Over a year ago in a commit ac96a1665a ("bcm53xx: update Disable MMU
and Dcache during decompression") we switched to Florian's patch for
workarounding CFE bug. The main difference was using a part of existing
__armv7_mmu_cache_flush instead of implementing flushing separately.
This worked well for Northstar devices but doesn't work for BCM53573 as
these devices simply don't start booting with Florian's patch. It's
because of the ldmfd ASM instruction in the __armv7_mmu_cache_flush.
So this commit switches back to using standalone implementation. This
time instead of copying Broadcom's copy of cache-v7.S, we just make a
copy of the original file on our own. Unfortunately we can't cross-dir
compile cache-v7.S from ../../mm/ as that one also adds __INITDATA with
define_cache_functions v7 which would just trigger
> Error: unrecognized/unsupported machine ID (r1 = 0x0000007f).
The only real change Broadcom did in copied .S file was modifying mcr
instruction to use c6 instead of c14. It isn't clear to me if we really
need it, but let's use it for now.
By the way we also update commit message of the
[PATCH] ARM: BCM5301X: Disable MMU and Dcache during decompression
This makes kernel booting on BCM53573 successfully.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
It was adding unused variable:
drivers/mtd/devices/m25p80.c: In function 'm25p80_write':
drivers/mtd/devices/m25p80.c:85:6: warning: unused variable 'i' [-Wunused-variable]
int i;
^
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
It adds and uses ILP clock that requires some other work (in progress)
for upstreaming it. Other than that it adds a quick fix for bcma to add
serial flash before trying to read SPROM.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
It wasn't accepted upstream as there was a discusson on Northstar vs.
BCM53573. Once we get a new ARM arch Kconfig entry it should be
possible to upstream it.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This generates proper images when using CONFIG_TARGET_MULTI_PROFILE and
CONFIG_TARGET_PER_DEVICE_ROOTFS.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Thanks to this images for SoftMAC devices don't get brcmfmac anymore and
b43 is added for devices with (quite poor) support only.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This allows building images for selected devices with brcmfmac only
(without b43 which is for SoftMAC devices).
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Reviewed-by: Jo-Philipp Wich <jo@mein.io>
It is used by a core build template, so the variable should be
initialized and added to DEVICE_VARS in the core.
Same for DEVICE_DTS_DIR
Signed-off-by: Felix Fietkau <nbd@nbd.name>
We call all commands normally, with standard echoing. It's useful for
debugging with V=s. Don't make lzma compression an exception, it's a bit
confusing this way.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
We don't want checksum to cover any part of UBI as even its part with
SquashFS may be changed due to e.g. flash wearing.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
This reverts commit b3be33f135.
CFE is known to fail in some non-standard cases, e.g. when using kernel
or format different that what was tested by Broadcom. This kernel change
triggered some problem with booting OpenWrt kernel stored in Seama.
As long as Seama checksum was covering enough data, it was working fine.
We need to change it however, because calculating MD5 over part of UBI
containing SquashFS is unsafe. UBI may move PEBs depending on flash
wearing level which would break CFE booting the image.
For some reason this kernel change was breaking CFE. Calculating MD5
over 0x47ffc0 B data or less (there is 0x40 B long header) was stopping
booting process at:
Starting program at 0x00008000
As our kernel is usually 0x3fffc0 B it was affected by this problem.
Reverting this change fixes sysupgrade (which already uses kernel size
for MD5 calculation) and will allow us to adjust "fixseama" command call
on the first boot.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
In kernel 4.7 there is upstreamed b53 driver using (mostly?) the same
symbols as our b53 does. Change our symbols so both drivers can coexist
in kernel tree.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Acked-by: Jonas Gorski <jonas.gorski@gmail.com>
Both scripts modified by this patch were added by me. First of all I
incorrectly added OpenWrt as Copyright holder. It was wrong because:
1) I simply can't transfer my moral rights according to the Polish law
2) Transfering copyrights (economic rights) requires an agreement which
I didn't sign with OpenWrt(.org).
Other than that I don't find these trivial scripts important enough to
put info about *my* copyrights in a header so this patch just drops them
completely.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
For some time now m25p80 supports 32 MiB flashes and we just needed to
patch spi-nor to support JEDEC incompatible w25q128. Also by switching
to m25p80 we gain accelerated SPI flash reads.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49260
It was already increased for NAND devices so do the same for the ones
with serial memory.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49207
This reverts commit r48961. It didn't fix PCIe host driver issue as
there are more aborts triggered by it.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49120
4.3 isn't oficially used anymore but it may be useful for debugging PCI
host driver regression. SPROM driver doesn't seem to work correctly with
4.3:
[ 2.336686] bcma: bus0: Using fallback SPROM failed (err -2)
[ 2.342366] bcma: bus0: No SPROM available
but it at least compiles now.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 49113
Initially 4.4 didn't work on bcm53xx because network was breaking few
seconds after setting interfaces. It's fixed since r48059 (switching
from 4.4-rc5 to 4.4-rc7).
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48510
