On 07/08/2013 04:49 PM, Mark Brown wrote: > On Mon, Jul 08, 2013 at 03:29:16PM +0200, Michal Simek wrote: >> Use devm_request_irq() for irq allocation which >> simplify driver code. > >> @@ -495,7 +493,6 @@ static int xilinx_spi_remove(struct platform_device >> *pdev) >> struct xilinx_spi *xspi = platform_get_drvdata(pdev); >> >> spi_bitbang_stop(&xspi->bitbang); >> - free_irq(xspi->irq, xspi); >> >> spi_master_put(xspi->bitbang.master); > > Is it definitely safe to leave the IRQ hanging around after the master > has been freed - there's no possibility of a late error interrupt or > something?
I think it is more generic question if this race condition is fine for all drivers which are using devres groups. I have just looked at it and devres_release_all() is called where driver is unload and irq are disabled there. It means that all handlers should be unregistered and if IRQ happen after it then it should be handled by Linux kernel if driver doesn't disable it. btw: What's the proper way for spi driver unregistration? spi_unregistered_master() (which also free private structure) and spi_master_put()? Thanks, Michal -- Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91 w: www.monstr.eu p: +42-0-721842854 Maintainer of Linux kernel - Microblaze cpu - http://www.monstr.eu/fdt/ Maintainer of Linux kernel - Xilinx Zynq ARM architecture Microblaze U-BOOT custodian and responsible for u-boot arm zynq platform
signature.asc
Description: OpenPGP digital signature
