On 06/18/2012 07:13 AM, R, Sricharan wrote:
Hi Tom,
On Fri, Jun 15, 2012 at 8:50 PM, Tom Rini trini@ti.com wrote:
On 06/15/2012 08:18 AM, R, Sricharan wrote:
Hi, [snip..]
> If it is a problem with unaligned regions, then that is the only > thing to be fixed > right ?. Just trying to understand why this change is required ?
The problem is that within the USB/network/filesystem stacks we have a lot of not cache safe alignments apparently. Without this every '#' of a tftp gives a screen full of error printfs. So tftp'ing a kernel takes minutes, not seconds, to complete.
Ok,but when buffers are not aligned isn't that you are going to have coherency problems ?
Exactly. USB requires dcache to be turned off today. With it turned off it spams these messages making tftp to an aligned address take minutes because the console is busy printing all of these messages about flushes it didn't need to do. The alternative is to build omap4/omap5/etc with dcache disabled or USB disabled.
Ok, but who is calling cache maintenance in your case afterturning off caches ?
The USB stack currently.
Ok, in that case if
- Alignment issue has to be fixed USB stack, (or)
This is the correct, long term solution.
- Better to turn off caches when say USB-EHCI is enabled ?
Doing that is what showed the issue to start with. Running 'dcache off' doesn't turn off attempted cache flushing.
(or)
- Turn of caches until cache library is fixed so that we will not require any maintenance after turning it off. ( I will work on this some time, this week)
This is also a good thing. It just needs to make sure attempted flushes don't spam the console, when cache is off :)
I am not sure which one to do, but definitely we should not prevent it from executing as in the above patch. This may lead to other issues.
OK. I'll see how badly boot time goes up when dcache is off to start with on a few boards.