[igt-dev] [PATCH i-g-t 1/1] lib/intel_allocator: Move the ioctl calls to client processes

Zbigniew Kempczyński zbigniew.kempczynski at intel.com
Tue May 25 06:19:54 UTC 2021


On Mon, May 24, 2021 at 06:34:29PM +0200, Andrzej Turko wrote:
> When running the allocator in multiprocess mode, all queries
> are processed in a designated thread in the parent process.
> However, a child process may request opening the allocator
> for a gpu using a file descriptor absent in the parent process.
> Hence, querying available gtt size must be done in the child
> instead of the parent process.
> 
> This modification has triggered slight changes to the
> creation of random and reloc allocators.
> 
> Signed-off-by: Andrzej Turko <andrzej.turko at linux.intel.com>
> Cc: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>
> ---
>  lib/intel_allocator.c        | 51 ++++++++++++++++++++++++++----------
>  lib/intel_allocator_random.c | 28 ++++++++------------
>  lib/intel_allocator_reloc.c  | 20 ++++----------
>  lib/intel_allocator_simple.c | 44 ++++---------------------------
>  4 files changed, 58 insertions(+), 85 deletions(-)

Add test which reveals the problem. I suggest to add it in 
first patch because I just want to see how allocator explodes.

> 
> diff --git a/lib/intel_allocator.c b/lib/intel_allocator.c
> index 96f839d4b..5313af174 100644
> --- a/lib/intel_allocator.c
> +++ b/lib/intel_allocator.c
> @@ -45,6 +45,12 @@ static inline const char *reqstr(enum reqtype request_type)
>  #define alloc_debug(...) {}
>  #endif
>  
> +/*
> + * We limit allocator space to avoid hang when batch would be
> + * pinned in the last page.
> + */
> +#define RESERVED 4096
> +
>  struct allocator {
>  	int fd;
>  	uint32_t ctx;
> @@ -58,12 +64,11 @@ struct handle_entry {
>  	struct allocator *al;
>  };
>  
> -struct intel_allocator *intel_allocator_reloc_create(int fd);
> -struct intel_allocator *intel_allocator_random_create(int fd);
> -struct intel_allocator *intel_allocator_simple_create(int fd);
> +struct intel_allocator *intel_allocator_reloc_create(int fd, uint64_t end);
> +struct intel_allocator *intel_allocator_random_create(int fd, uint64_t end);

Add start too, just if we want to start on different offset. 

>  struct intel_allocator *
> -intel_allocator_simple_create_full(int fd, uint64_t start, uint64_t end,
> -				   enum allocator_strategy strategy);
> +intel_allocator_simple_create(int fd, uint64_t start, uint64_t end,
> +			      enum allocator_strategy strategy);
>  
>  /*
>   * Instead of trying to find first empty handle just get new one. Assuming
> @@ -286,17 +291,14 @@ static struct intel_allocator *intel_allocator_create(int fd,
>  			     "We cannot use NONE allocator\n");
>  		break;
>  	case INTEL_ALLOCATOR_RELOC:
> -		ial = intel_allocator_reloc_create(fd);
> +		ial = intel_allocator_reloc_create(fd, end);
>  		break;
>  	case INTEL_ALLOCATOR_RANDOM:
> -		ial = intel_allocator_random_create(fd);
> +		ial = intel_allocator_random_create(fd, end);
>  		break;
>  	case INTEL_ALLOCATOR_SIMPLE:
> -		if (!start && !end)
> -			ial = intel_allocator_simple_create(fd);
> -		else
> -			ial = intel_allocator_simple_create_full(fd, start, end,
> -								 allocator_strategy);
> +		ial = intel_allocator_simple_create(fd, start, end,
> +						    allocator_strategy);
>  		break;
>  	default:
>  		igt_assert_f(ial, "Allocator type %d not implemented\n",
> @@ -877,6 +879,13 @@ static uint64_t __intel_allocator_open_full(int fd, uint32_t ctx,
>  				 .open.allocator_type = allocator_type,
>  				 .open.allocator_strategy = strategy };
>  	struct alloc_resp resp;
> +	uint64_t gtt_size = gem_aperture_size(fd);
> +
> +	igt_assert(end <= gtt_size);
> +	if (!gem_uses_full_ppgtt(fd))
> +		gtt_size /= 2;
> +	igt_assert(end - start <= gtt_size);
> +

Unnecessary extra space.

>  
>  	/* Get child_tid only once at open() */
>  	if (child_tid == -1)
> @@ -954,13 +963,27 @@ uint64_t intel_allocator_open_vm_full(int fd, uint32_t vm,
>   */
>  uint64_t intel_allocator_open(int fd, uint32_t ctx, uint8_t allocator_type)
>  {
> -	return intel_allocator_open_full(fd, ctx, 0, 0, allocator_type,
> +	uint64_t gtt_size = gem_aperture_size(fd);
> +
> +	if (!gem_uses_full_ppgtt(fd))
> +		gtt_size /= 2;
> +	else
> +		gtt_size -= RESERVED;
> +
> +	return intel_allocator_open_full(fd, ctx, 0, gtt_size, allocator_type,
>  					 ALLOC_STRATEGY_HIGH_TO_LOW);
>  }
>  
>  uint64_t intel_allocator_open_vm(int fd, uint32_t vm, uint8_t allocator_type)
>  {
> -	return intel_allocator_open_vm_full(fd, vm, 0, 0, allocator_type,
> +	uint64_t gtt_size = gem_aperture_size(fd);
> +
> +	if (!gem_uses_full_ppgtt(fd))
> +		gtt_size /= 2;
> +	else
> +		gtt_size -= RESERVED;
> +
> +	return intel_allocator_open_vm_full(fd, vm, 0, gtt_size, allocator_type,
>  					    ALLOC_STRATEGY_HIGH_TO_LOW);
>  }
>  
> diff --git a/lib/intel_allocator_random.c b/lib/intel_allocator_random.c
> index 3d9a78f17..52b167a12 100644
> --- a/lib/intel_allocator_random.c
> +++ b/lib/intel_allocator_random.c
> @@ -10,12 +10,12 @@
>  #include "igt_rand.h"
>  #include "intel_allocator.h"
>  
> -struct intel_allocator *intel_allocator_random_create(int fd);
> +struct intel_allocator *intel_allocator_random_create(int fd, uint64_t end);
>  
>  struct intel_allocator_random {
>  	uint64_t bias;
>  	uint32_t prng;
> -	uint64_t gtt_size;
> +	uint64_t address_mask;
>  	uint64_t start;
>  	uint64_t end;
>  
> @@ -23,12 +23,8 @@ struct intel_allocator_random {
>  	uint64_t allocated_objects;
>  };
>  
> -static uint64_t get_bias(int fd)
> -{
> -	(void) fd;
> +#define BIAS 256 << 10;
>  
> -	return 256 << 10;
> -}
>  
>  static void intel_allocator_random_get_address_range(struct intel_allocator *ial,
>  						     uint64_t *startp,
> @@ -57,8 +53,8 @@ static uint64_t intel_allocator_random_alloc(struct intel_allocator *ial,
>  	/* randomize the address, we try to avoid relocations */
>  	do {
>  		offset = hars_petruska_f54_1_random64(&ialr->prng);
> -		offset += ialr->bias; /* Keep the low 256k clear, for negative deltas */
> -		offset &= ialr->gtt_size - 1;
> +		offset |= ialr->bias; /* Keep the low 256k clear, for negative deltas */
> +		offset &= ialr->address_mask;
>  		offset &= ~(alignment - 1);
>  	} while (offset + size > ialr->end);
>  
> @@ -150,8 +146,7 @@ static bool intel_allocator_random_is_empty(struct intel_allocator *ial)
>  	return !ialr->allocated_objects;
>  }
>  
> -#define RESERVED 4096
> -struct intel_allocator *intel_allocator_random_create(int fd)
> +struct intel_allocator *intel_allocator_random_create(int fd, uint64_t end)
>  {
>  	struct intel_allocator *ial;
>  	struct intel_allocator_random *ialr;
> @@ -175,14 +170,13 @@ struct intel_allocator *intel_allocator_random_create(int fd)
>  	ialr = ial->priv = calloc(1, sizeof(*ialr));
>  	igt_assert(ial->priv);
>  	ialr->prng = (uint32_t) to_user_pointer(ial);
> -	ialr->gtt_size = gem_aperture_size(fd);
> -	igt_debug("Gtt size: %" PRId64 "\n", ialr->gtt_size);
> -	if (!gem_uses_full_ppgtt(fd))
> -		ialr->gtt_size /= 2;
> +	ialr->address_mask = 1;
> +	while (ialr->address_mask <= end) ialr->address_mask *= 2ULL;
> +	ialr->address_mask--;

If end is not power of two you can have sparse address mask. 
Calculate this mask using roundup_power_of_two(). Masking 
randomized offset is not enough but while() condition will
do the rest. 

--
Zbigniew

>  
> -	ialr->bias = get_bias(fd);
> +	ialr->bias = BIAS;
>  	ialr->start = ialr->bias;
> -	ialr->end = ialr->gtt_size - RESERVED;
> +	ialr->end = end;
>  
>  	ialr->allocated_objects = 0;
>  
> diff --git a/lib/intel_allocator_reloc.c b/lib/intel_allocator_reloc.c
> index e8af787b0..e856132a7 100644
> --- a/lib/intel_allocator_reloc.c
> +++ b/lib/intel_allocator_reloc.c
> @@ -10,12 +10,11 @@
>  #include "igt_rand.h"
>  #include "intel_allocator.h"
>  
> -struct intel_allocator *intel_allocator_reloc_create(int fd);
> +struct intel_allocator *intel_allocator_reloc_create(int fd, uint64_t end);
>  
>  struct intel_allocator_reloc {
>  	uint64_t bias;
>  	uint32_t prng;
> -	uint64_t gtt_size;
>  	uint64_t start;
>  	uint64_t end;
>  	uint64_t offset;
> @@ -24,12 +23,8 @@ struct intel_allocator_reloc {
>  	uint64_t allocated_objects;
>  };
>  
> -static uint64_t get_bias(int fd)
> -{
> -	(void) fd;
> +#define BIAS 256 << 10;
>  
> -	return 256 << 10;
> -}
>  
>  static void intel_allocator_reloc_get_address_range(struct intel_allocator *ial,
>  						    uint64_t *startp,
> @@ -152,8 +147,7 @@ static bool intel_allocator_reloc_is_empty(struct intel_allocator *ial)
>  	return !ialr->allocated_objects;
>  }
>  
> -#define RESERVED 4096
> -struct intel_allocator *intel_allocator_reloc_create(int fd)
> +struct intel_allocator *intel_allocator_reloc_create(int fd, uint64_t end)
>  {
>  	struct intel_allocator *ial;
>  	struct intel_allocator_reloc *ialr;
> @@ -177,14 +171,10 @@ struct intel_allocator *intel_allocator_reloc_create(int fd)
>  	ialr = ial->priv = calloc(1, sizeof(*ialr));
>  	igt_assert(ial->priv);
>  	ialr->prng = (uint32_t) to_user_pointer(ial);
> -	ialr->gtt_size = gem_aperture_size(fd);
> -	igt_debug("Gtt size: %" PRId64 "\n", ialr->gtt_size);
> -	if (!gem_uses_full_ppgtt(fd))
> -		ialr->gtt_size /= 2;
>  
> -	ialr->bias = ialr->offset = get_bias(fd);
> +	ialr->bias = ialr->offset = BIAS;
>  	ialr->start = ialr->bias;
> -	ialr->end = ialr->gtt_size - RESERVED;
> +	ialr->end = end;
>  
>  	ialr->allocated_objects = 0;
>  
> diff --git a/lib/intel_allocator_simple.c b/lib/intel_allocator_simple.c
> index 963d8d257..6022e832b 100644
> --- a/lib/intel_allocator_simple.c
> +++ b/lib/intel_allocator_simple.c
> @@ -11,17 +11,11 @@
>  #include "intel_bufops.h"
>  #include "igt_map.h"
>  
> -/*
> - * We limit allocator space to avoid hang when batch would be
> - * pinned in the last page.
> - */
> -#define RESERVED 4096
>  
>  /* Avoid compilation warning */
> -struct intel_allocator *intel_allocator_simple_create(int fd);
>  struct intel_allocator *
> -intel_allocator_simple_create_full(int fd, uint64_t start, uint64_t end,
> -				   enum allocator_strategy strategy);
> +intel_allocator_simple_create(int fd, uint64_t start, uint64_t end,
> +			      enum allocator_strategy strategy);
>  
>  struct simple_vma_heap {
>  	struct igt_list_head holes;
> @@ -734,9 +728,9 @@ static void intel_allocator_simple_print(struct intel_allocator *ial, bool full)
>  		 ials->allocated_objects, ials->reserved_areas);
>  }
>  
> -static struct intel_allocator *
> -__intel_allocator_simple_create(int fd, uint64_t start, uint64_t end,
> -				enum allocator_strategy strategy)
> +struct intel_allocator *
> +intel_allocator_simple_create(int fd, uint64_t start, uint64_t end,
> +			      enum allocator_strategy strategy)
>  {
>  	struct intel_allocator *ial;
>  	struct intel_allocator_simple *ials;
> @@ -777,31 +771,3 @@ __intel_allocator_simple_create(int fd, uint64_t start, uint64_t end,
>  
>  	return ial;
>  }
> -
> -struct intel_allocator *
> -intel_allocator_simple_create(int fd)
> -{
> -	uint64_t gtt_size = gem_aperture_size(fd);
> -
> -	if (!gem_uses_full_ppgtt(fd))
> -		gtt_size /= 2;
> -	else
> -		gtt_size -= RESERVED;
> -
> -	return __intel_allocator_simple_create(fd, 0, gtt_size,
> -					       ALLOC_STRATEGY_HIGH_TO_LOW);
> -}
> -
> -struct intel_allocator *
> -intel_allocator_simple_create_full(int fd, uint64_t start, uint64_t end,
> -				   enum allocator_strategy strategy)
> -{
> -	uint64_t gtt_size = gem_aperture_size(fd);
> -
> -	igt_assert(end <= gtt_size);
> -	if (!gem_uses_full_ppgtt(fd))
> -		gtt_size /= 2;
> -	igt_assert(end - start <= gtt_size);
> -
> -	return __intel_allocator_simple_create(fd, start, end, strategy);
> -}
> -- 
> 2.25.1
> 


More information about the igt-dev mailing list