summaryrefslogtreecommitdiff
path: root/drivers/nfc/pn533/Kconfig
diff options
context:
space:
mode:
authorLars Poeschel <poeschel@lemonage.de>2019-10-29 17:46:46 +0300
committerDavid S. Miller <davem@davemloft.net>2019-10-30 07:05:26 +0300
commit843cc92ed323692943c94d7e6ce97a0353b8c2d7 (patch)
treee2a9210b2aa95ba8f3ad4560092ad9fba4edb9a2 /drivers/nfc/pn533/Kconfig
parent0bf2840ccc6efcba82d83b224dcde19dea9f1ee3 (diff)
downloadlinux-843cc92ed323692943c94d7e6ce97a0353b8c2d7.tar.xz
nfc: pn533: Split pn533 init & nfc_register
There is a problem in the initialisation and setup of the pn533: It registers with nfc too early. It could happen, that it finished registering with nfc and someone starts using it. But setup of the pn533 is not yet finished. Bad or at least unintended things could happen. So I split out nfc registering (and unregistering) to seperate functions that have to be called late in probe then. i2c requires a bit more love: i2c requests an irq in it's probe function. 'Commit 32ecc75ded72 ("NFC: pn533: change order operations in dev registation")' shows, this can not happen too early. An irq can be served before structs are fully initialized. The way chosen to prevent this is to request the irq after nfc_alloc_device initialized the structs, but before nfc_register_device. So there is now this pn532_i2c_nfc_alloc function. Cc: Johan Hovold <johan@kernel.org> Cc: Claudiu Beznea <Claudiu.Beznea@microchip.com> Cc: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: Lars Poeschel <poeschel@lemonage.de> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'drivers/nfc/pn533/Kconfig')
0 files changed, 0 insertions, 0 deletions