[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <aOPBePFxNk5FP/ru@MiWiFi-R3L-srv>
Date: Mon, 6 Oct 2025 21:17:44 +0800
From: Baoquan He <bhe@...hat.com>
To: Uladzislau Rezki <urezki@...il.com>
Cc: linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>,
Michal Hocko <mhocko@...nel.org>,
LKML <linux-kernel@...r.kernel.org>, Michal Hocko <mhocko@...e.com>
Subject: Re: [PATCH v3 09/10] mm/vmalloc: Update __vmalloc_node_range()
documentation
On 10/06/25 at 12:06pm, Uladzislau Rezki wrote:
> On Sat, Oct 04, 2025 at 01:02:02PM +0800, Baoquan He wrote:
> > On 10/04/25 at 12:11pm, Baoquan He wrote:
> > > On 10/01/25 at 09:26pm, Uladzislau Rezki (Sony) wrote:
> > > > __vmalloc() function now supports non-blocking flags such as
> > > > GFP_ATOMIC and GFP_NOWAIT. Update the documentation accordingly.
> > > >
> > > > Acked-by: Michal Hocko <mhocko@...e.com>
> > > > Signed-off-by: Uladzislau Rezki (Sony) <urezki@...il.com>
> > > > ---
> > > > mm/vmalloc.c | 21 +++++++++++----------
> > > > 1 file changed, 11 insertions(+), 10 deletions(-)
> > > >
> > > > diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> > > > index d7e7049e01f8..2b45cd4ce119 100644
> > > > --- a/mm/vmalloc.c
> > > > +++ b/mm/vmalloc.c
> > > > @@ -3881,19 +3881,20 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
> > > > * @caller: caller's return address
> > > > *
> > > > * Allocate enough pages to cover @size from the page level
> > > > - * allocator with @gfp_mask flags. Please note that the full set of gfp
> > > > - * flags are not supported. GFP_KERNEL, GFP_NOFS and GFP_NOIO are all
> > > > - * supported.
> > > > - * Zone modifiers are not supported. From the reclaim modifiers
> > > > - * __GFP_DIRECT_RECLAIM is required (aka GFP_NOWAIT is not supported)
> > > > - * and only __GFP_NOFAIL is supported (i.e. __GFP_NORETRY and
> > > > - * __GFP_RETRY_MAYFAIL are not supported).
> > > > + * allocator with @gfp_mask flags and map them into contiguous
> > > > + * virtual range with protection @prot.
> > > > *
> > > > - * __GFP_NOWARN can be used to suppress failures messages.
> > > > + * Supported GFP classes: %GFP_KERNEL, %GFP_ATOMIC, %GFP_NOWAIT,
> > > > + * %GFP_NOFS and %GFP_NOIO. Zone modifiers are not supported.
> > > > + * Please note %GFP_ATOMIC and %GFP_NOWAIT are supported only
> > > > + * by __vmalloc().
> > > > +
> > > > + * Retry modifiers: only %__GFP_NOFAIL is supported; %__GFP_NORETRY
> > > > + * and %__GFP_RETRY_MAYFAIL are not supported.
> > >
> > > Do we need to update the documentation of __vmalloc_node_noprof()
> > > accordingly? I see it has below description about "Retry modifiers"
> > > where gfp_mask is passed down to __vmalloc_node_range_noprof() directly
> > > but have different description. Not sure if I missed anything.
> > >
> > > ===
> > > * Retry modifiers: only %__GFP_NOFAIL is supported; %__GFP_NORETRY
> > > * and %__GFP_RETRY_MAYFAIL are not supported.
> > > ===
> >
> > Sorry, I copied the wrong sentences. Below is copied from documentation
> > of __vmalloc_node_noprof().
> > ====
> > * Reclaim modifiers in @gfp_mask - __GFP_NORETRY, __GFP_RETRY_MAYFAIL
> > * and __GFP_NOFAIL are not supported
> > ====
> >
> >
> > >
> > > > *
> > > > - * Map them into contiguous kernel virtual space, using a pagetable
> > > > - * protection of @prot.
> > > > + * %__GFP_NOWARN can be used to suppress failure messages.
> > > > *
> > > > + * Can not be called from interrupt nor NMI contexts.
> > > > * Return: the address of the area or %NULL on failure
> > > > */
> > > > void *__vmalloc_node_range_noprof(unsigned long size, unsigned long align,
> > > > --
> > > > 2.47.3
> > > >
> > >
> >
> We need. But i am not sure it should be fully copy-pasted from the
> __vmalloc_node_range_noprof(). At least __GFP_NOFAIL is supported
> and thus stating that it is not - is wrong.
>
> It has to be fixed but not by this series because when __GFP_NOFAIL
> support was introduced the doc. should have to be updated accordingly.
Maybe just remove the documentation for __vmalloc_node_noprof() since
it's only a wrapper of __vmalloc_node_range_noprof()? Surely this should
be done in another standalone patch later.
Powered by blists - more mailing lists