On Mon, Dec 21, 2015 at 03:05:25PM +0800, Pingbo Wen wrote:

Does anyone run KASan on db410c successfully? I merged arm64-upstream branch over integration-linux-qcomlt, and enable KASan feature, compiled with gcc 5.1. But the kernel hanged before start_kernel(), no log output.

Linus (CCed) was looking at testing KASan on ARMv8. I think he was using Juno rather than Dragonboard but he might have some hints about gotchas.

On Mon, Jan 4, 2016 at 9:20 PM, Mark Brown broonie@kernel.org wrote:

On Mon, Jan 04, 2016 at 08:17:04PM +0000, Mark Brown wrote:

...

On Mon, Dec 21, 2015 at 03:05:25PM +0800, Pingbo Wen wrote:

...

Does anyone run KASan on db410c successfully? I merged arm64-upstream branch over integration-linux-qcomlt, and enable KASan feature, compiled with gcc 5.1. But the kernel hanged before start_kernel(), no log output.

Linus (CCed) was looking at testing KASan on ARMv8. I think he was using Juno rather than Dragonboard but he might have some hints about gotchas.

Really CCing Linus this time (hopefully).

There are a few things to think about with KASan, but on v8 it should mostly work out of the box.

I configure it like so:

$(CURDIR)/scripts/config --file $(config_file) \ --enable KASAN \ --enable KASAN_OUTLINE \ --enable STACKTRACE \ --enable SLUB_DEBUG_ON \ --enable TEST_KASAN

I think you can use KASAN_INLINE with GCC 5.1 too.

If need be I can test linux-next on the Juno.

Yours, Linus Walleij

On Wed, Jan 6, 2016 at 10:17 AM, Pingbo Wen pingbo.wen@linaro.org wrote:

Hi, Mark & Linus

Thanks for your reply.

...

在 2016年1月5日，17:18，Linus Walleij linus.walleij@linaro.org 写道： There are a few things to think about with KASan, but on v8 it should mostly work out of the box.

I configure it like so:

   $(CURDIR)/scripts/config --file $(config_file) \
   --enable KASAN \
   --enable KASAN_OUTLINE \
   --enable STACKTRACE \
   --enable SLUB_DEBUG_ON \
   --enable TEST_KASAN

I think you can use KASAN_INLINE with GCC 5.1 too.

The result is same, kernel hang in dark:(

I’m going to debug it via JTAG, but should wait for my flyswatter2 JTAG debugger first(I don’t know if it works in armv8, the OpenOCD have some limit support for it). I will update this thread if I get something new:)

Note that the kernel takes a *LONG* time to boot on KASan, so have a lot of patience. Have a coffee and come back. Also note that it populates all shadow memory before even getting to start_kernel() so yeah, debugging using JTAG and/or emulators such as QEMU may be necessary to get anywhere.

Yours, Linus Walleij