Re: [PATCH 4.6 00/96] 4.6.6-stable review

[Paul Burton]

On 09/08/16 09:24, Greg Kroah-Hartman wrote:
> On Mon, Aug 08, 2016 at 10:03:39PM -0700, Guenter Roeck wrote:
> > On 08/08/2016 12:10 PM, Greg Kroah-Hartman wrote:
> > > This is the start of the stable review cycle for the 4.6.6 release.
> > > There are 96 patches in this series, all will be posted as a response
> > > to this one.  If anyone has any issues with these being applied, please
> > > let me know.
> > >
> > > Responses should be made by Wed Aug 10 18:02:27 UTC 2016.
> > > Anything received after that time might be too late.
> >
> > Build results:
> > 	total: 148 pass: 146 fail: 2
> > Failed builds:
> > 	unicore32:defconfig
> > 	unicore32:allnoconfig
> >
> > Qemu test results:
> > 	total: 107 pass: 103 fail: 4
> > Failed tests:
> > 	mips:malta_defconfig:nosmp
> > 	mips64:malta_defconfig:nosmp
> > 	mipsel:malta_defconfig:nosmp
> > 	mipsel64:malta_defconfig:nosmp
> >
> > The unicore32 build error is still not fixed in mainline.
> >
> > mips:
> >
> > Building mips:malta_defconfig:nosmp ... failed
> > ------------
> > Error log:
> > drivers/built-in.o: In function `gic_shared_irq_domain_map.isra.0':
> > irq-mips-gic.c:(.text+0x1d58): undefined reference to `smp_num_siblings'
> > irq-mips-gic.c:(.text+0x1d5c): undefined reference to `smp_num_siblings'
> > irq-mips-gic.c:(.text+0x1de0): undefined reference to `smp_num_siblings'
> > irq-mips-gic.c:(.text+0x1de8): undefined reference to `smp_num_siblings'
> >
> > Only seen with non-SMP builds. Tricky one, required bisect.
> >
> > # bad: [0353fa20cb40ea75a7e3b7293b533bde4fa013e9] Linux 4.6.6-rc1
> > # good: [bed4c611a3b80fae04d75b9f85fcaf174bac1b06] Linux 4.6.5
> > git bisect start 'HEAD' 'v4.6.5'
> > # good: [8d3b41c18cff2db7ccea71243b945a33610b8cb5] cgroup: remove redundant cleanup in css_create
> > git bisect good 8d3b41c18cff2db7ccea71243b945a33610b8cb583c9f9e994f5c963fc652a52befc120699dcc591
> > # good: [3a255979bc75ffcc263928045df4b5f46df8a8ca] sched/fair: Fix effective_load() to consistently use smoothed load
> > git bisect good 3a255979bc75ffcc263928045df4b5f46df8a8ca
> > # bad: [6eafb1f777de375b52c8a83d262da2885a28579a] posix_cpu_timer: Exit early when process has been reaped
> > git bisect bad 6eafb1f777de375b52c8a83d262da2885a28579a
> > # bad: [83c9f9e994f5c963fc652a52befc120699dcc591] irqchip/mips-gic: Map to VPs using HW VPNum
> > git bisect bad 83c9f9e994f5c963fc652a52befc120699dcc591
> > # good: [3a1e9944db1570f9b654b7a187f13898e5c59529] can: fix handling of unmodifiable configuration options fix
> > git bisect good 3a1e9944db1570f9b654b7a187f13898e5c59529
> > # good: [f06b9b8071ae6e383931795e2a459d6abc0bc7bd] RDS: fix rds_tcp_init() error path
> > git bisect good f06b9b8071ae6e383931795e2a459d6abc0bc7bd
> > # first bad commit: [83c9f9e994f5c963fc652a52befc120699dcc591] irqchip/mips-gic: Map to VPs using HW VPNum
> >
> > Reverting 83c9f9e994f5c963fc652a52befc120699dcc591 fixes the problem.
> > Its description suggests that it fixes a real bug, though, so copying Paul for input.
>
> Thanks for the bisection and report.
>
> Paul, any ideas?  Why doesn't this also fail in 4.7?
>
> thanks,
>
> greg k-h

Hi Greg, Guenter,

The easiest way to fix this is likely to backport a60ae81e5e59 ("MIPS: 
CM: Fix mips_cm_max_vp_width for UP kernels") which went into mainline 
in the 4.7 cycle. Apologies for overlooking the need for that one in stable.

Thanks,
    Paul