure using --param arg=value
                    in CFLAGS
           Product: glibc
           Version: 2.23
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: build
          Assignee: unassigned at sourceware dot org
          Reporter: snorkellingcactus at gmail dot com
                CC: carlos at redhat dot com
  Target Milestone: ---

Created attachment 9431
  --> https://sourceware.org/bugzilla/attachment.cgi?id=9431&action=edit
GCC Version

Hi.
 I have obtained my custom cflags, trough a script that shows the output of gcc
and -march=native when it call cc1. In these cflags, are cpu-specific options,
like --param l1-cache-size=512.
 When i use these cflags, i get the error below, where i see that cflags are
very similar ( almost duplicated , maybe due to the $(CFLAGS) $(CPPFLAGS) like
lines in Makerules and stuff ).

 However, the first block of cflags, miss the --param parts, and put one
before.

Here a simplification:

when running gcc -S -o ..etc

#First (missing) block of C[XX]FLAGS:

-std=gnu11 -fgnu89-inline  --param -O2 -Wall -Wundef l1-cache-line-size=32
l1-cache-size=16 l2-cache-size=256

#Second (valid) block of C[XX]FLAGS:

-std=gnu11 -fgnu89-inline -O2 -Wall -Wundef --param l1-cache-line-size=32
--param l1-cache-size=16 --param l2-cache-size=256

Result error:

i686-pc-linux-gnu-gcc: error: l1-cache-line-size=32: No such file or directory
i686-pc-linux-gnu-gcc: error: l1-cache-size=16: No such file or directory
i686-pc-linux-gnu-gcc: error: l2-cache-size=256: No such file or directory


Note: I now that my custom CFLAGS are some discouraged, however, the flags
related to de issue ( --param ) are the same that gcc uses.

I'm replacing -march=native by gcc flags, by enable my compiler wrapper
(icecream) to distribute the workload, but the same results are obtained
without the wrapper.

Thanks you

-- 
You are receiving this mail because:
You are on the CC list for the bug.