* [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt [not found] ` <20150324161358.GA694@kahuna> @ 2015-03-24 18:42 ` grygorii.strashko 2015-03-24 20:00 ` Murali Karicheri ` (2 more replies) [not found] ` <20150326003939.GA25368@verge.net.au> 1 sibling, 3 replies; 21+ messages in thread From: grygorii.strashko @ 2015-03-24 18:42 UTC (permalink / raw) To: linux, Linus Walleij, Alexandre Courbot Cc: m-karicheri2, ssantosh, linux-gpio, linux-kernel, Grygorii Strashko From: Grygorii Strashko <grygorii.strashko@linaro.org> Now GPIO syscon driver produces bunch of warnings during the boot of Kesytone 2 SoCs: gpio-syscon soc:keystone_dsp_gpio@02620240: can't read the dir register offset! gpio-syscon soc:keystone_dsp_gpio@2620244: can't read the dir register offset! This message unintentionally was added using dev_err(), but its actual log level is debug, because third cell of "ti,syscon-dev" is optional. Hence change it to dev_dbg() as it should be. This patch fixes commit: 5a3e3f8 ("gpio: syscon: retriave syscon node and regs offsets from dt") Reported-by: Russell King <linux@arm.linux.org.uk> Signed-off-by: Grygorii Strashko <grygorii.strashko@linaro.org> --- drivers/gpio/gpio-syscon.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpio/gpio-syscon.c b/drivers/gpio/gpio-syscon.c index 257e298..045a952 100644 --- a/drivers/gpio/gpio-syscon.c +++ b/drivers/gpio/gpio-syscon.c @@ -219,7 +219,7 @@ static int syscon_gpio_probe(struct platform_device *pdev) ret = of_property_read_u32_index(np, "gpio,syscon-dev", 2, &priv->dir_reg_offset); if (ret) - dev_err(dev, "can't read the dir register offset!\n"); + dev_dbg(dev, "can't read the dir register offset!\n"); priv->dir_reg_offset <<= 3; } -- 1.9.1 ^ permalink raw reply related [flat|nested] 21+ messages in thread
* Re: [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt 2015-03-24 18:42 ` [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt grygorii.strashko @ 2015-03-24 20:00 ` Murali Karicheri 2015-03-24 20:52 ` santosh shilimkar 2015-03-27 10:18 ` Linus Walleij 2 siblings, 0 replies; 21+ messages in thread From: Murali Karicheri @ 2015-03-24 20:00 UTC (permalink / raw) To: grygorii.strashko Cc: linux, Linus Walleij, Alexandre Courbot, ssantosh, linux-gpio, linux-kernel On 03/24/2015 02:42 PM, grygorii.strashko@linaro.org wrote: > From: Grygorii Strashko<grygorii.strashko@linaro.org> > > Now GPIO syscon driver produces bunch of warnings during the > boot of Kesytone 2 SoCs: > gpio-syscon soc:keystone_dsp_gpio@02620240: can't read the dir register offset! > gpio-syscon soc:keystone_dsp_gpio@2620244: can't read the dir register offset! > > This message unintentionally was added using dev_err(), but its > actual log level is debug, because third cell of "ti,syscon-dev" is > optional. > > Hence change it to dev_dbg() as it should be. > > This patch fixes commit: > 5a3e3f8 ("gpio: syscon: retriave syscon node and regs offsets from dt") > > Reported-by: Russell King<linux@arm.linux.org.uk> > Signed-off-by: Grygorii Strashko<grygorii.strashko@linaro.org> > --- > drivers/gpio/gpio-syscon.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpio/gpio-syscon.c b/drivers/gpio/gpio-syscon.c > index 257e298..045a952 100644 > --- a/drivers/gpio/gpio-syscon.c > +++ b/drivers/gpio/gpio-syscon.c > @@ -219,7 +219,7 @@ static int syscon_gpio_probe(struct platform_device *pdev) > ret = of_property_read_u32_index(np, "gpio,syscon-dev", 2, > &priv->dir_reg_offset); > if (ret) > - dev_err(dev, "can't read the dir register offset!\n"); > + dev_dbg(dev, "can't read the dir register offset!\n"); > > priv->dir_reg_offset<<= 3; > } Tested this on K2HK EVM and found the error log is not appearing any more. Tested-by: Murali Karicheri <m-karicheri2@ti.com> -- Murali Karicheri Linux Kernel, Texas Instruments ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt 2015-03-24 18:42 ` [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt grygorii.strashko 2015-03-24 20:00 ` Murali Karicheri @ 2015-03-24 20:52 ` santosh shilimkar 2015-03-27 10:18 ` Linus Walleij 2 siblings, 0 replies; 21+ messages in thread From: santosh shilimkar @ 2015-03-24 20:52 UTC (permalink / raw) To: grygorii.strashko, linux, Linus Walleij, Alexandre Courbot Cc: m-karicheri2, ssantosh, linux-gpio, linux-kernel On 3/24/2015 11:42 AM, grygorii.strashko@linaro.org wrote: > From: Grygorii Strashko <grygorii.strashko@linaro.org> > > Now GPIO syscon driver produces bunch of warnings during the > boot of Kesytone 2 SoCs: > gpio-syscon soc:keystone_dsp_gpio@02620240: can't read the dir register offset! > gpio-syscon soc:keystone_dsp_gpio@2620244: can't read the dir register offset! > > This message unintentionally was added using dev_err(), but its > actual log level is debug, because third cell of "ti,syscon-dev" is > optional. > > Hence change it to dev_dbg() as it should be. > > This patch fixes commit: > 5a3e3f8 ("gpio: syscon: retriave syscon node and regs offsets from dt") > > Reported-by: Russell King <linux@arm.linux.org.uk> > Signed-off-by: Grygorii Strashko <grygorii.strashko@linaro.org> > --- Acked-by: Santosh Shilimkar <ssantosh@kernel.org> ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt 2015-03-24 18:42 ` [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt grygorii.strashko 2015-03-24 20:00 ` Murali Karicheri 2015-03-24 20:52 ` santosh shilimkar @ 2015-03-27 10:18 ` Linus Walleij 2 siblings, 0 replies; 21+ messages in thread From: Linus Walleij @ 2015-03-27 10:18 UTC (permalink / raw) To: Grygorii Strashko Cc: Russell King - ARM Linux, Alexandre Courbot, Muralidharan Karicheri, Santosh Shilimkar, linux-gpio, linux-kernel On Tue, Mar 24, 2015 at 7:42 PM, <grygorii.strashko@linaro.org> wrote: > From: Grygorii Strashko <grygorii.strashko@linaro.org> > > Now GPIO syscon driver produces bunch of warnings during the > boot of Kesytone 2 SoCs: > gpio-syscon soc:keystone_dsp_gpio@02620240: can't read the dir register offset! > gpio-syscon soc:keystone_dsp_gpio@2620244: can't read the dir register offset! > > This message unintentionally was added using dev_err(), but its > actual log level is debug, because third cell of "ti,syscon-dev" is > optional. > > Hence change it to dev_dbg() as it should be. > > This patch fixes commit: > 5a3e3f8 ("gpio: syscon: retriave syscon node and regs offsets from dt") > > Reported-by: Russell King <linux@arm.linux.org.uk> > Signed-off-by: Grygorii Strashko <grygorii.strashko@linaro.org> OK sorry for missing this when asking about the strange messages. Patch applied for fixes with these test tags and ack. Yours, Linus Walleij ^ permalink raw reply [flat|nested] 21+ messages in thread
[parent not found: <20150326003939.GA25368@verge.net.au>]
[parent not found: <20150326133631.GB2805@arm.com>]
[parent not found: <CANMBJr68dsbYvvHUzy6U4m4fEM6nq8dVHBH4kLQ=0c4QNOhLPQ@mail.gmail.com>]
[parent not found: <20150327002554.GA5527@verge.net.au>]
[parent not found: <20150327100612.GB1562@arm.com>]
[parent not found: <7hbnj99epe.fsf@deeprootsystems.com>]
[parent not found: <CAKv+Gu_ZHZFm-1eXn+r7fkEHOxqSmj+Q+Mmy7k6LK531vSfAjQ@mail.gmail.com>]
[parent not found: <7h8uec95t2.fsf@deeprootsystems.com>]
* [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 (was: Re: Possible regression in gcc 4.7.3 next-20150323 due to "ARM, arm64: kvm: get rid of the bounce page") [not found] ` <7h8uec95t2.fsf@deeprootsystems.com> @ 2015-04-01 9:37 ` Geert Uytterhoeven 2015-04-01 9:47 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 Marc Zyngier 2015-04-01 19:27 ` Kevin Hilman 0 siblings, 2 replies; 21+ messages in thread From: Geert Uytterhoeven @ 2015-04-01 9:37 UTC (permalink / raw) To: Kevin Hilman Cc: Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Marc Zyngier, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Andrew Morton, Linux Kernel Development, linux-mm Hi Kevin, On Tue, 31 Mar 2015, Kevin Hilman wrote: > Ard Biesheuvel <ard.biesheuvel@linaro.org> writes: > Nope, that branch is already part of linux-next, and linux-next still > fails to compile for 20+ defconfigs[1] > > > Could you elaborate on the issue please? What is the error you are > > getting, and can you confirm that is is caused by ld choking on the > > linker script? If not, this is another error than the one we have been > > trying to fix > > It's definitely not linker script related. > > Using "arm-linux-gnueabi-gcc (Ubuntu/Linaro 4.7.3-12ubuntu1) 4.7.3", > here's the error when building for multi_v7_defconfig (full log > available[2]): > > ../mm/migrate.c: In function 'migrate_pages': > ../mm/migrate.c:1148:1: internal compiler error: in push_minipool_fix, at config/arm/arm.c:13101 > Please submit a full bug report, > with preprocessed source if appropriate. > See <file:///usr/share/doc/gcc-4.7/README.Bugs> for instructions. > Preprocessed source stored into /tmp/ccO1Nz1m.out file, please attach > this to your bugreport. > make[2]: *** [mm/migrate.o] Error 1 > make[2]: Target `__build' not remade because of errors. > make[1]: *** [mm] Error 2 > > build bisect points to commit 21f992084aeb[3], but that doesn't revert > cleanly so I haven't got any further than that yet. I installed gcc-arm-linux-gnueabi (4:4.7.2-1 from Ubuntu 14.04 LTS) and could reproduce the ICE. I came up with the workaround below. Does this work for you? >From 7ebe83316eaf1952e55a76754ce7a5832e461b8c Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven <geert+renesas@glider.be> Date: Wed, 1 Apr 2015 11:22:51 +0200 Subject: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With gcc version 4.7.3 (Ubuntu/Linaro 4.7.3-12ubuntu1) : mm/migrate.c: In function ‘migrate_pages’: mm/migrate.c:1148:1: internal compiler error: in push_minipool_fix, at config/arm/arm.c:13500 Please submit a full bug report, with preprocessed source if appropriate. See <file:///usr/share/doc/gcc-4.7/README.Bugs> for instructions. Preprocessed source stored into /tmp/ccPoM1tr.out file, please attach this to your bugreport. make[1]: *** [mm/migrate.o] Error 1 make: *** [mm/migrate.o] Error 2 Mark unmap_and_move() (which is used in a single place only) "noinline" to work around this compiler bug. Reported-by: Kevin Hilman <khilman@kernel.org> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> --- mm/migrate.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/mm/migrate.c b/mm/migrate.c index 114602a68111d809..98f8574456c2010c 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -904,9 +904,10 @@ out: * Obtain the lock on page, remove all ptes and migrate the page * to the newly allocated page in newpage. */ -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, - unsigned long private, struct page *page, int force, - enum migrate_mode mode) +static noinline int unmap_and_move(new_page_t get_new_page, + free_page_t put_new_page, + unsigned long private, struct page *page, + int force, enum migrate_mode mode) { int rc = 0; int *result = NULL; -- 1.9.1 Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds ^ permalink raw reply related [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 9:37 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 (was: Re: Possible regression in gcc 4.7.3 next-20150323 due to "ARM, arm64: kvm: get rid of the bounce page") Geert Uytterhoeven @ 2015-04-01 9:47 ` Marc Zyngier 2015-04-01 19:40 ` Andrew Morton 2015-04-01 19:27 ` Kevin Hilman 1 sibling, 1 reply; 21+ messages in thread From: Marc Zyngier @ 2015-04-01 9:47 UTC (permalink / raw) To: Geert Uytterhoeven, Kevin Hilman Cc: Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Andrew Morton, Linux Kernel Development, linux-mm On 01/04/15 10:37, Geert Uytterhoeven wrote: > Hi Kevin, > > On Tue, 31 Mar 2015, Kevin Hilman wrote: >> Ard Biesheuvel <ard.biesheuvel@linaro.org> writes: >> Nope, that branch is already part of linux-next, and linux-next still >> fails to compile for 20+ defconfigs[1] >> >>> Could you elaborate on the issue please? What is the error you are >>> getting, and can you confirm that is is caused by ld choking on the >>> linker script? If not, this is another error than the one we have been >>> trying to fix >> >> It's definitely not linker script related. >> >> Using "arm-linux-gnueabi-gcc (Ubuntu/Linaro 4.7.3-12ubuntu1) 4.7.3", >> here's the error when building for multi_v7_defconfig (full log >> available[2]): >> >> ../mm/migrate.c: In function 'migrate_pages': >> ../mm/migrate.c:1148:1: internal compiler error: in push_minipool_fix, at config/arm/arm.c:13101 >> Please submit a full bug report, >> with preprocessed source if appropriate. >> See <file:///usr/share/doc/gcc-4.7/README.Bugs> for instructions. >> Preprocessed source stored into /tmp/ccO1Nz1m.out file, please attach >> this to your bugreport. >> make[2]: *** [mm/migrate.o] Error 1 >> make[2]: Target `__build' not remade because of errors. >> make[1]: *** [mm] Error 2 >> >> build bisect points to commit 21f992084aeb[3], but that doesn't revert >> cleanly so I haven't got any further than that yet. > > I installed gcc-arm-linux-gnueabi (4:4.7.2-1 from Ubuntu 14.04 LTS) and could > reproduce the ICE. I came up with the workaround below. > Does this work for you? > > From 7ebe83316eaf1952e55a76754ce7a5832e461b8c Mon Sep 17 00:00:00 2001 > From: Geert Uytterhoeven <geert+renesas@glider.be> > Date: Wed, 1 Apr 2015 11:22:51 +0200 > Subject: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in > gcc 4.7.3 > MIME-Version: 1.0 > Content-Type: text/plain; charset=UTF-8 > Content-Transfer-Encoding: 8bit > > With gcc version 4.7.3 (Ubuntu/Linaro 4.7.3-12ubuntu1) : > > mm/migrate.c: In function ‘migrate_pages’: > mm/migrate.c:1148:1: internal compiler error: in push_minipool_fix, at config/arm/arm.c:13500 > Please submit a full bug report, > with preprocessed source if appropriate. > See <file:///usr/share/doc/gcc-4.7/README.Bugs> for instructions. > Preprocessed source stored into /tmp/ccPoM1tr.out file, please attach this to your bugreport. > make[1]: *** [mm/migrate.o] Error 1 > make: *** [mm/migrate.o] Error 2 > > Mark unmap_and_move() (which is used in a single place only) "noinline" > to work around this compiler bug. > > Reported-by: Kevin Hilman <khilman@kernel.org> > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> > --- > mm/migrate.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/mm/migrate.c b/mm/migrate.c > index 114602a68111d809..98f8574456c2010c 100644 > --- a/mm/migrate.c > +++ b/mm/migrate.c > @@ -904,9 +904,10 @@ out: > * Obtain the lock on page, remove all ptes and migrate the page > * to the newly allocated page in newpage. > */ > -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, > - unsigned long private, struct page *page, int force, > - enum migrate_mode mode) > +static noinline int unmap_and_move(new_page_t get_new_page, > + free_page_t put_new_page, > + unsigned long private, struct page *page, > + int force, enum migrate_mode mode) > { > int rc = 0; > int *result = NULL; > Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of the parameters on the stack, which is not going to help performance either (not that this would be useful on 32bit ARM anyway...). Any chance you could make this dependent on some compiler detection mechanism? Thanks, M. -- Jazz is not dead. It just smells funny... ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 9:47 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 Marc Zyngier @ 2015-04-01 19:40 ` Andrew Morton 2015-04-01 21:54 ` Kevin Hilman 0 siblings, 1 reply; 21+ messages in thread From: Andrew Morton @ 2015-04-01 19:40 UTC (permalink / raw) To: Marc Zyngier Cc: Geert Uytterhoeven, Kevin Hilman, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm On Wed, 01 Apr 2015 10:47:49 +0100 Marc Zyngier <marc.zyngier@arm.com> wrote: > > -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, > > - unsigned long private, struct page *page, int force, > > - enum migrate_mode mode) > > +static noinline int unmap_and_move(new_page_t get_new_page, > > + free_page_t put_new_page, > > + unsigned long private, struct page *page, > > + int force, enum migrate_mode mode) > > { > > int rc = 0; > > int *result = NULL; > > > > Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of > the parameters on the stack, which is not going to help performance > either (not that this would be useful on 32bit ARM anyway...). > > Any chance you could make this dependent on some compiler detection > mechanism? With my arm compiler (gcc-4.4.4) the patch makes no difference - unmap_and_move() isn't being inlined anyway. How does this look? Kevin, could you please retest? I might have fat-fingered something... --- a/mm/migrate.c~mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473-fix +++ a/mm/migrate.c @@ -901,10 +901,20 @@ out: } /* + * gcc-4.7.3 on arm gets an ICE when inlining unmap_and_move(). Work around + * it. + */ +#if GCC_VERSION == 40703 && defined(CONFIG_ARM) +#define ICE_noinline noinline +#else +#define ICE_noinline +#endif + +/* * Obtain the lock on page, remove all ptes and migrate the page * to the newly allocated page in newpage. */ -static noinline int unmap_and_move(new_page_t get_new_page, +static ICE_noinline int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, unsigned long private, struct page *page, int force, enum migrate_mode mode) _ ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 19:40 ` Andrew Morton @ 2015-04-01 21:54 ` Kevin Hilman 2015-04-01 21:59 ` Russell King - ARM Linux ` (2 more replies) 0 siblings, 3 replies; 21+ messages in thread From: Kevin Hilman @ 2015-04-01 21:54 UTC (permalink / raw) To: Andrew Morton Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm Andrew Morton <akpm@linux-foundation.org> writes: > On Wed, 01 Apr 2015 10:47:49 +0100 Marc Zyngier <marc.zyngier@arm.com> wrote: > >> > -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, >> > - unsigned long private, struct page *page, int force, >> > - enum migrate_mode mode) >> > +static noinline int unmap_and_move(new_page_t get_new_page, >> > + free_page_t put_new_page, >> > + unsigned long private, struct page *page, >> > + int force, enum migrate_mode mode) >> > { >> > int rc = 0; >> > int *result = NULL; >> > >> >> Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of >> the parameters on the stack, which is not going to help performance >> either (not that this would be useful on 32bit ARM anyway...). >> >> Any chance you could make this dependent on some compiler detection >> mechanism? > > With my arm compiler (gcc-4.4.4) the patch makes no difference - > unmap_and_move() isn't being inlined anyway. > > How does this look? > > Kevin, could you please retest? I might have fat-fingered something... Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. However, I'm not sure how specific we can be on the versions. /me goes to test a few more compilers... OK... ICE: 4.7.1, 4.7.3, 4.8.3 OK: 4.6.3, 4.9.2, 4.9.3 The diff below[2] on top of yours compiles fine here and at least covers the compilers I *know* to trigger the ICE. Kevin [1] diff --git a/mm/migrate.c b/mm/migrate.c index 25fd7f6291de..6e15ae3248e0 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -901,10 +901,10 @@ out: } /* - * gcc-4.7.3 on arm gets an ICE when inlining unmap_and_move(). Work around + * gcc 4.7 and 4.8 on arm gets an ICE when inlining unmap_and_move(). Work around * it. */ -#if GCC_VERSION == 40703 && defined(CONFIG_ARM) +#if (GCC_VERSION >= 40700 && GCC_VERSION < 40900) && defined(CONFIG_ARM) #define ICE_noinline noinline #else #define ICE_noinline ^ permalink raw reply related [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 21:54 ` Kevin Hilman @ 2015-04-01 21:59 ` Russell King - ARM Linux 2015-04-02 7:17 ` Geert Uytterhoeven 2015-04-02 19:12 ` Lina Iyer 2015-04-07 17:57 ` Kevin Hilman 2 siblings, 1 reply; 21+ messages in thread From: Russell King - ARM Linux @ 2015-04-01 21:59 UTC (permalink / raw) To: Kevin Hilman Cc: Andrew Morton, Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm On Wed, Apr 01, 2015 at 02:54:59PM -0700, Kevin Hilman wrote: > Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. > However, I'm not sure how specific we can be on the versions. > > /me goes to test a few more compilers... OK... > > ICE: 4.7.1, 4.7.3, 4.8.3 > OK: 4.6.3, 4.9.2, 4.9.3 > > The diff below[2] on top of yours compiles fine here and at least covers > the compilers I *know* to trigger the ICE. Interesting. I'm using stock gcc 4.7.4 here, though I'm not building -next (only mainline + my tree + arm-soc) and it hasn't shown a problem yet. I think we need to ask the question: is the bug in stock GCC or Linaro GCC? If it's not in stock GCC, then it's a GCC vendor problem :) -- FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up according to speedtest.net. ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 21:59 ` Russell King - ARM Linux @ 2015-04-02 7:17 ` Geert Uytterhoeven 0 siblings, 0 replies; 21+ messages in thread From: Geert Uytterhoeven @ 2015-04-02 7:17 UTC (permalink / raw) To: Russell King - ARM Linux Cc: Kevin Hilman, Andrew Morton, Marc Zyngier, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm Hi Russell, On Wed, Apr 1, 2015 at 11:59 PM, Russell King - ARM Linux <linux@arm.linux.org.uk> wrote: > On Wed, Apr 01, 2015 at 02:54:59PM -0700, Kevin Hilman wrote: >> Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. >> However, I'm not sure how specific we can be on the versions. >> >> /me goes to test a few more compilers... OK... >> >> ICE: 4.7.1, 4.7.3, 4.8.3 >> OK: 4.6.3, 4.9.2, 4.9.3 >> >> The diff below[2] on top of yours compiles fine here and at least covers >> the compilers I *know* to trigger the ICE. > > Interesting. I'm using stock gcc 4.7.4 here, though I'm not building > -next (only mainline + my tree + arm-soc) and it hasn't shown a problem > yet. Mainline doesn't fail. > I think we need to ask the question: is the bug in stock GCC or Linaro > GCC? If it's not in stock GCC, then it's a GCC vendor problem :) Can you please try -next (e.g. next-20150320)? make bockw_defconfig make mm/migrate.o Thanks! Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 21:54 ` Kevin Hilman 2015-04-01 21:59 ` Russell King - ARM Linux @ 2015-04-02 19:12 ` Lina Iyer 2015-04-02 21:12 ` Kevin Hilman 2015-04-07 17:57 ` Kevin Hilman 2 siblings, 1 reply; 21+ messages in thread From: Lina Iyer @ 2015-04-02 19:12 UTC (permalink / raw) To: Kevin Hilman Cc: Andrew Morton, Nishanth Menon, Magnus Damm, Russell King - ARM Linux, Arnd Bergmann, Ard Biesheuvel, Marc Zyngier, Tyler Baker, linux-sh, Will Deacon, Linux Kernel Development, linux-mm, Simon Horman, Geert Uytterhoeven, Catalin Marinas, grygorii.strashko, linux-omap, linux-arm-kernel On Wed, Apr 01 2015 at 15:57 -0600, Kevin Hilman wrote: >Andrew Morton <akpm@linux-foundation.org> writes: > >> On Wed, 01 Apr 2015 10:47:49 +0100 Marc Zyngier <marc.zyngier@arm.com> wrote: >> >>> > -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, >>> > - unsigned long private, struct page *page, int force, >>> > - enum migrate_mode mode) >>> > +static noinline int unmap_and_move(new_page_t get_new_page, >>> > + free_page_t put_new_page, >>> > + unsigned long private, struct page *page, >>> > + int force, enum migrate_mode mode) >>> > { >>> > int rc = 0; >>> > int *result = NULL; >>> > >>> >>> Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of >>> the parameters on the stack, which is not going to help performance >>> either (not that this would be useful on 32bit ARM anyway...). >>> >>> Any chance you could make this dependent on some compiler detection >>> mechanism? >> >> With my arm compiler (gcc-4.4.4) the patch makes no difference - >> unmap_and_move() isn't being inlined anyway. >> >> How does this look? >> >> Kevin, could you please retest? I might have fat-fingered something... > >Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. >However, I'm not sure how specific we can be on the versions. > >/me goes to test a few more compilers... OK... > >ICE: 4.7.1, 4.7.3, 4.8.3 >OK: 4.6.3, 4.9.2, 4.9.3 > >The diff below[2] on top of yours compiles fine here and at least covers >the compilers I *know* to trigger the ICE. I see ICE on arm-linux-gnueabi-gcc (Ubuntu/Linaro 4.7.4-2ubuntu1) 4.7.4 > >Kevin > > >[1] >diff --git a/mm/migrate.c b/mm/migrate.c >index 25fd7f6291de..6e15ae3248e0 100644 >--- a/mm/migrate.c >+++ b/mm/migrate.c >@@ -901,10 +901,10 @@ out: > } > > /* >- * gcc-4.7.3 on arm gets an ICE when inlining unmap_and_move(). Work around >+ * gcc 4.7 and 4.8 on arm gets an ICE when inlining unmap_and_move(). Work around > * it. > */ >-#if GCC_VERSION == 40703 && defined(CONFIG_ARM) >+#if (GCC_VERSION >= 40700 && GCC_VERSION < 40900) && defined(CONFIG_ARM) > #define ICE_noinline noinline > #else > #define ICE_noinline > >_______________________________________________ >linux-arm-kernel mailing list >linux-arm-kernel@lists.infradead.org >http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-02 19:12 ` Lina Iyer @ 2015-04-02 21:12 ` Kevin Hilman 2015-04-02 21:53 ` Lina Iyer 0 siblings, 1 reply; 21+ messages in thread From: Kevin Hilman @ 2015-04-02 21:12 UTC (permalink / raw) To: Lina Iyer Cc: Kevin Hilman, Andrew Morton, Nishanth Menon, Magnus Damm, Russell King - ARM Linux, Arnd Bergmann, Ard Biesheuvel, Marc Zyngier, Tyler Baker, linux-sh, Will Deacon, Linux Kernel Development, linux-mm, Simon Horman, Geert Uytterhoeven, Catalin Marinas, grygorii.strashko, linux-omap, linux-arm-kernel On Thu, Apr 2, 2015 at 12:12 PM, Lina Iyer <lina.iyer@linaro.org> wrote: > On Wed, Apr 01 2015 at 15:57 -0600, Kevin Hilman wrote: >> >> Andrew Morton <akpm@linux-foundation.org> writes: >> >>> On Wed, 01 Apr 2015 10:47:49 +0100 Marc Zyngier <marc.zyngier@arm.com> >>> wrote: >>> >>>> > -static int unmap_and_move(new_page_t get_new_page, free_page_t >>>> > put_new_page, >>>> > - unsigned long private, struct page *page, int >>>> > force, >>>> > - enum migrate_mode mode) >>>> > +static noinline int unmap_and_move(new_page_t get_new_page, >>>> > + free_page_t put_new_page, >>>> > + unsigned long private, struct page >>>> > *page, >>>> > + int force, enum migrate_mode mode) >>>> > { >>>> > int rc = 0; >>>> > int *result = NULL; >>>> > >>>> >>>> Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of >>>> the parameters on the stack, which is not going to help performance >>>> either (not that this would be useful on 32bit ARM anyway...). >>>> >>>> Any chance you could make this dependent on some compiler detection >>>> mechanism? >>> >>> >>> With my arm compiler (gcc-4.4.4) the patch makes no difference - >>> unmap_and_move() isn't being inlined anyway. >>> >>> How does this look? >>> >>> Kevin, could you please retest? I might have fat-fingered something... >> >> >> Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. >> However, I'm not sure how specific we can be on the versions. >> >> /me goes to test a few more compilers... OK... >> >> ICE: 4.7.1, 4.7.3, 4.8.3 >> OK: 4.6.3, 4.9.2, 4.9.3 >> >> The diff below[2] on top of yours compiles fine here and at least covers >> the compilers I *know* to trigger the ICE. > > > I see ICE on arm-linux-gnueabi-gcc (Ubuntu/Linaro 4.7.4-2ubuntu1) 4.7.4 > Thanks for checking. I'm assuming my patch fixes it for your since that should catch any 4.7.x compiler. Kevin ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-02 21:12 ` Kevin Hilman @ 2015-04-02 21:53 ` Lina Iyer 0 siblings, 0 replies; 21+ messages in thread From: Lina Iyer @ 2015-04-02 21:53 UTC (permalink / raw) To: Kevin Hilman Cc: Andrew Morton, Nishanth Menon, Magnus Damm, Russell King - ARM Linux, Arnd Bergmann, Ard Biesheuvel, Marc Zyngier, Tyler Baker, linux-sh, Will Deacon, Linux Kernel Development, linux-mm, Simon Horman, Geert Uytterhoeven, Catalin Marinas, grygorii.strashko, linux-omap, linux-arm-kernel On Thu, Apr 02 2015 at 15:12 -0600, Kevin Hilman wrote: >On Thu, Apr 2, 2015 at 12:12 PM, Lina Iyer <lina.iyer@linaro.org> wrote: >> On Wed, Apr 01 2015 at 15:57 -0600, Kevin Hilman wrote: >>> >>> Andrew Morton <akpm@linux-foundation.org> writes: >>> >>>> On Wed, 01 Apr 2015 10:47:49 +0100 Marc Zyngier <marc.zyngier@arm.com> >>>> wrote: >>>> >>>>> > -static int unmap_and_move(new_page_t get_new_page, free_page_t >>>>> > put_new_page, >>>>> > - unsigned long private, struct page *page, int >>>>> > force, >>>>> > - enum migrate_mode mode) >>>>> > +static noinline int unmap_and_move(new_page_t get_new_page, >>>>> > + free_page_t put_new_page, >>>>> > + unsigned long private, struct page >>>>> > *page, >>>>> > + int force, enum migrate_mode mode) >>>>> > { >>>>> > int rc = 0; >>>>> > int *result = NULL; >>>>> > >>>>> >>>>> Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of >>>>> the parameters on the stack, which is not going to help performance >>>>> either (not that this would be useful on 32bit ARM anyway...). >>>>> >>>>> Any chance you could make this dependent on some compiler detection >>>>> mechanism? >>>> >>>> >>>> With my arm compiler (gcc-4.4.4) the patch makes no difference - >>>> unmap_and_move() isn't being inlined anyway. >>>> >>>> How does this look? >>>> >>>> Kevin, could you please retest? I might have fat-fingered something... >>> >>> >>> Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. >>> However, I'm not sure how specific we can be on the versions. >>> >>> /me goes to test a few more compilers... OK... >>> >>> ICE: 4.7.1, 4.7.3, 4.8.3 >>> OK: 4.6.3, 4.9.2, 4.9.3 >>> >>> The diff below[2] on top of yours compiles fine here and at least covers >>> the compilers I *know* to trigger the ICE. >> >> >> I see ICE on arm-linux-gnueabi-gcc (Ubuntu/Linaro 4.7.4-2ubuntu1) 4.7.4 >> > >Thanks for checking. I'm assuming my patch fixes it for your since >that should catch any 4.7.x compiler. Yes, thank you. This fixes it on 4.7.4 > >Kevin ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 21:54 ` Kevin Hilman 2015-04-01 21:59 ` Russell King - ARM Linux 2015-04-02 19:12 ` Lina Iyer @ 2015-04-07 17:57 ` Kevin Hilman 2015-04-07 20:17 ` Andrew Morton 2 siblings, 1 reply; 21+ messages in thread From: Kevin Hilman @ 2015-04-07 17:57 UTC (permalink / raw) To: Kevin Hilman Cc: Andrew Morton, Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm Hi Andrew, On Wed, Apr 1, 2015 at 2:54 PM, Kevin Hilman <khilman@kernel.org> wrote: > Andrew Morton <akpm@linux-foundation.org> writes: > >> On Wed, 01 Apr 2015 10:47:49 +0100 Marc Zyngier <marc.zyngier@arm.com> wrote: >> >>> > -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, >>> > - unsigned long private, struct page *page, int force, >>> > - enum migrate_mode mode) >>> > +static noinline int unmap_and_move(new_page_t get_new_page, >>> > + free_page_t put_new_page, >>> > + unsigned long private, struct page *page, >>> > + int force, enum migrate_mode mode) >>> > { >>> > int rc = 0; >>> > int *result = NULL; >>> > >>> >>> Ouch. That's really ugly. And on 32bit ARM, we end-up spilling half of >>> the parameters on the stack, which is not going to help performance >>> either (not that this would be useful on 32bit ARM anyway...). >>> >>> Any chance you could make this dependent on some compiler detection >>> mechanism? >> >> With my arm compiler (gcc-4.4.4) the patch makes no difference - >> unmap_and_move() isn't being inlined anyway. >> >> How does this look? >> >> Kevin, could you please retest? I might have fat-fingered something... > > Your patch on top of Geert's still compiles fine for me with gcc-4.7.3. > However, I'm not sure how specific we can be on the versions. > > /me goes to test a few more compilers... OK... > > ICE: 4.7.1, 4.7.3, 4.8.3 > OK: 4.6.3, 4.9.2, 4.9.3 > > The diff below[2] on top of yours compiles fine here and at least covers > the compilers I *know* to trigger the ICE. I see my fix in your mmots since last Thurs (4/2), but it's not in mmotm (last updated today) so today's linux-next still has the ICE for anything other than gcc-4.7.3. Just checking to see when you plan to update mmotm. Thanks, Kevin ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-07 17:57 ` Kevin Hilman @ 2015-04-07 20:17 ` Andrew Morton 2015-04-07 22:41 ` Kevin Hilman 0 siblings, 1 reply; 21+ messages in thread From: Andrew Morton @ 2015-04-07 20:17 UTC (permalink / raw) To: Kevin Hilman Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm On Tue, 7 Apr 2015 10:57:52 -0700 Kevin Hilman <khilman@kernel.org> wrote: > > The diff below[2] on top of yours compiles fine here and at least covers > > the compilers I *know* to trigger the ICE. > > I see my fix in your mmots since last Thurs (4/2), but it's not in > mmotm (last updated today) so today's linux-next still has the ICE for > anything other than gcc-4.7.3. Just checking to see when you plan to > update mmotm. It should all be there today? ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-07 20:17 ` Andrew Morton @ 2015-04-07 22:41 ` Kevin Hilman 2015-04-07 22:53 ` Andrew Morton 0 siblings, 1 reply; 21+ messages in thread From: Kevin Hilman @ 2015-04-07 22:41 UTC (permalink / raw) To: Andrew Morton Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm Andrew Morton <akpm@linux-foundation.org> writes: > On Tue, 7 Apr 2015 10:57:52 -0700 Kevin Hilman <khilman@kernel.org> wrote: > >> > The diff below[2] on top of yours compiles fine here and at least covers >> > the compilers I *know* to trigger the ICE. >> >> I see my fix in your mmots since last Thurs (4/2), but it's not in >> mmotm (last updated today) so today's linux-next still has the ICE for >> anything other than gcc-4.7.3. Just checking to see when you plan to >> update mmotm. > > It should all be there today? Nope. In mmotm, only the original patch plus your first fix is there: $ curl -sO http://www.ozlabs.org/~akpm/mmotm/broken-out.tar.gz $ tar -tavf broken-out.tar.gz |grep gcc-473 -rw-r----- akpm/eng 1838 2015-04-01 14:41 broken-out/mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473.patch -rw-r----- akpm/eng 1309 2015-04-01 14:41 broken-out/mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473-fix.patch but in mmots, the additional ptch from me, plus another comment fixup from you are also there: $ curl -sO http://www.ozlabs.org/~akpm/mmots/broken-out.tar.gz $ tar -tavf broken-out.tar.gz |grep gcc-473 -rw-r----- akpm/eng 1882 2015-04-06 16:24 broken-out/mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473.patch -rw-r----- akpm/eng 1271 2015-04-06 16:24 broken-out/mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473-fix.patch -rw-r----- akpm/eng 1382 2015-04-06 16:24 broken-out/mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473-fix-fix.patch -rw-r----- akpm/eng 968 2015-04-06 16:24 broken-out/mm-migrate-mark-unmap_and_move-noinline-to-avoid-ice-in-gcc-473-fix-fix-fix.patch Kevin ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-07 22:41 ` Kevin Hilman @ 2015-04-07 22:53 ` Andrew Morton 2015-04-07 23:27 ` Kevin Hilman 0 siblings, 1 reply; 21+ messages in thread From: Andrew Morton @ 2015-04-07 22:53 UTC (permalink / raw) To: Kevin Hilman Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm On Tue, 07 Apr 2015 15:41:32 -0700 Kevin Hilman <khilman@kernel.org> wrote: > Andrew Morton <akpm@linux-foundation.org> writes: > > > On Tue, 7 Apr 2015 10:57:52 -0700 Kevin Hilman <khilman@kernel.org> wrote: > > > >> > The diff below[2] on top of yours compiles fine here and at least covers > >> > the compilers I *know* to trigger the ICE. > >> > >> I see my fix in your mmots since last Thurs (4/2), but it's not in > >> mmotm (last updated today) so today's linux-next still has the ICE for > >> anything other than gcc-4.7.3. Just checking to see when you plan to > >> update mmotm. > > > > It should all be there today? > > Nope. huh, I swear I did an mmotm yesterday. Let me see if I can sort out the watchdog mess and produce something releasable... ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-07 22:53 ` Andrew Morton @ 2015-04-07 23:27 ` Kevin Hilman 2015-04-07 23:36 ` Andrew Morton 0 siblings, 1 reply; 21+ messages in thread From: Kevin Hilman @ 2015-04-07 23:27 UTC (permalink / raw) To: Andrew Morton Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm Andrew Morton <akpm@linux-foundation.org> writes: > On Tue, 07 Apr 2015 15:41:32 -0700 Kevin Hilman <khilman@kernel.org> wrote: > >> Andrew Morton <akpm@linux-foundation.org> writes: >> >> > On Tue, 7 Apr 2015 10:57:52 -0700 Kevin Hilman <khilman@kernel.org> wrote: >> > >> >> > The diff below[2] on top of yours compiles fine here and at least covers >> >> > the compilers I *know* to trigger the ICE. >> >> >> >> I see my fix in your mmots since last Thurs (4/2), but it's not in >> >> mmotm (last updated today) so today's linux-next still has the ICE for >> >> anything other than gcc-4.7.3. Just checking to see when you plan to >> >> update mmotm. >> > >> > It should all be there today? >> >> Nope. > > huh, I swear I did an mmotm yesterday. Well, based on the timestamp of the mmotm dir on ozlabs, it appears you did. That's why I was confused why the gcc-473 patches from mmots aren't there. > Let me see if I can sort out the watchdog mess and produce something > releasable... OK, thanks. Kevin ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-07 23:27 ` Kevin Hilman @ 2015-04-07 23:36 ` Andrew Morton 2015-04-08 0:15 ` Kevin Hilman 0 siblings, 1 reply; 21+ messages in thread From: Andrew Morton @ 2015-04-07 23:36 UTC (permalink / raw) To: Kevin Hilman Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm On Tue, 07 Apr 2015 16:27:44 -0700 Kevin Hilman <khilman@kernel.org> wrote: > >> > It should all be there today? > >> > >> Nope. > > > > huh, I swear I did an mmotm yesterday. > > Well, based on the timestamp of the mmotm dir on ozlabs, it appears you > did. That's why I was confused why the gcc-473 patches from mmots aren't > there. Things look a bit better now. ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-07 23:36 ` Andrew Morton @ 2015-04-08 0:15 ` Kevin Hilman 0 siblings, 0 replies; 21+ messages in thread From: Kevin Hilman @ 2015-04-08 0:15 UTC (permalink / raw) To: Andrew Morton Cc: Marc Zyngier, Geert Uytterhoeven, Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Linux Kernel Development, linux-mm Andrew Morton <akpm@linux-foundation.org> writes: > On Tue, 07 Apr 2015 16:27:44 -0700 Kevin Hilman <khilman@kernel.org> wrote: > >> >> > It should all be there today? >> >> >> >> Nope. >> > >> > huh, I swear I did an mmotm yesterday. >> >> Well, based on the timestamp of the mmotm dir on ozlabs, it appears you >> did. That's why I was confused why the gcc-473 patches from mmots aren't >> there. > > Things look a bit better now. Yup, I can confirm all 4 patches are there now. Things should be in good shape for the next -next. Thanks, Kevin ^ permalink raw reply [flat|nested] 21+ messages in thread
* Re: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 2015-04-01 9:37 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 (was: Re: Possible regression in gcc 4.7.3 next-20150323 due to "ARM, arm64: kvm: get rid of the bounce page") Geert Uytterhoeven 2015-04-01 9:47 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 Marc Zyngier @ 2015-04-01 19:27 ` Kevin Hilman 1 sibling, 0 replies; 21+ messages in thread From: Kevin Hilman @ 2015-04-01 19:27 UTC (permalink / raw) To: Geert Uytterhoeven Cc: Ard Biesheuvel, Will Deacon, Simon Horman, Tyler Baker, Nishanth Menon, Russell King - ARM Linux, Arnd Bergmann, linux-sh, Marc Zyngier, Catalin Marinas, Magnus Damm, grygorii.strashko, linux-omap, linux-arm-kernel, Andrew Morton, Linux Kernel Development, linux-mm Geert Uytterhoeven <geert@linux-m68k.org> writes: [...] >> build bisect points to commit 21f992084aeb[3], but that doesn't revert >> cleanly so I haven't got any further than that yet. > > I installed gcc-arm-linux-gnueabi (4:4.7.2-1 from Ubuntu 14.04 LTS) and could > reproduce the ICE. I came up with the workaround below. Awesome, thanks! > Does this work for you? Yes, that patch works well and fixes the regression. Build results for all the defconfigs here: http://kernelci.org/build/khilman/kernel/v4.0-rc6-8294-g2ef3958cc27e/ and the remaining issues arent' realted to this ICE. > From 7ebe83316eaf1952e55a76754ce7a5832e461b8c Mon Sep 17 00:00:00 2001 > From: Geert Uytterhoeven <geert+renesas@glider.be> > Date: Wed, 1 Apr 2015 11:22:51 +0200 > Subject: [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in > gcc 4.7.3 > MIME-Version: 1.0 > Content-Type: text/plain; charset=UTF-8 > Content-Transfer-Encoding: 8bit > > With gcc version 4.7.3 (Ubuntu/Linaro 4.7.3-12ubuntu1) : > > mm/migrate.c: In function ‘migrate_pages’: > mm/migrate.c:1148:1: internal compiler error: in push_minipool_fix, at config/arm/arm.c:13500 > Please submit a full bug report, > with preprocessed source if appropriate. > See <file:///usr/share/doc/gcc-4.7/README.Bugs> for instructions. > Preprocessed source stored into /tmp/ccPoM1tr.out file, please attach this to your bugreport. > make[1]: *** [mm/migrate.o] Error 1 > make: *** [mm/migrate.o] Error 2 > > Mark unmap_and_move() (which is used in a single place only) "noinline" > to work around this compiler bug. > > Reported-by: Kevin Hilman <khilman@kernel.org> > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Tested-by: Kevin Hilman <khilman@linaro.org> > --- > mm/migrate.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/mm/migrate.c b/mm/migrate.c > index 114602a68111d809..98f8574456c2010c 100644 > --- a/mm/migrate.c > +++ b/mm/migrate.c > @@ -904,9 +904,10 @@ out: > * Obtain the lock on page, remove all ptes and migrate the page > * to the newly allocated page in newpage. > */ > -static int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, > - unsigned long private, struct page *page, int force, > - enum migrate_mode mode) > +static noinline int unmap_and_move(new_page_t get_new_page, > + free_page_t put_new_page, > + unsigned long private, struct page *page, > + int force, enum migrate_mode mode) > { > int rc = 0; > int *result = NULL; ^ permalink raw reply [flat|nested] 21+ messages in thread
end of thread, other threads:[~2015-04-08 0:16 UTC | newest] Thread overview: 21+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- [not found] <20150324004537.GA24816@verge.net.au> [not found] ` <CAKv+Gu-0jPk=KQ4gY32ELc+BVbe=1QdcrwQ+Pb=RkdwO9K3Vkw@mail.gmail.com> [not found] ` <20150324161358.GA694@kahuna> 2015-03-24 18:42 ` [PATCH] gpio: syscon: reduce message level when direction reg offset not in dt grygorii.strashko 2015-03-24 20:00 ` Murali Karicheri 2015-03-24 20:52 ` santosh shilimkar 2015-03-27 10:18 ` Linus Walleij [not found] ` <20150326003939.GA25368@verge.net.au> [not found] ` <20150326133631.GB2805@arm.com> [not found] ` <CANMBJr68dsbYvvHUzy6U4m4fEM6nq8dVHBH4kLQ=0c4QNOhLPQ@mail.gmail.com> [not found] ` <20150327002554.GA5527@verge.net.au> [not found] ` <20150327100612.GB1562@arm.com> [not found] ` <7hbnj99epe.fsf@deeprootsystems.com> [not found] ` <CAKv+Gu_ZHZFm-1eXn+r7fkEHOxqSmj+Q+Mmy7k6LK531vSfAjQ@mail.gmail.com> [not found] ` <7h8uec95t2.fsf@deeprootsystems.com> 2015-04-01 9:37 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 (was: Re: Possible regression in gcc 4.7.3 next-20150323 due to "ARM, arm64: kvm: get rid of the bounce page") Geert Uytterhoeven 2015-04-01 9:47 ` [PATCH] mm/migrate: Mark unmap_and_move() "noinline" to avoid ICE in gcc 4.7.3 Marc Zyngier 2015-04-01 19:40 ` Andrew Morton 2015-04-01 21:54 ` Kevin Hilman 2015-04-01 21:59 ` Russell King - ARM Linux 2015-04-02 7:17 ` Geert Uytterhoeven 2015-04-02 19:12 ` Lina Iyer 2015-04-02 21:12 ` Kevin Hilman 2015-04-02 21:53 ` Lina Iyer 2015-04-07 17:57 ` Kevin Hilman 2015-04-07 20:17 ` Andrew Morton 2015-04-07 22:41 ` Kevin Hilman 2015-04-07 22:53 ` Andrew Morton 2015-04-07 23:27 ` Kevin Hilman 2015-04-07 23:36 ` Andrew Morton 2015-04-08 0:15 ` Kevin Hilman 2015-04-01 19:27 ` Kevin Hilman
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).