Re: [PATCH i-g-t v3 1/1] test/intel/xe_vm:Add oversubscribe concurrent bind stress subtest

["Hellstrom, Thomas" <thomas.hellstrom@intel.com>]

On Tue, 2026-04-14 at 19:27 +0200, Kamil Konieczny wrote:
> Hi Sobin,
> On 2026-04-13 at 04:33:18 +0000, Sobin Thomas wrote:
> 
> small nit about subject, add space after ':' so it will be:
> 
> [PATCH i-g-t v3 1/1] test/intel/xe_vm: Add oversubscribe concurrent
> bind stress subtest
> 
> Few more nits below.
> 
> > Add test for oversubscribing VRAM in multi process environment that
> > creates VM, bind large BOs and submit workloads nearly
> > simultaneously.
> > 
> > Previous coverage lacked a scenario combining multi-process bind
> > with VRAM oversubscription. This generates memory pressure with
> > multi-process VM Bind activity and concurrent submission,
> > exercising
> > the bind pipeline under eviction pressure.
> > 
> > v2: Removed helper APIs usage clock_nanosleep and commented
> > code.(Nishit)
> > 
> > v3: Refactored code to smaller functions.
> >     Added check for available SRAM usage and keep the max process
> > to 20.
> > 
> > v4: Remove explicit macros definition
> >     Replace Bind ioctl with library calls.(Thomas)
> > 
> > Signed-off-by: Sobin Thomas <sobin.thomas@intel.com>
> > ---
> >  tests/intel/xe_vm.c | 433
> > ++++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 433 insertions(+)
> > 
> > diff --git a/tests/intel/xe_vm.c b/tests/intel/xe_vm.c
> > index d75b0730d..42effbd48 100644
> > --- a/tests/intel/xe_vm.c
> > +++ b/tests/intel/xe_vm.c
> > @@ -19,8 +19,100 @@
> >  #include "xe/xe_ioctl.h"
> >  #include "xe/xe_query.h"
> >  #include "xe/xe_spin.h"
> > +#include <inttypes.h>
> >  #include <string.h>
> 
> Could you move both to proper place?
> 
> >  
> > +#define GB(x) (1024ULL * 1024ULL * 1024ULL * (x))
> > +#define MAX_THREADS 20
> > +struct gem_bo {
> > +	uint32_t handle;
> > +	uint64_t size;
> > +	int *ptr;
> > +	uint64_t addr;
> > +};
> > +
> > +struct xe_test_ctx {
> > +	int fd;
> > +	uint32_t vm_id;
> > +	uint32_t exec_queue_id;
> > +	uint16_t sram_instance;
> > +	uint16_t vram_instance;
> > +	bool has_vram;
> > +};
> > +
> > +struct mem_bind_sync {
> > +	struct gem_bo *bufs;
> > +	int n_bufs;
> > +	uint64_t *binds_ufence;
> > +};
> > +
> > +static uint64_t align_to_page_size(uint64_t size)
> 
> Here imho add this as param or rename to explicit 4K:
> 
> static uint64_t align_to_4K_page_size(uint64_t size)
> 
> or
> 
> static uint64_t align_to_page_size(uint64_t size, uint64_t page_size)
> 
> > +{
> > +	return (size + 4095UL) & ~4095UL;
> > +}
> > +

Perhaps use the ALIGN() macro?
/Thomas


> 
> Regards,
> Kamil 
> 
> [cut]