Re: [PATCH RFC] mm: Add debug_virt_to_phys()

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 11/14/2016 10:45 AM, Laura Abbott wrote:
> On 11/11/2016 04:44 PM, Florian Fainelli wrote:
>> When CONFIG_DEBUG_VM is turned on, virt_to_phys() maps to
>> debug_virt_to_phys() which helps catch vmalloc space addresses being
>> passed. This is helpful in debugging bogus drivers that just assume
>> linear mappings all over the place.
>>
>> For ARM, ARM64, Unicore32 and Microblaze, the architectures define
>> __virt_to_phys() as being the functional implementation of the address
>> translation, so we special case the debug stub to call into
>> __virt_to_phys directly.
>>
>> Signed-off-by: Florian Fainelli <f.fainelli@xxxxxxxxx>
>> ---
>>  arch/arm/include/asm/memory.h      |  4 ++++
>>  arch/arm64/include/asm/memory.h    |  4 ++++
>>  include/asm-generic/memory_model.h |  4 ++++
>>  mm/debug.c                         | 15 +++++++++++++++
>>  4 files changed, 27 insertions(+)
>>
>> diff --git a/arch/arm/include/asm/memory.h b/arch/arm/include/asm/memory.h
>> index 76cbd9c674df..448dec9b8b00 100644
>> --- a/arch/arm/include/asm/memory.h
>> +++ b/arch/arm/include/asm/memory.h
>> @@ -260,11 +260,15 @@ static inline unsigned long __phys_to_virt(phys_addr_t x)
>>   * translation for translating DMA addresses.  Use the driver
>>   * DMA support - see dma-mapping.h.
>>   */
>> +#ifndef CONFIG_DEBUG_VM
>>  #define virt_to_phys virt_to_phys
>>  static inline phys_addr_t virt_to_phys(const volatile void *x)
>>  {
>>  	return __virt_to_phys((unsigned long)(x));
>>  }
>> +#else
>> +#define virt_to_phys debug_virt_to_phys
>> +#endif
>>  
>>  #define phys_to_virt phys_to_virt
>>  static inline void *phys_to_virt(phys_addr_t x)
>> diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h
>> index b71086d25195..c9e436b28523 100644
>> --- a/arch/arm64/include/asm/memory.h
>> +++ b/arch/arm64/include/asm/memory.h
>> @@ -186,11 +186,15 @@ extern u64			kimage_voffset;
>>   * translation for translating DMA addresses.  Use the driver
>>   * DMA support - see dma-mapping.h.
>>   */
>> +#ifndef CONFIG_DEBUG_VM
>>  #define virt_to_phys virt_to_phys
>>  static inline phys_addr_t virt_to_phys(const volatile void *x)
>>  {
>>  	return __virt_to_phys((unsigned long)(x));
>>  }
>> +#else
>> +#define virt_to_phys debug_virt_to_phys
>> +#endif
>>  
>>  #define phys_to_virt phys_to_virt
>>  static inline void *phys_to_virt(phys_addr_t x)
>> diff --git a/include/asm-generic/memory_model.h b/include/asm-generic/memory_model.h
>> index 5148150cc80b..426085757258 100644
>> --- a/include/asm-generic/memory_model.h
>> +++ b/include/asm-generic/memory_model.h
>> @@ -80,6 +80,10 @@
>>  #define page_to_pfn __page_to_pfn
>>  #define pfn_to_page __pfn_to_page
>>  
>> +#ifdef CONFIG_DEBUG_VM
>> +unsigned long debug_virt_to_phys(volatile void *address);
>> +#endif /* CONFIG_DEBUG_VM */
>> +
>>  #endif /* __ASSEMBLY__ */
>>  
>>  #endif
>> diff --git a/mm/debug.c b/mm/debug.c
>> index 9feb699c5d25..72b2ca9b11f4 100644
>> --- a/mm/debug.c
>> +++ b/mm/debug.c
>> @@ -161,4 +161,19 @@ void dump_mm(const struct mm_struct *mm)
>>  	);
>>  }
>>  
>> +#include <asm/memory.h>
>> +#include <linux/mm.h>
>> +
>> +unsigned long debug_virt_to_phys(volatile void *address)
>> +{
>> +	BUG_ON(is_vmalloc_addr((const void *)address));
>> +#if defined(CONFIG_ARM) || defined(CONFIG_ARM64) || defined(CONFIG_UNICORE32) || \
>> +	defined(CONFIG_MICROBLAZE)
>> +	return __virt_to_phys(address);
>> +#else
>> +	return virt_to_phys(address);
>> +#endif
>> +}
>> +EXPORT_SYMBOL(debug_virt_to_phys);
>> +
>>  #endif		/* CONFIG_DEBUG_VM */
>>
> 
> is_vmalloc_addr is necessary but not sufficient. This misses
> cases like module addresses.

Indeed, thanks.

> The x86 version (CONFIG_DEBUG_VIRTUAL)
> bounds checks against the known linear map to catch all cases.
> I'm for a generic approach to this if it can catch all cases
> that an architecture specific version would catch.

For one, my patch causes an early BUG to occur on ARM64 during
arch/arm64/kernel/setup.c::setup_arch when we call
cpu_uninstall_idmap(). I suspect there could be a bunch of little checks
like these where we'd have to have an architecture specific "is this
physical/virtual address valid" that may make a generic implementation
hard to come up with.
-- 
Florian

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>



[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]