2021-02-05 16:05:37

by Andrey Konovalov

Subject: [PATCH v2 12/12] arm64: kasan: export MTE symbols for KASAN tests

Export mte_enable_kernel_sync() and mte_set_report_once() to fix:

ERROR: modpost: "mte_enable_kernel_sync" [lib/test_kasan.ko] undefined!
ERROR: modpost: "mte_set_report_once" [lib/test_kasan.ko] undefined!

Reviewed-by: Catalin Marinas <[email protected]>
Signed-off-by: Andrey Konovalov <[email protected]>
---
arch/arm64/kernel/mte.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/arch/arm64/kernel/mte.c b/arch/arm64/kernel/mte.c
index 8b27b70e1aac..fc753fb77940 100644
--- a/arch/arm64/kernel/mte.c
+++ b/arch/arm64/kernel/mte.c
@@ -120,6 +120,7 @@ void mte_enable_kernel_sync(void)
{
__mte_enable_kernel("synchronous", SCTLR_ELx_TCF_SYNC);
}
+EXPORT_SYMBOL_GPL(mte_enable_kernel_sync);

void mte_enable_kernel_async(void)
{
@@ -130,6 +131,7 @@ void mte_set_report_once(bool state)
{
WRITE_ONCE(report_fault_once, state);
}
+EXPORT_SYMBOL_GPL(mte_set_report_once);

bool mte_report_once(void)
{
--
2.30.0.365.g02bc693789-goog

2021-02-08 19:16:00

by Christoph Hellwig

Subject: Re: [PATCH v2 12/12] arm64: kasan: export MTE symbols for KASAN tests

On Fri, Feb 05, 2021 at 04:39:13PM +0100, Andrey Konovalov wrote:
> Export mte_enable_kernel_sync() and mte_set_report_once() to fix:
>
> ERROR: modpost: "mte_enable_kernel_sync" [lib/test_kasan.ko] undefined!
> ERROR: modpost: "mte_set_report_once" [lib/test_kasan.ko] undefined!

Please put this under an ifdef for the testing option that pull the
symbols in.

2021-02-10 05:38:09

by Andrey Konovalov

Subject: Re: [PATCH v2 12/12] arm64: kasan: export MTE symbols for KASAN tests

On Mon, Feb 8, 2021 at 6:42 PM Christoph Hellwig <[email protected]> wrote:
>
> On Fri, Feb 05, 2021 at 04:39:13PM +0100, Andrey Konovalov wrote:
> > Export mte_enable_kernel_sync() and mte_set_report_once() to fix:
> >
> > ERROR: modpost: "mte_enable_kernel_sync" [lib/test_kasan.ko] undefined!
> > ERROR: modpost: "mte_set_report_once" [lib/test_kasan.ko] undefined!
>
> Please put this under an ifdef for the testing option that pull the
> symbols in.

Sure. I'll send this change as a fix-up patch instead of resending the
whole series.

Thanks!