| Message ID | 1345222980-16965-1-git-send-email-jonathan.austin@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, 17 Aug 2012, Jonathan Austin wrote: > With !HIGHMEM, sanity_check_meminfo checks for banks that completely or > partially overlap the vmalloc region. The test for partial overlap checks > __va(bank->start + bank->size) > vmalloc_min. This is not appropriate if > there is a non-linear translation between virtual and physical addresses, > as bank->start + bank->size is actually in the bank following the one being > interrogated. > > In most cases, even when using SPARSEMEM, this is not problematic as the > subsequent bank will start at a higher va than the one in question. However > if the physical to virtual address conversion is not monotonic increasing, > the incorrect test could result in a bank not being truncated when it > should be. > > This patch ensures we perform the va-pa conversion on memory from the > bank we are interested in, not the following one. > > Reported-by: 湛振波 (Steve) <zhanzhenbo@gmail.com> > Signed-off-by: Jonathan Austin <jonathan.austin@arm.com> Acked-by: Nicolas Pitre <nico@linaro.org> > --- > > This is modified based on Russell's feedback in the original > discussion (Why the region area don't decrease 1 in function > sanity_check_meminfo?) > > arch/arm/mm/mmu.c | 4 ++-- > 1 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c > index 4c2d045..dbd9ff6 100644 > --- a/arch/arm/mm/mmu.c > +++ b/arch/arm/mm/mmu.c > @@ -961,8 +961,8 @@ void __init sanity_check_meminfo(void) > * Check whether this memory bank would partially overlap > * the vmalloc area. > */ > - if (__va(bank->start + bank->size) > vmalloc_min || > - __va(bank->start + bank->size) < __va(bank->start)) { > + if (__va(bank->start + bank->size - 1) >= vmalloc_min || > + __va(bank->start + bank->size - 1) <= __va(bank->start)) { > unsigned long newsize = vmalloc_min - __va(bank->start); > printk(KERN_NOTICE "Truncating RAM at %.8llx-%.8llx " > "to -%.8llx (vmalloc region overlap).\n", > -- > 1.7.1 > > > > _______________________________________________ > linux-arm-kernel mailing list > linux-arm-kernel@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/linux-arm-kernel >
diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c index 4c2d045..dbd9ff6 100644 --- a/arch/arm/mm/mmu.c +++ b/arch/arm/mm/mmu.c @@ -961,8 +961,8 @@ void __init sanity_check_meminfo(void) * Check whether this memory bank would partially overlap * the vmalloc area. */ - if (__va(bank->start + bank->size) > vmalloc_min || - __va(bank->start + bank->size) < __va(bank->start)) { + if (__va(bank->start + bank->size - 1) >= vmalloc_min || + __va(bank->start + bank->size - 1) <= __va(bank->start)) { unsigned long newsize = vmalloc_min - __va(bank->start); printk(KERN_NOTICE "Truncating RAM at %.8llx-%.8llx " "to -%.8llx (vmalloc region overlap).\n",
With !HIGHMEM, sanity_check_meminfo checks for banks that completely or partially overlap the vmalloc region. The test for partial overlap checks __va(bank->start + bank->size) > vmalloc_min. This is not appropriate if there is a non-linear translation between virtual and physical addresses, as bank->start + bank->size is actually in the bank following the one being interrogated. In most cases, even when using SPARSEMEM, this is not problematic as the subsequent bank will start at a higher va than the one in question. However if the physical to virtual address conversion is not monotonic increasing, the incorrect test could result in a bank not being truncated when it should be. This patch ensures we perform the va-pa conversion on memory from the bank we are interested in, not the following one. Reported-by: 湛振波 (Steve) <zhanzhenbo@gmail.com> Signed-off-by: Jonathan Austin <jonathan.austin@arm.com> --- This is modified based on Russell's feedback in the original discussion (Why the region area don't decrease 1 in function sanity_check_meminfo?) arch/arm/mm/mmu.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)