Mike Rapoport <rppt@xxxxxxxxxx> writes: > From: Mike Rapoport <rppt@xxxxxxxxxxxxx> > > Some powerpc platforms (e.g. 85xx) limit DMA-able memory way below 4G. If a > system has more physical memory than this limit, the swiotlb buffer is not > addressable because it is allocated from memblock using top-down mode. > > Force memblock to bottom-up mode before calling swiotlb_init() to ensure > that the swiotlb buffer is DMA-able. > > Link: https://lkml.kernel.org/r/F1EBB706-73DF-430E-9020-C214EC8ED5DA@xxxxxxxxxxx This wasn't bisected, but I thought it was a regression. Do we know what commit caused it? Was it 25078dc1f74b ("powerpc: use mm zones more sensibly") ? Or was that a red herring? cheers > Reported-by: Christian Zigotzky <chzigotzky@xxxxxxxxxxx> > Signed-off-by: Mike Rapoport <rppt@xxxxxxxxxxxxx> > Cc: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx> > Cc: Christoph Hellwig <hch@xxxxxx> > Cc: Darren Stevens <darren@xxxxxxxxxxxxxxxx> > Cc: mad skateman <madskateman@xxxxxxxxx> > Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx> > Cc: Nicolas Saenz Julienne <nsaenzjulienne@xxxxxxx> > Cc: Paul Mackerras <paulus@xxxxxxxxx> > Cc: Robin Murphy <robin.murphy@xxxxxxx> > Cc: Rob Herring <robh+dt@xxxxxxxxxx> > --- > arch/powerpc/mm/mem.c | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c > index be941d382c8d..14c2c53e3f9e 100644 > --- a/arch/powerpc/mm/mem.c > +++ b/arch/powerpc/mm/mem.c > @@ -260,6 +260,14 @@ void __init mem_init(void) > BUILD_BUG_ON(MMU_PAGE_COUNT > 16); > > #ifdef CONFIG_SWIOTLB > + /* > + * Some platforms (e.g. 85xx) limit DMA-able memory way below > + * 4G. We force memblock to bottom-up mode to ensure that the > + * memory allocated in swiotlb_init() is DMA-able. > + * As it's the last memblock allocation, no need to reset it > + * back to to-down. > + */ > + memblock_set_bottom_up(true); > swiotlb_init(0); > #endif > > -- > 2.24.0
