From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Vetter Date: Mon, 25 Apr 2016 09:11:28 +0000 Subject: Re: [PATCH v2 5/8] fbdev: fb_defio: Export fb_deferred_io_mmap Message-Id: <20160425091128.GR2510@phenom.ffwll.local> List-Id: References: <1461530942-22485-1-git-send-email-noralf@tronnes.org> <1461530942-22485-6-git-send-email-noralf@tronnes.org> In-Reply-To: <1461530942-22485-6-git-send-email-noralf@tronnes.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To: Noralf =?iso-8859-1?Q?Tr=F8nnes?= Cc: linux-fbdev@vger.kernel.org, linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, tomi.valkeinen@ti.com, laurent.pinchart@ideasonboard.com On Sun, Apr 24, 2016 at 10:48:59PM +0200, Noralf Tr=F8nnes wrote: > Export fb_deferred_io_mmap so drivers can change vma->vm_page_prot. > When the framebuffer memory is allocated using dma_alloc_writecombine() > instead of vmalloc(), I get cache syncing problems on ARM. > This solves it: >=20 > static int drm_fbdev_cma_deferred_io_mmap(struct fb_info *info, > struct vm_area_struct *vma) > { > fb_deferred_io_mmap(info, vma); > vma->vm_page_prot =3D pgprot_writecombine(vma->vm_page_prot); >=20 > return 0; > } >=20 > Could this have been done in the core? > Drivers that don't set (struct fb_ops *)->fb_mmap, gets a call to > fb_pgprotect() at the end of the default fb_mmap implementation > (drivers/video/fbdev/core/fbmem.c). This is an architecture specific > function that on many platforms uses pgprot_writecombine(), but not on > all. And looking at some of the fb_mmap implementations, some of them > sets vm_page_prot to nocache for instance, so I think the safest bet is > to do this in the driver and not in the fbdev core. And we can't call > fb_pgprotect() from fb_deferred_io_mmap() either because we don't have > access to the file pointer that powerpc needs. >=20 > Signed-off-by: Noralf Tr=F8nnes lgtm, but needs an ack from Tomi for merging through drm trees. Tomi? Aside: When resending patches please include all the r-b/acks and Cc: people who participated in the discussion explicitly. That includes conditional r-b tags if you do the changes (like some of the r-b I'm doing right now). Makes life _much_ easier for maintainers ;-) -Daniel > --- >=20 > Changes since v1: > - Expand commit message >=20 > drivers/video/fbdev/core/fb_defio.c | 3 ++- > include/linux/fb.h | 1 + > 2 files changed, 3 insertions(+), 1 deletion(-) >=20 > diff --git a/drivers/video/fbdev/core/fb_defio.c b/drivers/video/fbdev/co= re/fb_defio.c > index 57721c7..74b5bca 100644 > --- a/drivers/video/fbdev/core/fb_defio.c > +++ b/drivers/video/fbdev/core/fb_defio.c > @@ -164,7 +164,7 @@ static const struct address_space_operations fb_defer= red_io_aops =3D { > .set_page_dirty =3D fb_deferred_io_set_page_dirty, > }; >=20 > -static int fb_deferred_io_mmap(struct fb_info *info, struct vm_area_stru= ct *vma) > +int fb_deferred_io_mmap(struct fb_info *info, struct vm_area_struct *vma) > { > vma->vm_ops =3D &fb_deferred_io_vm_ops; > vma->vm_flags |=3D VM_DONTEXPAND | VM_DONTDUMP; > @@ -173,6 +173,7 @@ static int fb_deferred_io_mmap(struct fb_info *info, = struct vm_area_struct *vma) > vma->vm_private_data =3D info; > return 0; > } > +EXPORT_SYMBOL(fb_deferred_io_mmap); >=20 > /* workqueue callback */ > static void fb_deferred_io_work(struct work_struct *work) > diff --git a/include/linux/fb.h b/include/linux/fb.h > index dfe8835..a964d07 100644 > --- a/include/linux/fb.h > +++ b/include/linux/fb.h > @@ -673,6 +673,7 @@ static inline void __fb_pad_aligned_buffer(u8 *dst, u= 32 d_pitch, > } >=20 > /* drivers/video/fb_defio.c */ > +int fb_deferred_io_mmap(struct fb_info *info, struct vm_area_struct *vma= ); > extern void fb_deferred_io_init(struct fb_info *info); > extern void fb_deferred_io_open(struct fb_info *info, > struct inode *inode, > -- > 2.2.2 >=20 --=20 Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch