Re: [PATCH] spi: mark device nodes only in case of successful instantiation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Oct 17, 2016 at 10:20:47PM +0300, Pantelis Antoniou wrote:
> Hi Ralf,
> 
> > On Oct 16, 2016, at 12:55 , Ralf Ramsauer <ralf@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> > 
> > Hi Geert,
> > 
> > On 10/16/2016 10:49 AM, Geert Uytterhoeven wrote:
> >> Hi Ralf,
> >> 
> >> (Cc i2c)
> >> 
> >> On Fri, Oct 14, 2016 at 9:31 PM, Ralf Ramsauer
> >> <ralf@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> >>> Instantiated SPI device nodes are marked with OF_POPULATE. This was
> >>> introduced in bd6c164. On unloading, loaded device nodes will of course
> >>> be unmarked. The problem are nodes the fail during initialisation: If a
> >>> node failed during registration, it won't be unloaded and hence never be
> >>> unmarked again.
> >>> 
> >>> So if a SPI driver module is unloaded and reloaded, it will skip nodes
> >>> that failed before.
> >>> 
> >>> Skip device nodes that are already populated and mark them only in case
> >>> of success.
> >>> 
> >>> Fixes: bd6c164 ("spi: Mark instantiated device nodes with OF_POPULATE")
> >>> Signed-off-by: Ralf Ramsauer <ralf@xxxxxxxxxxxxxxxxxxxxxx>
> >>> Cc: Geert Uytterhoeven <geert+renesas@xxxxxxxxx>
> >>> ---
> >>> Hi,
> >>> 
> >>> imagine the following situation: you loaded a spi driver as module, but
> >>> it fails to instantiate, because of some reasons (e.g. some resources,
> >>> like gpios, might be in use in userspace).
> >>> 
> >>> When reloading the driver, _all_ nodes, including previously failed
> >>> ones, should be probed again. This is not the case at the moment.
> >>> Current behaviour only re-registers nodes that were previously
> >>> successfully loaded.
> >>> 
> >>> This small patches fixes this behaviour. I stumbled over this while
> >>> working on a spi driver.
> >> 
> >> Thanks for your patch!
> >> 
> >>> drivers/spi/spi.c | 7 +++++--
> >>> 1 file changed, 5 insertions(+), 2 deletions(-)
> >>> 
> >>> diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
> >>> index 200ca22..f96a04e 100644
> >>> --- a/drivers/spi/spi.c
> >>> +++ b/drivers/spi/spi.c
> >>> @@ -1604,12 +1604,15 @@ static void of_register_spi_devices(struct spi_master *master)
> >>>                return;
> >>> 
> >>>        for_each_available_child_of_node(master->dev.of_node, nc) {
> >>> -               if (of_node_test_and_set_flag(nc, OF_POPULATED))
> >>> +               if (of_node_check_flag(nc, OF_POPULATED))
> >>>                        continue;
> >>>                spi = of_register_spi_device(master, nc);
> >>> -               if (IS_ERR(spi))
> >>> +               if (IS_ERR(spi)) {
> >>>                        dev_warn(&master->dev, "Failed to create SPI device for %s\n",
> >>>                                nc->full_name);
> >>> +                       continue;
> >>> +               }
> >>> +               of_node_set_flag(nc, OF_POPULATED);
> >> 
> >> I think it's safer to keep the atomic test-and-set, but clear the flag on
> >> failure, cfr. of_platform_device_create_pdata() and of_amba_device_create().
> > Ack, no prob. Let me change this in the next version.
> >> 

> Thanks for this. This is a very rare case that’s easy to slip through.
> It is good to be consistent :)

I read this as acked-by for the series?

Attachment: signature.asc
Description: PGP signature


[Index of Archives]     [Linux Kernel]     [Linux ARM (vger)]     [Linux ARM MSM]     [Linux Omap]     [Linux Arm]     [Linux Tegra]     [Fedora ARM]     [Linux for Samsung SOC]     [eCos]     [Linux Fastboot]     [Gcc Help]     [Git]     [DCCP]     [IETF Announce]     [Security]     [Linux MIPS]     [Yosemite Campsites]

  Powered by Linux