ARM L2C-310 cache controller interconnect question
Christopher Kenna <[email protected]> Tue, 9 Oct 2012 21:35:06 -0400
| Newsgroups | gmane.linux.ports.arm.general |
|---|---|
| Message-ID | <CACdM24fcvKzi+PK0F8tsf3D6vzg+WcR07wbKpp1cdHOHijHLxA@mail.gmail.com> |
Hello, I was reading about cache lockdown in ARM's L2C-310 (or PL310) cache controller in [1]. Based on the documentation, it looks like the cache lockdown by master (CPU) is implemented depending upon how the signal AyUSERSx[7:5] is connected. It looks like it makes the most sense to hook up this signal like in the example, "System including four Cortex-A9 MPCore processors and L2C-310" so that the lockdown register 0 determines what cache ways are available for CPU0, lockdown register 1 determines what cache ways are available for CPU 1, etc. The problem is that I cannot find out how AyUSERSx[7:5] is connected in my system, and I cannot determine if cache lockdown by master is supported (the documentation says that this is an optional feature). I believe that these options are determined by the SoC manufacturers, and are not available to the public. I am specifically curious about how this is set up in the PandaBoard ES, which uses a TI OMAP4460, and the ODROID-X, which uses a Samsung Exynos 4412. Do you know where I can find this information out? Thanks very much, Chris Kenna [1] http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0246h/ch02s03s06.html