On Tue, May 25, 2021 at 6:51 AM Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Mon, May 24, 2021 at 03:58:29PM -0700, Andrii Nakryiko wrote:
> > > It took me a while to reliably bisect this, but it clearly points to
> > > this commit:
> > >
> > > e481fac7d80b ("mm/page_alloc: convert per-cpu list protection to local_lock")
> > >
> > > One commit before it, 676535512684 ("mm/page_alloc: split per cpu page
> > > lists and zone stats -fix"), works just fine.
> > >
> > > I'll have to spend more time debugging what exactly is happening, but
> > > the immediate problem is two different definitions of numa_node
> > > per-cpu variable. They both are at the same offset within
> > > .data..percpu ELF section, they both have the same name, but one of
> > > them is marked as static and another as global. And one is int
> > > variable, while another is struct pagesets. I'll look some more
> > > tomorrow, but adding Jiri and Arnaldo for visibility.
> > >
> > > [110907] DATASEC '.data..percpu' size=178904 vlen=303
> > > ...
> > > type_id=27753 offset=163976 size=4 (VAR 'numa_node')
> > > type_id=27754 offset=163976 size=4 (VAR 'numa_node')
> > >
> > > [27753] VAR 'numa_node' type_id=27556, linkage=static
> > > [27754] VAR 'numa_node' type_id=20, linkage=global
> > >
> > > [20] INT 'int' size=4 bits_offset=0 nr_bits=32 encoding=SIGNED
> > >
> > > [27556] STRUCT 'pagesets' size=0 vlen=1
> > > 'lock' type_id=507 bits_offset=0
> > >
> > > [506] STRUCT '(anon)' size=0 vlen=0
> > > [507] TYPEDEF 'local_lock_t' type_id=506
> > >
> > > So also something weird about those zero-sized struct pagesets and
> > > local_lock_t inside it.
> >
> > Ok, so nothing weird about them. local_lock_t is designed to be
> > zero-sized unless CONFIG_DEBUG_LOCK_ALLOC is defined.
> >
> > But such zero-sized per-CPU variables are confusing pahole during BTF
> > generation, as now two different variables "occupy" the same address.
> >
> > Given this seems to be the first zero-sized per-CPU variable, I wonder
> > if it would be ok to make sure it's never zero-sized, while pahole
> > gets fixed and it's latest version gets widely packaged and
> > distributed.
> >
> > Mel, what do you think about something like below? Or maybe you can
> > advise some better solution?
> >
>
> Ouch, something like that may never go away. How about just this?
Yeah, that would work just fine, thanks! Would you like me to send a
formal patch or you'd like to do it?
>
> diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
> index 58426acf5983..dce2df33d823 100644
> --- a/lib/Kconfig.debug
> +++ b/lib/Kconfig.debug
> @@ -338,6 +338,9 @@ config DEBUG_INFO_BTF
> config PAHOLE_HAS_SPLIT_BTF
> def_bool $(success, test `$(PAHOLE) --version | sed -E 's/v([0-9]+)\.([0-9]+)/\1\2/'` -ge "119")
>
> +config PAHOLE_HAS_ZEROSIZE_PERCPU_SUPPORT
> + def_bool $(success, test `$(PAHOLE) --version | sed -E 's/v([0-9]+)\.([0-9]+)/\1\2/'` -ge "122")
> +
> config DEBUG_INFO_BTF_MODULES
> def_bool y
> depends on DEBUG_INFO_BTF && MODULES && PAHOLE_HAS_SPLIT_BTF
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 1599985e0ee1..cb1f84848c99 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -124,6 +124,17 @@ static DEFINE_MUTEX(pcp_batch_high_lock);
>
> struct pagesets {
> local_lock_t lock;
> +#if defined(CONFIG_DEBUG_INFO_BTF) && \
> + !defined(CONFIG_DEBUG_LOCK_ALLOC) && \
> + !defined(CONFIG_PAHOLE_HAS_ZEROSIZE_PERCPU_SUPPORT)
> + /*
> + * pahole 1.21 and earlier gets confused by zero-sized per-CPU
> + * variables and produces invalid BTF. Ensure that
> + * sizeof(struct pagesets) != 0 for older versions of pahole.
> + */
> + char __pahole_hack;
> + #warning "pahole too old to support zero-sized struct pagesets"
> +#endif
> };
> static DEFINE_PER_CPU(struct pagesets, pagesets) = {
> .lock = INIT_LOCAL_LOCK(lock),
> diff --git a/scripts/rust-version.sh b/scripts/rust-version.sh
> old mode 100644
> new mode 100755
Probably didn't intend to include this?
> --
> Mel Gorman
> SUSE Labs
