| Message ID | 1504029821-38267-1-git-send-email-f.fainelli@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 08/29/2017 11:03 AM, Florian Fainelli wrote: > Building a big-endian kernel for ARCH_BRCMSTB revealed that we would not > be correctly polling for the right bit in the busyuart macro, turns out > there are a few transformations needed to work with big-endian kernels. > First we need to swap the value we read from SUN_TOP_CTRL to properly > compare it against our local tables. Then, just like 8250.S we need to > swap the value before storing it, and conversely swap it after a load. > > Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Applied to soc/next.
diff --git a/arch/arm/include/debug/brcmstb.S b/arch/arm/include/debug/brcmstb.S index 52aaed2b936f..c826f15d2f80 100644 --- a/arch/arm/include/debug/brcmstb.S +++ b/arch/arm/include/debug/brcmstb.S @@ -58,6 +58,7 @@ /* Check SUN_TOP_CTRL base */ ldr \rp, =SUN_TOP_CTRL_BASE @ load SUN_TOP_CTRL PA ldr \rv, [\rp, #0] @ get register contents +ARM_BE8( rev \rv, \rv ) and \rv, \rv, #0xffffff00 @ strip revision bits [7:0] /* Chip specific detection starts here */ @@ -98,11 +99,13 @@ .endm .macro store, rd, rx:vararg +ARM_BE8( rev \rd, \rd ) str \rd, \rx .endm .macro load, rd, rx:vararg ldr \rd, \rx +ARM_BE8( rev \rd, \rd ) .endm .macro senduart,rd,rx
Building a big-endian kernel for ARCH_BRCMSTB revealed that we would not be correctly polling for the right bit in the busyuart macro, turns out there are a few transformations needed to work with big-endian kernels. First we need to swap the value we read from SUN_TOP_CTRL to properly compare it against our local tables. Then, just like 8250.S we need to swap the value before storing it, and conversely swap it after a load. Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> --- arch/arm/include/debug/brcmstb.S | 3 +++ 1 file changed, 3 insertions(+)