OMAP baseline test results for v3.19-rc1
Aaro Koskinen
aaro.koskinen at iki.fi
Sun Jan 4 11:55:51 PST 2015
Hi,
On Sun, Jan 04, 2015 at 05:46:31PM +0100, Arnd Bergmann wrote:
> On Saturday 03 January 2015 15:44:38 Aaro Koskinen wrote:
> > On Sat, Jan 03, 2015 at 01:51:53PM +0100, Arnd Bergmann wrote:
> > > On Saturday 03 January 2015 07:21:49 Paul Walmsley wrote:
> > > > Another ~300KB kernel object size increase for omap2plus_defconfig
> > > > kernels.
> > >
> > > 300kb seems like a lot for a single release. Have you looked into
> > > where this is coming from?
> >
> > In 3.19-rc1, all initrd decompressors are by default enabled...
>
> Interesting, this seems to be the opposite of the intention of the
> change that caused it, which tried to let you disable the algorithms
> without selecting CONFIG_EXPERT. For some reason that is impossible
> to know now, they already defaulted to 'n' when EXPERT was set, but
> now they default to 'y'.
>
> Does this indeed cause the entire change in size?
Actually, no. It seems a bigger contributor is enabling SATA
in omap2plus_defconfig. But there still something else... Below is some
data based on quick checks.
v3.18 / omap2plus_defconfig:
text data bss dec hex filename
8811980 790500 8423536 18026016 1130e20 vmlinux
v3.19-rc1 / omap2plus_defconfig:
text data bss dec hex filename
9093858 800316 8419072 18313246 117701e vmlinux
v3.19-rc1 / omap2plus_defconfig + revert ec72c666fb34 ("usr/Kconfig: make
initrd compression algorithm selection not expert"):
text data bss dec hex filename
9070939 800260 8419072 18290271 117165f vmlinux
v3.19-rc1 / omap2plus_defconfig + revert 661ea91b676b ("ARM:
omap2plus_defconfig: Enable AHCI_PLATFORM driver"):
text data bss dec hex filename
8928148 792700 8414784 18135632 114ba50 vmlinux
v3.19-rc1 / omap2plus_defconfig + revert ec72c666fb34 & 661ea91b676b:
text data bss dec hex filename
8901129 792644 8414784 18108557 114508d vmlinux
A.
More information about the linux-arm-kernel
mailing list