Less file system overhead on the flash side and no memory used to manage
the structures needed to insmod / rmmod.
Exactly. Maybe also a tighter memory layout since each module doesn't
have to be in its own set of ELF sections, IIUC.
Also, the linkage would be (presumably) static rather than
While I have seen the size listed in lsmod and /proc/modules, I don't
really know what it represents. I am assuming that it means code size
and not memory.
I suspect so as well. Still, it may point to an opportunity.
Make sure that you are counting the size of the files in /lib/modules
Hi,
I wonder IF linux can be made small enough for some of the
embedded chips to boot and then attach an SD card as the file system?
warm regards,
wiz
On Mon, 10 May 2010 edgar.sol...@web.de wrote:
On 10.05.2010 23:47, Bernhard Loos wrote:
2010/5/10 Bernhard Loos
On 11.05.2010 17:25, Stefan Monnier wrote:
but wouldn't the increase in the kernel image actually equal the
decrease in the squash image and therefore the size of the rootfs_data
stay the same? Both are lzma compressed.
I expect that a kernel with some modules built-in will be smaller
(both in
but wouldn't the increase in the kernel image actually equal the
decrease in the squash image and therefore the size of the rootfs_data
stay the same? Both are lzma compressed.
I expect that a kernel with some modules built-in will be smaller
(both in terms of flash space and in terms of RAM
On 5/11/2010 11:27 AM, edgar.sol...@web.de wrote:
On 11.05.2010 17:25, Stefan Monnier wrote:
but wouldn't the increase in the kernel image actually equal the
decrease in the squash image and therefore the size of the rootfs_data
stay the same? Both are lzma compressed.
I expect that a kernel
there is partly in
https://dev.openwrt.org/browser/trunk/include/kernel-defaults.mk
line 101
you can actually add
CONFIG_KERNEL_*
entries to your .config and they are copied over as CONFIG_* to the
kernel config. All that's missing is a menuconfig interface for that.
But at least for routers
It may not downsize the packages themselves, but moving kernel mods from
rootfs to the kernel image will reduce the rootfs size. Leaving more
space for JFFS2 rootfs_data.
Having a K setting for kernel options would be great.
Will try to have a look at it during the next weeks.
Thanks for all
but wouldn't the increase in the kernel image actually equal the
decrease in the squash image and therefore the size of the rootfs_data
stay the same? Both are lzma compressed.
..ede
On 10.05.2010 19:16, Matthias Buecher / Germany wrote:
It may not downsize the packages themselves, but moving
Kernel and rootfs are in two different mtd partitions on the WRT54G:
# dmesg
...
Creating 5 MTD partitions on Physically mapped flash:
0x-0x0004 : cfe
0x0004-0x003f : linux
0x000bc000-0x0021 : rootfs
mtd: partition rootfs doesn't start on an erase block boundary --
force
are these sizes fixed or calculated according the space requirement?
Looks like the linux size is fixed, what is the maximum size for a
kernel on wrt54g?
.. ede
On 10.05.2010 23:28, Matthias Buecher / Germany wrote:
Kernel and rootfs are in two different mtd partitions on the WRT54G:
# dmesg
The linux partition spans over the kernel and the complete rootfs for
flashing.
The maximum kernel size is 0x000bc000 (begin of rootfs) minus
0x0004 (begin of linux) equals 0x0007c000 (496KB).
Maddes
On 10.05.2010 23:34, edgar.sol...@web.de wrote:
are these sizes fixed or calculated
2010/5/10 Matthias Buecher / Germany m...@maddes.net:
The linux partition spans over the kernel and the complete rootfs for
flashing.
The maximum kernel size is 0x000bc000 (begin of rootfs) minus
0x0004 (begin of linux) equals 0x0007c000 (496KB).
Maddes
This is not the maximum kernel
Then it'd make total sense to stuff in whatever fits there .. thanks for
the enlightenment, ede
On 10.05.2010 23:38, Matthias Buecher / Germany wrote:
The linux partition spans over the kernel and the complete rootfs for
flashing.
The maximum kernel size is 0x000bc000 (begin of rootfs) minus
2010/5/10 Bernhard Loos bernhardl...@googlemail.com:
2010/5/10 Matthias Buecher / Germany m...@maddes.net:
The linux partition spans over the kernel and the complete rootfs for
flashing.
The maximum kernel size is 0x000bc000 (begin of rootfs) minus
0x0004 (begin of linux) equals
On 10.05.2010 23:47, Bernhard Loos wrote:
2010/5/10 Bernhard Loos bernhardl...@googlemail.com:
2010/5/10 Matthias Buecher / Germany m...@maddes.net:
The linux partition spans over the kernel and the complete rootfs for
flashing.
The maximum kernel size is 0x000bc000 (begin of rootfs) minus
On 10.05.2010 23:47, Bernhard Loos wrote:
2010/5/10 Bernhard Loos bernhardl...@googlemail.com:
2010/5/10 Matthias Buecher / Germany m...@maddes.net:
The linux partition spans over the kernel and the complete rootfs for
flashing.
The maximum kernel size is 0x000bc000 (begin of rootfs) minus
I could finally install everything onto my 4MB WRT54g v2.2.
Avoided ntpclient plus dependencies with the patch from #7316.
Added BusyBox's ether-wake to the build.
Put all kmods (2) I needed into the kernel.
This freed up enough JFFS2 space to normally install OpenVPN and a
web-interface (here
Put all kmods (2) I needed into the kernel.
How?
Stefan
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
That's how it works. Y adds kernel modules to the kernel, and programs
to the image.
..ede
On 09.05.2010 18:10, Matthias Buecher / Germany wrote:
In 'make menuconfig' I included them with 'Y' instead of 'M'.
According to my (newbie) knowledge that adds them to the kernel image.
Can somebody
Actually, Y in menuconfig doesn't add the modules into the kernel,
they will still be kernel modules. It only places the modules directly
in the image and is in no way different from Y for programs. M creates
packages for the modules/program without placing it in the image.
If you want to compile
true, my bad, actually one can see it in the end of a make run, how the
packages (incl. kmods) are installed to the image filesystem.
thanks ede
On 09.05.2010 18:19, Bernhard Loos wrote:
Actually, Y in menuconfig doesn't add the modules into the kernel,
they will still be kernel modules. It
If you are really bored, you could use gcc 4.5 and link time optimization.
Support for it is already enabled in the openwrt gcc, but you have to
add the compiler flags to use it to packages and/or the kernel. In
theory, it could reduce the binary size, but I have no idea how well
this will work in
Re-checked the configs from my last build (available at
ftp://ftp.maddes.net/openwrt/backfire/brcm-2.4/build_21402/squashfs/)
In the menu config it is CONFIG_PACKAGE_kmod-tun=y, but in the kernel
and final/used config it is CONFIG_TUN=m.
So the better squashfs compression was what helped me.
Can I save space in the flash mem when I add all needed packages
directly to the image?
Or is the JFFS2 compression as good as the compression of the read-only
squashfs?
I'm also going to replace ntpclient with rdate.
As the clock of the WRT54G is very inaccurate it will not be sufficient
to
* Matthias Buecher / Germany m...@maddes.net [04.05.2010 16:30]:
Can I save space in the flash mem when I add all needed packages
directly to the image?
yes.
Or is the JFFS2 compression as good as the compression of the read-only
squashfs?
no. bye, Bastian.
signature.asc
Description:
I wanted to upgrade my mom's WRT54G v2.2 from WR 0.9 to Backfire 10.03.
But during installation the flash space got full and OpenVPN couldn't be
installed.
First I built an image without luci, then all packages fit into the
flash mem.
But when installing webif I got the same disk full error for
* Matthias Buecher / Germany m...@maddes.net [30.04.2010 22:30]:
Can OpenVPN be compiled to a smaller size?
nossl/nolzo - phew!
bye, Bastian
signature.asc
Description: Digital signature
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
28 matches
Mail list logo