Re: [PATCH v9] mm,kfence: decouple kfence from page granularity mapping judgement

[Zhenhua Huang <quic_zhenhuah@quicinc.com>]

On 2023/3/16 17:58, Pavan Kondeti wrote:
> On Thu, Mar 16, 2023 at 04:50:20PM +0800, Zhenhua Huang wrote:
>> Kfence only needs its pool to be mapped as page granularity, if it is
>> inited early. Previous judgement was a bit over protected. From [1], Mark
>> suggested to "just map the KFENCE region a page granularity". So I
>> decouple it from judgement and do page granularity mapping for kfence
>> pool only. Need to be noticed that late init of kfence pool still requires
>> page granularity mapping.
>>
>> Page granularity mapping in theory cost more(2M per 1GB) memory on arm64
>> platform. Like what I've tested on QEMU(emulated 1GB RAM) with
>> gki_defconfig, also turning off rodata protection:
>> Before:
>> [root@liebao ]# cat /proc/meminfo
>> MemTotal:         999484 kB
>> After:
>> [root@liebao ]# cat /proc/meminfo
>> MemTotal:        1001480 kB
>>
>> To implement this, also relocate the kfence pool allocation before the
>> linear mapping setting up, arm64_kfence_alloc_pool is to allocate phys
>> addr, __kfence_pool is to be set after linear mapping set up.
>>
>> LINK: [1] https://lore.kernel.org/linux-arm-kernel/Y+IsdrvDNILA59UN@FVFF77S0Q05N/
>> Suggested-by: Mark Rutland <mark.rutland@arm.com>
>> Signed-off-by: Zhenhua Huang <quic_zhenhuah@quicinc.com>
>> ---
>>   arch/arm64/include/asm/kfence.h | 16 +++++++++++
>>   arch/arm64/mm/mmu.c             | 59 +++++++++++++++++++++++++++++++++++++++++
>>   arch/arm64/mm/pageattr.c        |  9 +++++--
>>   include/linux/kfence.h          |  1 +
>>   mm/kfence/core.c                |  4 +++
>>   5 files changed, 87 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/arm64/include/asm/kfence.h b/arch/arm64/include/asm/kfence.h
>> index aa855c6..8143c91 100644
>> --- a/arch/arm64/include/asm/kfence.h
>> +++ b/arch/arm64/include/asm/kfence.h
>> @@ -10,6 +10,22 @@
>>   
>>   #include <asm/set_memory.h>
>>   
>> +extern phys_addr_t early_kfence_pool;
>> +
>> +#ifdef CONFIG_KFENCE
>> +
>> +extern char *__kfence_pool;
>> +static inline void kfence_set_pool(phys_addr_t addr)
>> +{
>> +	__kfence_pool = phys_to_virt(addr);
>> +}
>> +
>> +#else
>> +
>> +static inline void kfence_set_pool(phys_addr_t addr) { }
>> +
>> +#endif
>> +
>>   static inline bool arch_kfence_init_pool(void) { return true; }
>>   
>>   static inline bool kfence_protect_page(unsigned long addr, bool protect)
>> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
>> index 6f9d889..61944c70 100644
>> --- a/arch/arm64/mm/mmu.c
>> +++ b/arch/arm64/mm/mmu.c
>> @@ -24,6 +24,7 @@
>>   #include <linux/mm.h>
>>   #include <linux/vmalloc.h>
>>   #include <linux/set_memory.h>
>> +#include <linux/kfence.h>
>>   
>>   #include <asm/barrier.h>
>>   #include <asm/cputype.h>
>> @@ -38,6 +39,7 @@
>>   #include <asm/ptdump.h>
>>   #include <asm/tlbflush.h>
>>   #include <asm/pgalloc.h>
>> +#include <asm/kfence.h>
>>   
>>   #define NO_BLOCK_MAPPINGS	BIT(0)
>>   #define NO_CONT_MAPPINGS	BIT(1)
>> @@ -525,6 +527,48 @@ static int __init enable_crash_mem_map(char *arg)
>>   }
>>   early_param("crashkernel", enable_crash_mem_map);
>>   
>> +#ifdef CONFIG_KFENCE
>> +
>> +static bool kfence_early_init __initdata = !!CONFIG_KFENCE_SAMPLE_INTERVAL;
>> +/*
>> + * early_param can be parsed before linear mapping
>> + * set up
>> + */
>> +static int __init parse_kfence_early_init(char *p)
>> +{
>> +	int val;
>> +
>> +	if (get_option(&p, &val))
>> +		kfence_early_init = !!val;
>> +	return 0;
>> +}
>> +early_param("kfence.sample_interval", parse_kfence_early_init);
>> +
>> +static phys_addr_t arm64_kfence_alloc_pool(void)
>> +{
>> +	phys_addr_t kfence_pool;
>> +
>> +	if (!kfence_early_init)
>> +		return 0;
>> +
>> +	kfence_pool = memblock_phys_alloc(KFENCE_POOL_SIZE, PAGE_SIZE);
>> +	if (!kfence_pool)
>> +		pr_err("failed to allocate kfence pool\n");
>> +
>> +	return kfence_pool;
>> +}
>> +
>> +#else
>> +
>> +static phys_addr_t arm64_kfence_alloc_pool(void)
>> +{
>> +	return 0;
>> +}
>> +
>> +#endif
>> +
>> +phys_addr_t early_kfence_pool;
>> +
>>   static void __init map_mem(pgd_t *pgdp)
>>   {
>>   	static const u64 direct_map_end = _PAGE_END(VA_BITS_MIN);
>> @@ -543,6 +587,10 @@ static void __init map_mem(pgd_t *pgdp)
>>   	 */
>>   	BUILD_BUG_ON(pgd_index(direct_map_end - 1) == pgd_index(direct_map_end));
>>   
>> +	early_kfence_pool = arm64_kfence_alloc_pool();
>> +	if (early_kfence_pool)
>> +		memblock_mark_nomap(early_kfence_pool, KFENCE_POOL_SIZE);
>> +
>>   	if (can_set_direct_map())
>>   		flags |= NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS;
>>   
>> @@ -608,6 +656,17 @@ static void __init map_mem(pgd_t *pgdp)
>>   		}
>>   	}
>>   #endif
>> +
>> +	/* Kfence pool needs page-level mapping */
>> +	if (early_kfence_pool) {
>> +		__map_memblock(pgdp, early_kfence_pool,
>> +			early_kfence_pool + KFENCE_POOL_SIZE,
>> +			pgprot_tagged(PAGE_KERNEL),
>> +			NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS);
>> +		memblock_clear_nomap(early_kfence_pool, KFENCE_POOL_SIZE);
>> +		/* kfence_pool really mapped now */
>> +		kfence_set_pool(early_kfence_pool);
>> +	}
> 
> Why not wrap this under CONFIG_KFENCE ? early_kfence_pool can also go in
> there?

Because I didn't want to add CONFIG_KFENCE in function.. in the case of 
w/o CONFIG_KFENCE, early_kfence_pool should be always NULL.

Thanks,
Zhenhua

> 
> Thanks,
> Pavan