Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1431964 > unrolled thread

Re: next build: 142 builds: 6 failed, 136 passed, 17 errors, 111 warnings (next-20160624)

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: next build: 142 builds: 6 failed, 136 passed, 17 errors, 111 warnings (next-20160624) Arnd Bergmann <arnd@arndb.de> - 2016-06-27 13:40 +0200
    Re: next build: 142 builds: 6 failed, 136 passed, 17 errors, 111 warnings (next-20160624) Kevin Hilman <khilman@baylibre.com> - 2016-06-27 22:40 +0200
      Re: next build: 142 builds: 6 failed, 136 passed, 17 errors, 111 warnings (next-20160624) Arnd Bergmann <arnd@arndb.de> - 2016-06-27 23:10 +0200

#1431964 — Re: next build: 142 builds: 6 failed, 136 passed, 17 errors, 111 warnings (next-20160624)

On Friday, June 24, 2016 2:28:59 PM CEST kernelci. org bot wrote:
> 
> Errors summary:
> 
>       3  arch/arm/include/asm/jump_label.h:13:7: error: impossible constraint in 'asm'
>      66  arch/arm/include/asm/jump_label.h:13:7: warning: asm operand 0 probably doesn't match constraints

I've just sent a patch for this, hopefully it works again soon.

>       1  fs/fat/dir.c:758:424: internal compiler error: Segmentation fault

I'm pretty sure this is fixed with gcc-6.1.1, but I cannot reproduce the problem
here. I've tried various compilers.

>       2  drivers/char/mwave/smapi.h:53:0: warning: "FALSE" redefined [enabled by default]
>       2  drivers/char/mwave/smapi.h:52:0: warning: "TRUE" redefined [enabled by default]

This should be fixed in tomorrow's linux-next, I did two different patches for it,
and the other one should get merged as well.

>       1  arch/arm/configs/mps2_defconfig:102:warning: symbol value 'y' invalid for PRINTK_TIME

I should do a patch for this.

>       1  .config:972:warning: override: NOHIGHMEM changes choice state
>       1  .config:971:warning: override: SLOB changes choice state
>       1  .config:969:warning: override: KERNEL_XZ changes choice state
>       1  .config:968:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
>       1  .config:944:warning: override: SLOB changes choice state
>       1  .config:941:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
>       1  .config:877:warning: override: SLOB changes choice state
>       1  .config:875:warning: override: KERNEL_XZ changes choice state
>       1  .config:874:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state

The tinyconfig warnings are really annoying :(

> --------------------------------------------------------------------------------
> allmodconfig (arm64) — FAIL, 2 errors, 3 warnings, 0 section mismatches
> 
> Errors:
>     binder.c:(.text+0x294f54): undefined reference to `usb_ep_set_maxpacket_limit'
>     binder.c:(.text+0x295030): undefined reference to `usb_ep_set_maxpacket_limit'

I sent a fix.

>     net/rxrpc/peer_object.c:57:15: warning: 'p' may be used uninitialized in this function [-Wmaybe-uninitialized]

My fix was merged, should be good today.

>     drivers/media/dvb-frontends/cxd2841er.c:3217:552: warning: 'carrier_offset' may be used uninitialized in this function [-Wmaybe-uninitialized]

I don't see this one. It's a false positive.

> --------------------------------------------------------------------------------
> allmodconfig (x86) — PASS, 0 errors, 14 warnings, 0 section mismatches
> 
> Warnings:
>     net/bluetooth/6lowpan.c:608:570: warning: 'addr_type' may be used uninitialized in this function [-Wmaybe-uninitialized]
>     lib/test_hash.c:224:7: warning: "HAVE_ARCH__HASH_32" is not defined [-Wundef]
>     lib/test_hash.c:229:7: warning: "HAVE_ARCH_HASH_32" is not defined [-Wundef]
>     lib/test_hash.c:234:7: warning: "HAVE_ARCH_HASH_64" is not defined [-Wundef]
>     lib/test_hash.c:146:2: warning: missing braces around initializer [-Wmissing-braces]
>     lib/test_hash.c:146:2: warning: (near initialization for 'hash_or[0]') [-Wmissing-braces]

The problem only exists with gcc-4.9 or earlier. Geert has sent a patch a month
ago but nothing happened.

>     drivers/extcon/extcon.c:435:6: warning: 'idx' may be used uninitialized in this function [-Wmaybe-uninitialized]
>     drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c:89:90: warning: 'type' may be used uninitialized in this function [-Wmaybe-uninitialized]
>     drivers/media/usb/dvb-usb/dvb-usb-dvb.c:323:5: warning: 'ret' may be used uninitialized in this function [-Wmaybe-uninitialized]
>     drivers/media/dvb-frontends/cxd2841er.c:3217:569: warning: 'carrier_offset' may be used uninitialized in this function [-Wmaybe-uninitialized]
>     drivers/net/xen-netback/hash.c:127:6: warning: 'val' may be used uninitialized in this function [-Wmaybe-uninitialized]

I think these are all the result of using a pre-4.9 gcc and they are
all false positives. Maybe we can just call those older versions with
-Wno-maybe-uninitialized in the kernelci infrastructure?

> --------------------------------------------------------------------------------
> allmodconfig+CONFIG_OF=n (x86) — FAIL, 4 errors, 14 warnings, 0 section mismatches
> 
> Errors:

>     drivers/pinctrl/bcm/pinctrl-iproc-gpio.c:381:20: error: 'pinconf_generic_dt_node_to_map_pin' undeclared here (not in a function)
>     drivers/pinctrl/bcm/pinctrl-cygnus-mux.c:739:20: error: 'pinconf_generic_dt_node_to_map_group' undeclared here (not in a function)
>     drivers/pinctrl/bcm/pinctrl-nsp-gpio.c:365:20: error: 'pinconf_generic_dt_node_to_map_pin' undeclared here (not in a function)

Haven't looked at the x86 specific problem yet, and I don't immediately see
what's wrong here. We appear to always set GENERIC_PINCONF from the driver.

> --------------------------------------------------------------------------------
> multi_v7_defconfig+CONFIG_SMP=n (arm) — FAIL, 4 errors, 0 warnings, 0 section mismatches
> 
> Errors:
>     dss-common.c:(.text+0x52bc): undefined reference to `secondary_startup'
>     dss-common.c:(.text+0x52e8): undefined reference to `secondary_startup'
>     dss-common.c:(.text+0x530c): undefined reference to `secondary_startup'
>     dss-common.c:(.text+0x5340): undefined reference to `secondary_startup'

I reported the problem, but don't have an idea for a good fix. This broke
with 0573b957fc21 ("ARM: OMAP4+: Prevent CPU1 related hang with kexec").

> --------------------------------------------------------------------------------
> pxa_defconfig (arm) — FAIL, 3 errors, 0 warnings, 0 section mismatches
> 
> Errors:
>     /home/buildslave/workspace/khilman-kbuilder/next/build/../drivers/of/of_mdio.c:435: undefined reference to `fixed_phy_register'
>     /home/buildslave/workspace/khilman-kbuilder/next/build/../drivers/of/of_mdio.c:413: undefined reference to `fixed_phy_register'
>     /home/buildslave/workspace/khilman-kbuilder/next/build/../drivers/of/of_mdio.c:447: undefined reference to `fixed_phy_register'

I sent a patch for this.

	Arnd

[toc] | [next] | [standalone]

#1432302

Arnd Bergmann <arnd@arndb.de> writes:

> On Friday, June 24, 2016 2:28:59 PM CEST kernelci. org bot wrote:

[...]

>>       1  .config:972:warning: override: NOHIGHMEM changes choice state
>>       1  .config:971:warning: override: SLOB changes choice state
>>       1  .config:969:warning: override: KERNEL_XZ changes choice state
>>       1  .config:968:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
>>       1  .config:944:warning: override: SLOB changes choice state
>>       1  .config:941:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
>>       1  .config:877:warning: override: SLOB changes choice state
>>       1  .config:875:warning: override: KERNEL_XZ changes choice state
>>       1  .config:874:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
>
> The tinyconfig warnings are really annoying :(
>

Hmm, yeah.  We should probably just filter out all the kconfig warnings
that happen when we use config fragments, since the point of the
fragment is to override things and change state.

Right now, our check for warnings is a rather simple grep, but I'll add
this to the TODO list.

Kevin

[toc] | [prev] | [next] | [standalone]

#1432313

On Monday, June 27, 2016 1:30:17 PM CEST Kevin Hilman wrote:
> Arnd Bergmann <arnd@arndb.de> writes:
> 
> > On Friday, June 24, 2016 2:28:59 PM CEST kernelci. org bot wrote:
> 
> [...]
> 
> >>       1  .config:972:warning: override: NOHIGHMEM changes choice state
> >>       1  .config:971:warning: override: SLOB changes choice state
> >>       1  .config:969:warning: override: KERNEL_XZ changes choice state
> >>       1  .config:968:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
> >>       1  .config:944:warning: override: SLOB changes choice state
> >>       1  .config:941:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
> >>       1  .config:877:warning: override: SLOB changes choice state
> >>       1  .config:875:warning: override: KERNEL_XZ changes choice state
> >>       1  .config:874:warning: override: CC_OPTIMIZE_FOR_SIZE changes choice state
> >
> > The tinyconfig warnings are really annoying 
> >
> 
> Hmm, yeah.  We should probably just filter out all the kconfig warnings
> that happen when we use config fragments, since the point of the
> fragment is to override things and change state.
> 
> Right now, our check for warnings is a rather simple grep, but I'll add
> this to the TODO list.

I've looked at this before and sadly, the scripts/kconfig tool can do it
right if you pass in both the base config file (defconfig, or flag for
allnoconfig/allmod/...) *and* the fragment, but the way that the Makefile
works for minconfig is that we first generate a .config file and
then add the fragment for minconfig to that, and this causes the
warnings above.

This should be fixable by patching the kernel Makefiles, but I have
not come up with a nice way to do that.

	Arnd

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web