linux-next: build failure after merge of the vfio tree

linux-next: build failure after merge of the vfio tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 352 bytes --]

Hi all,

After merging the vfio tree, today's linux-next build (x86_64
allmodconfig) failed like this:


Caused by commit

  cc0ee20bd969 ("vfio/fsl-mc: trigger an interrupt via eventfd")
  ac93ab2bf69a ("vfio/fsl-mc: Add support for device reset")

I have used the vfio tree from next-20201012 for today.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the vfio tree

[Diana Craciun OSS]

Hi,

How does it fail? What's the error?

Thanks,
Diana


On 10/13/2020 6:07 AM, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the vfio tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
> 
> 
> Caused by commit
> 
>    cc0ee20bd969 ("vfio/fsl-mc: trigger an interrupt via eventfd")
>    ac93ab2bf69a ("vfio/fsl-mc: Add support for device reset")
> 
> I have used the vfio tree from next-20201012 for today.
> 

Re: linux-next: build failure after merge of the vfio tree

[Alex Williamson]

On Tue, 13 Oct 2020 18:56:07 +0300
Diana Craciun OSS <diana.craciun@oss.nxp.com> wrote:

> Hi,
> 
> How does it fail? What's the error?
> 
> Thanks,
> Diana
> 
> 
> On 10/13/2020 6:07 AM, Stephen Rothwell wrote:
> > Hi all,
> > 
> > After merging the vfio tree, today's linux-next build (x86_64
> > allmodconfig) failed like this:
> > 
> > 
> > Caused by commit
> > 
> >    cc0ee20bd969 ("vfio/fsl-mc: trigger an interrupt via eventfd")
> >    ac93ab2bf69a ("vfio/fsl-mc: Add support for device reset")
> > 
> > I have used the vfio tree from next-20201012 for today.

Thanks, Stephen.  Diana has posted a 32bit build fix which I've merged,
maybe that was the error.  Also Diana's series in my branch is currently
dependent on fsl-bus support in GregKH's char-misc-next branch.  Looking
at the log from the successful build, I wonder if our branches are just
in the wrong order (vfio/next processed on line 341, char-misc-next
processed on 387).  I don't know if you regularly re-order for this
sort of thing, otherwise it should work out when Greg's branch gets
merged, but testing sooner in next would be preferred.  Thanks,

Alex

Re: linux-next: build failure after merge of the vfio tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 3855 bytes --]

Hi Diana,

On Tue, 13 Oct 2020 18:56:07 +0300 Diana Craciun OSS <diana.craciun@oss.nxp.com> wrote:
>
> Hi,
> 
> How does it fail? What's the error?

Sorry about that:

drivers/vfio/fsl-mc/vfio_fsl_mc_intr.c: In function 'vfio_fsl_mc_set_irq_trigger':
drivers/vfio/fsl-mc/vfio_fsl_mc_intr.c:121:8: error: implicit declaration of function 'fsl_mc_populate_irq_pool' [-Werror=implicit-function-declaration]
  121 |  ret = fsl_mc_populate_irq_pool(mc_cont,
      |        ^~~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc_intr.c:122:4: error: 'FSL_MC_IRQ_POOL_MAX_TOTAL_IRQS' undeclared (first use in this function)
  122 |    FSL_MC_IRQ_POOL_MAX_TOTAL_IRQS);
      |    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_release':
drivers/vfio/fsl-mc/vfio_fsl_mc.c:178:9: error: implicit declaration of function 'dprc_reset_container' [-Werror=implicit-function-declaration]
  178 |   ret = dprc_reset_container(mc_cont->mc_io, 0,
      |         ^~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:181:6: error: 'DPRC_RESET_OPTION_NON_RECURSIVE' undeclared (first use in this function)
  181 |      DPRC_RESET_OPTION_NON_RECURSIVE);
      |      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:181:6: note: each undeclared identifier is reported only once for each function it appears in
drivers/vfio/fsl-mc/vfio_fsl_mc.c:191:3: error: implicit declaration of function 'fsl_mc_cleanup_irq_pool' [-Werror=implicit-function-declaration]
  191 |   fsl_mc_cleanup_irq_pool(mc_cont);
      |   ^~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_ioctl':
drivers/vfio/fsl-mc/vfio_fsl_mc.c:316:9: error: 'DPRC_RESET_OPTION_NON_RECURSIVE' undeclared (first use in this function)
  316 |         DPRC_RESET_OPTION_NON_RECURSIVE);
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_mmap_mmio':
drivers/vfio/fsl-mc/vfio_fsl_mc.c:455:36: error: 'FSL_MC_REGION_CACHEABLE' undeclared (first use in this function)
  455 |  region_cacheable = (region.type & FSL_MC_REGION_CACHEABLE) &&
      |                                    ^~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:456:22: error: 'FSL_MC_REGION_SHAREABLE' undeclared (first use in this function)
  456 |       (region.type & FSL_MC_REGION_SHAREABLE);
      |                      ^~~~~~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_bus_notifier':
drivers/vfio/fsl-mc/vfio_fsl_mc.c:522:9: error: 'struct fsl_mc_device' has no member named 'driver_override'
  522 |   mc_dev->driver_override = kasprintf(GFP_KERNEL, "%s",
      |         ^~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:524:14: error: 'struct fsl_mc_device' has no member named 'driver_override'
  524 |   if (!mc_dev->driver_override)
      |              ^~
drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_init_device':
drivers/vfio/fsl-mc/vfio_fsl_mc.c:561:8: error: implicit declaration of function 'dprc_setup' [-Werror=implicit-function-declaration]
  561 |  ret = dprc_setup(mc_dev);
      |        ^~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:567:8: error: implicit declaration of function 'dprc_scan_container' [-Werror=implicit-function-declaration]
  567 |  ret = dprc_scan_container(mc_dev, false);
      |        ^~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:576:2: error: implicit declaration of function 'dprc_remove_devices' [-Werror=implicit-function-declaration]
  576 |  dprc_remove_devices(mc_dev, NULL, 0);
      |  ^~~~~~~~~~~~~~~~~~~
drivers/vfio/fsl-mc/vfio_fsl_mc.c:577:2: error: implicit declaration of function 'dprc_cleanup' [-Werror=implicit-function-declaration]
  577 |  dprc_cleanup(mc_dev);
      |  ^~~~~~~~~~~~

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 484 bytes --]

Re: linux-next: build failure after merge of the vfio tree

[Diana Craciun OSS]

Thank you. The errors indicate that, most likely, the fsl-mc-bus patches 
from char-misc-next are missing at this point. I have added the vfio 
patches on top of linux-next (which already contains the fsl-mc-bus 
patches) and built x86_64 allmodconfig. There were no errors.

Thanks,
Diana

On 10/14/2020 1:16 AM, Stephen Rothwell wrote:
> Hi Diana,
> 
> On Tue, 13 Oct 2020 18:56:07 +0300 Diana Craciun OSS <diana.craciun@oss.nxp.com> wrote:
>>
>> Hi,
>>
>> How does it fail? What's the error?
> 
> Sorry about that:
> 
> drivers/vfio/fsl-mc/vfio_fsl_mc_intr.c: In function 'vfio_fsl_mc_set_irq_trigger':
> drivers/vfio/fsl-mc/vfio_fsl_mc_intr.c:121:8: error: implicit declaration of function 'fsl_mc_populate_irq_pool' [-Werror=implicit-function-declaration]
>    121 |  ret = fsl_mc_populate_irq_pool(mc_cont,
>        |        ^~~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc_intr.c:122:4: error: 'FSL_MC_IRQ_POOL_MAX_TOTAL_IRQS' undeclared (first use in this function)
>    122 |    FSL_MC_IRQ_POOL_MAX_TOTAL_IRQS);
>        |    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_release':
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:178:9: error: implicit declaration of function 'dprc_reset_container' [-Werror=implicit-function-declaration]
>    178 |   ret = dprc_reset_container(mc_cont->mc_io, 0,
>        |         ^~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:181:6: error: 'DPRC_RESET_OPTION_NON_RECURSIVE' undeclared (first use in this function)
>    181 |      DPRC_RESET_OPTION_NON_RECURSIVE);
>        |      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:181:6: note: each undeclared identifier is reported only once for each function it appears in
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:191:3: error: implicit declaration of function 'fsl_mc_cleanup_irq_pool' [-Werror=implicit-function-declaration]
>    191 |   fsl_mc_cleanup_irq_pool(mc_cont);
>        |   ^~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_ioctl':
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:316:9: error: 'DPRC_RESET_OPTION_NON_RECURSIVE' undeclared (first use in this function)
>    316 |         DPRC_RESET_OPTION_NON_RECURSIVE);
>        |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_mmap_mmio':
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:455:36: error: 'FSL_MC_REGION_CACHEABLE' undeclared (first use in this function)
>    455 |  region_cacheable = (region.type & FSL_MC_REGION_CACHEABLE) &&
>        |                                    ^~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:456:22: error: 'FSL_MC_REGION_SHAREABLE' undeclared (first use in this function)
>    456 |       (region.type & FSL_MC_REGION_SHAREABLE);
>        |                      ^~~~~~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_bus_notifier':
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:522:9: error: 'struct fsl_mc_device' has no member named 'driver_override'
>    522 |   mc_dev->driver_override = kasprintf(GFP_KERNEL, "%s",
>        |         ^~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:524:14: error: 'struct fsl_mc_device' has no member named 'driver_override'
>    524 |   if (!mc_dev->driver_override)
>        |              ^~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c: In function 'vfio_fsl_mc_init_device':
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:561:8: error: implicit declaration of function 'dprc_setup' [-Werror=implicit-function-declaration]
>    561 |  ret = dprc_setup(mc_dev);
>        |        ^~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:567:8: error: implicit declaration of function 'dprc_scan_container' [-Werror=implicit-function-declaration]
>    567 |  ret = dprc_scan_container(mc_dev, false);
>        |        ^~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:576:2: error: implicit declaration of function 'dprc_remove_devices' [-Werror=implicit-function-declaration]
>    576 |  dprc_remove_devices(mc_dev, NULL, 0);
>        |  ^~~~~~~~~~~~~~~~~~~
> drivers/vfio/fsl-mc/vfio_fsl_mc.c:577:2: error: implicit declaration of function 'dprc_cleanup' [-Werror=implicit-function-declaration]
>    577 |  dprc_cleanup(mc_dev);
>        |  ^~~~~~~~~~~~
> 

Re: linux-next: build failure after merge of the vfio tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1016 bytes --]

Hi Alex,

On Tue, 13 Oct 2020 13:20:16 -0600 Alex Williamson <alex.williamson@redhat.com> wrote:
>
> Thanks, Stephen.  Diana has posted a 32bit build fix which I've merged,
> maybe that was the error.  Also Diana's series in my branch is currently
> dependent on fsl-bus support in GregKH's char-misc-next branch.  Looking
> at the log from the successful build, I wonder if our branches are just
> in the wrong order (vfio/next processed on line 341, char-misc-next
> processed on 387).  I don't know if you regularly re-order for this
> sort of thing, otherwise it should work out when Greg's branch gets
> merged, but testing sooner in next would be preferred.

I have put the vfio tree after the char-misc tree today (so hopefully
it will build).  The proper way to do this is for you and Greg to have
a shared branch with the commits you both depend on and bot merge that
branch.  That way, it doesn't matter what order the tress are merged
(by me or Linus).

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 484 bytes --]

Re: linux-next: build failure after merge of the vfio tree

[Alex Williamson]

On Thu, 15 Oct 2020 11:59:03 +1100
Stephen Rothwell <sfr@canb.auug.org.au> wrote:

> Hi Alex,
> 
> On Tue, 13 Oct 2020 13:20:16 -0600 Alex Williamson <alex.williamson@redhat.com> wrote:
> >
> > Thanks, Stephen.  Diana has posted a 32bit build fix which I've merged,
> > maybe that was the error.  Also Diana's series in my branch is currently
> > dependent on fsl-bus support in GregKH's char-misc-next branch.  Looking
> > at the log from the successful build, I wonder if our branches are just
> > in the wrong order (vfio/next processed on line 341, char-misc-next
> > processed on 387).  I don't know if you regularly re-order for this
> > sort of thing, otherwise it should work out when Greg's branch gets
> > merged, but testing sooner in next would be preferred.  
> 
> I have put the vfio tree after the char-misc tree today (so hopefully
> it will build).  The proper way to do this is for you and Greg to have
> a shared branch with the commits you both depend on and bot merge that
> branch.  That way, it doesn't matter what order the tress are merged
> (by me or Linus).

Hi Stephen,

Well that seems like the obviously correct solution in retrospect ;)
thanks for the explanation.  I'll check-in with Greg to see if his pull
request is imminent or we can share a branch.  Thanks,

Alex