[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20130709144704.GY27646@sirena.org.uk>
Date: Tue, 9 Jul 2013 15:47:04 +0100
From: Mark Brown <broonie@...nel.org>
To: Michal Simek <monstr@...str.eu>
Cc: Michal Simek <michal.simek@...inx.com>,
linux-kernel@...r.kernel.org,
linux-spi <linux-spi@...r.kernel.org>,
Grant Likely <grant.likely@...aro.org>,
spi-devel-general@...ts.sourceforge.net
Subject: Re: [PATCH v1 3/4] spi/xilinx: Simplify irq allocation
On Tue, Jul 09, 2013 at 04:15:13PM +0200, Michal Simek wrote:
> 4. spi_master_release()
> 5. devres_release() with irq
> If interrupt happends between 4 and 5 than it can be the problem.
> Do I understand you correctly?
Yes.
> If this is problematic case we can disable local and global interrupts
> and add it between 2 and 3 or 3-4.
> /* Disable all the interrupts just in case */
> xspi->write_fn(0, regs_base + XIPIF_V123B_IIER_OFFSET);
> /* Disable the global IPIF interrupt */
> xspi->write_fn(0, regs_base + XIPIF_V123B_DGIER_OFFSET);
> What do you think about this solution?
That's fine, though if just manually freeing the IRQ works that's also
OK...
> I have also tried to run one thing with and without this patch
> and the results are below.
> When I add this irq disable function between 1 and 2 then
> module removing stucks.
You'll need to wait until the device is quiesced at least if it's
relying on the interrupt to complete operations.
Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)
Powered by blists - more mailing lists