| Message ID | 20171102092155.737712-1-arnd@arndb.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 02/11/17 09:21, Arnd Bergmann wrote: > The reworked MPU code produces a new warning in some configurations, > presumably starting with the code move after the compiler now makes > different inlining decisions: > > arch/arm/mm/pmsa-v7.c: In function 'adjust_lowmem_bounds_mpu': > arch/arm/mm/pmsa-v7.c:310:5: error: 'specified_mem_size' may be used uninitialized in this function [-Werror=maybe-uninitialized] > > This appears to be harmless, as we know that there is always at > least one memblock, and the only way this could get triggered is > if the for_each_memblock() loop was never entered. > > I could not come up with a better workaround than initializing > the specified_mem_size to zero, but at least that is the value > that the variable would have in the hypothetical case of no > memblocks. > > Fixes: 877ec119dbbf ("ARM: 8706/1: NOMMU: Move out MPU setup in separate module") > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > Vladimir, if this looks good to you, can you forward it to Russell's > patch tracker, or otherwise suggest a different fix? Accepted as patch 8719/1. Not sure if Russell will pick it up or not... anyway, thanks for the patch! Vladimir > --- > arch/arm/mm/pmsa-v7.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm/mm/pmsa-v7.c b/arch/arm/mm/pmsa-v7.c > index 106ae1c435a3..976df60ac426 100644 > --- a/arch/arm/mm/pmsa-v7.c > +++ b/arch/arm/mm/pmsa-v7.c > @@ -234,7 +234,7 @@ static int __init allocate_region(phys_addr_t base, phys_addr_t size, > /* MPU initialisation functions */ > void __init adjust_lowmem_bounds_mpu(void) > { > - phys_addr_t specified_mem_size, total_mem_size = 0; > + phys_addr_t specified_mem_size = 0, total_mem_size = 0; > struct memblock_region *reg; > bool first = true; > phys_addr_t mem_start; >
On Thu, Nov 02, 2017 at 12:25:47PM +0000, Vladimir Murzin wrote: > On 02/11/17 09:21, Arnd Bergmann wrote: > > The reworked MPU code produces a new warning in some configurations, > > presumably starting with the code move after the compiler now makes > > different inlining decisions: > > > > arch/arm/mm/pmsa-v7.c: In function 'adjust_lowmem_bounds_mpu': > > arch/arm/mm/pmsa-v7.c:310:5: error: 'specified_mem_size' may be used uninitialized in this function [-Werror=maybe-uninitialized] > > > > This appears to be harmless, as we know that there is always at > > least one memblock, and the only way this could get triggered is > > if the for_each_memblock() loop was never entered. > > > > I could not come up with a better workaround than initializing > > the specified_mem_size to zero, but at least that is the value > > that the variable would have in the hypothetical case of no > > memblocks. > > > > Fixes: 877ec119dbbf ("ARM: 8706/1: NOMMU: Move out MPU setup in separate module") > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > --- > > Vladimir, if this looks good to you, can you forward it to Russell's > > patch tracker, or otherwise suggest a different fix? > > Accepted as patch 8719/1. Not sure if Russell will pick it up or not... anyway, > thanks for the patch! I'd prefer not to at this stage. Same things that apply that I said in the "ARM: early_printk: use printascii() rather than printch()" thread - I need to ensure that what I have is stable before 4.14 is released, so I'm not accepting anything further unless it is _really_ urgent. This isn't urgent.
On 02/11/17 13:07, Russell King - ARM Linux wrote: > On Thu, Nov 02, 2017 at 12:25:47PM +0000, Vladimir Murzin wrote: >> On 02/11/17 09:21, Arnd Bergmann wrote: >>> The reworked MPU code produces a new warning in some configurations, >>> presumably starting with the code move after the compiler now makes >>> different inlining decisions: >>> >>> arch/arm/mm/pmsa-v7.c: In function 'adjust_lowmem_bounds_mpu': >>> arch/arm/mm/pmsa-v7.c:310:5: error: 'specified_mem_size' may be used uninitialized in this function [-Werror=maybe-uninitialized] >>> >>> This appears to be harmless, as we know that there is always at >>> least one memblock, and the only way this could get triggered is >>> if the for_each_memblock() loop was never entered. >>> >>> I could not come up with a better workaround than initializing >>> the specified_mem_size to zero, but at least that is the value >>> that the variable would have in the hypothetical case of no >>> memblocks. >>> >>> Fixes: 877ec119dbbf ("ARM: 8706/1: NOMMU: Move out MPU setup in separate module") >>> Signed-off-by: Arnd Bergmann <arnd@arndb.de> >>> --- >>> Vladimir, if this looks good to you, can you forward it to Russell's >>> patch tracker, or otherwise suggest a different fix? >> >> Accepted as patch 8719/1. Not sure if Russell will pick it up or not... anyway, >> thanks for the patch! > > I'd prefer not to at this stage. Same things that apply that I said > in the "ARM: early_printk: use printascii() rather than printch()" > thread - I need to ensure that what I have is stable before 4.14 is > released, so I'm not accepting anything further unless it is _really_ > urgent. Yes, I saw that thread, it is why that "not sure" thing. > > This isn't urgent. > Should I resubmit after -rc1 or you'll apply it directly from patch tracker? Thanks Vladimir
On Thu, Nov 02, 2017 at 01:11:23PM +0000, Vladimir Murzin wrote: > On 02/11/17 13:07, Russell King - ARM Linux wrote: > > I'd prefer not to at this stage. Same things that apply that I said > > in the "ARM: early_printk: use printascii() rather than printch()" > > thread - I need to ensure that what I have is stable before 4.14 is > > released, so I'm not accepting anything further unless it is _really_ > > urgent. > > Yes, I saw that thread, it is why that "not sure" thing. > > > > > This isn't urgent. > > > > Should I resubmit after -rc1 or you'll apply it directly from patch tracker? If you put it in the patch tracker now, I'll apply it when appropriate. Thanks.
diff --git a/arch/arm/mm/pmsa-v7.c b/arch/arm/mm/pmsa-v7.c index 106ae1c435a3..976df60ac426 100644 --- a/arch/arm/mm/pmsa-v7.c +++ b/arch/arm/mm/pmsa-v7.c @@ -234,7 +234,7 @@ static int __init allocate_region(phys_addr_t base, phys_addr_t size, /* MPU initialisation functions */ void __init adjust_lowmem_bounds_mpu(void) { - phys_addr_t specified_mem_size, total_mem_size = 0; + phys_addr_t specified_mem_size = 0, total_mem_size = 0; struct memblock_region *reg; bool first = true; phys_addr_t mem_start;
The reworked MPU code produces a new warning in some configurations, presumably starting with the code move after the compiler now makes different inlining decisions: arch/arm/mm/pmsa-v7.c: In function 'adjust_lowmem_bounds_mpu': arch/arm/mm/pmsa-v7.c:310:5: error: 'specified_mem_size' may be used uninitialized in this function [-Werror=maybe-uninitialized] This appears to be harmless, as we know that there is always at least one memblock, and the only way this could get triggered is if the for_each_memblock() loop was never entered. I could not come up with a better workaround than initializing the specified_mem_size to zero, but at least that is the value that the variable would have in the hypothetical case of no memblocks. Fixes: 877ec119dbbf ("ARM: 8706/1: NOMMU: Move out MPU setup in separate module") Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- Vladimir, if this looks good to you, can you forward it to Russell's patch tracker, or otherwise suggest a different fix? --- arch/arm/mm/pmsa-v7.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)