Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

Hi Andrew, >filter_irq_stacks() is too large to be inlined.   Ok we are thinking to move definations in kernel/stacktrace.c as a normal global function.   >in_irqentry_text() is probably too large to be inlined as well, and >should return bool. We can declare it as static funciton in

Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

Hi Andrew, >filter_irq_stacks() is too large to be inlined.   Ok we are thinking to move definations in kernel/stacktrace.c as a normal global function.   >in_irqentry_text() is probably too large to be inlined as well, and >should return bool. We can declare it as static funciton in

Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

On Mon, 26 Mar 2018 19:47:17 +0530 Vaneet Narang wrote: > Hi Dmitry, > > >Every user of stack_depot should filter out irq frames, without that > >stack_depot will run out of memory sooner or later. so this is a > >change in the right direction. > > > >Do we need to define

Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

On Mon, 26 Mar 2018 19:47:17 +0530 Vaneet Narang wrote: > Hi Dmitry, > > >Every user of stack_depot should filter out irq frames, without that > >stack_depot will run out of memory sooner or later. so this is a > >change in the right direction. > > > >Do we need to define empty version of

RE: Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

Hi Dmitry, >Every user of stack_depot should filter out irq frames, without that >stack_depot will run out of memory sooner or later. so this is a >change in the right direction. > >Do we need to define empty version of in_irqentry_text? Shouldn't only >filter_irq_stacks be used by kernel code?

RE: Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

Hi Dmitry, >Every user of stack_depot should filter out irq frames, without that >stack_depot will run out of memory sooner or later. so this is a >change in the right direction. > >Do we need to define empty version of in_irqentry_text? Shouldn't only >filter_irq_stacks be used by kernel code?

Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

On Mon, Mar 26, 2018 at 11:58 AM, Maninder Singh wrote: > Check whether the allocation happens in an IRQ handler. > This lets us strip everything below the IRQ entry point to reduce the > number of unique stack traces needed to be stored. > > so moved code of KASAN in

Re: [PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

On Mon, Mar 26, 2018 at 11:58 AM, Maninder Singh wrote: > Check whether the allocation happens in an IRQ handler. > This lets us strip everything below the IRQ entry point to reduce the > number of unique stack traces needed to be stored. > > so moved code of KASAN in generic file so that

[PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

Check whether the allocation happens in an IRQ handler. This lets us strip everything below the IRQ entry point to reduce the number of unique stack traces needed to be stored. so moved code of KASAN in generic file so that page_owner can also do same filteration. Initial KASAN commit

[PATCH v2] mm/page_owner: ignore everything below the IRQ entry point

Check whether the allocation happens in an IRQ handler. This lets us strip everything below the IRQ entry point to reduce the number of unique stack traces needed to be stored. so moved code of KASAN in generic file so that page_owner can also do same filteration. Initial KASAN commit