incomplete mips patch made 3.10.55, remains broken in 3.10.58

incomplete mips patch made 3.10.55, remains broken in 3.10.58

[Alexandre Oliva]

[-- Attachment #1: Type: text/plain, Size: 719 bytes --]

Greg,

Commit ff522058bd717506b2fa066fa564657f2b86477e was merged into 3.10.55
stable as commit 4f91cb537d2f7fa700a2b6d86a2cc77d20ee2616.

Without the complement, commit 5596b0b245fb9d2cefb5023b11061050351c1398,
included below, cache invalidation functions modified by the former
patch may return between preempt_disable() and preempt_enable(), causing
such machines as yeeloongs to go down in flames early in the boot.

The complement patch had already made v3.12-rc4, and it's quite
obviously needed and correct.  I've also tested that it fixes the
regression on the yeeloong.

So, would you please merge it into the 3.10 stable series, at your
earlier convenience, so as to fix this regression?

Thanks in advance,

Re: incomplete mips patch made 3.10.55, remains broken in 3.10.58

[Ralf Baechle]

Greg, please apply.

Thanks,

  Ralf

On Thu, Oct 16, 2014 at 03:57:33AM -0300, Alexandre Oliva wrote:

> Commit ff522058bd717506b2fa066fa564657f2b86477e was merged into 3.10.55
> stable as commit 4f91cb537d2f7fa700a2b6d86a2cc77d20ee2616.
> 
> Without the complement, commit 5596b0b245fb9d2cefb5023b11061050351c1398,
> included below, cache invalidation functions modified by the former
> patch may return between preempt_disable() and preempt_enable(), causing
> such machines as yeeloongs to go down in flames early in the boot.
> 
> The complement patch had already made v3.12-rc4, and it's quite
> obviously needed and correct.  I've also tested that it fixes the
> regression on the yeeloong.
> 
> So, would you please merge it into the 3.10 stable series, at your
> earlier convenience, so as to fix this regression?
> 
> Thanks in advance,
> 

> >From 5596b0b245fb9d2cefb5023b11061050351c1398 Mon Sep 17 00:00:00 2001
> From: Yoichi Yuasa <yuasa@linux-mips.org>
> Date: Wed, 2 Oct 2013 15:03:03 +0900
> Subject: [PATCH] MIPS: Fix forgotten preempt_enable() when CPU has inclusive
>  pcaches
> 
> [    1.904000] BUG: scheduling while atomic: swapper/1/0x00000002
> [    1.908000] Modules linked in:
> [    1.916000] CPU: 0 PID: 1 Comm: swapper Not tainted 3.12.0-rc2-lemote-los.git-5318619-dirty #1
> [    1.920000] Stack : 0000000031aac000 ffffffff810d0000 0000000000000052 ffffffff802730a4
>           0000000000000000 0000000000000001 ffffffff810cdf90 ffffffff810d0000
>           ffffffff8068b968 ffffffff806f5537 ffffffff810cdf90 980000009f0782e8
>           0000000000000001 ffffffff80720000 ffffffff806b0000 980000009f078000
>           980000009f290000 ffffffff805f312c 980000009f05b5d8 ffffffff80233518
>           980000009f05b5e8 ffffffff80274b7c 980000009f078000 ffffffff8068b968
>           0000000000000000 0000000000000000 0000000000000000 0000000000000000
>           0000000000000000 980000009f05b520 0000000000000000 ffffffff805f2f6c
>           0000000000000000 ffffffff80700000 ffffffff80700000 ffffffff806fc758
>           ffffffff80700000 ffffffff8020be98 ffffffff806fceb0 ffffffff805f2f6c
>           ...
> [    2.028000] Call Trace:
> [    2.032000] [<ffffffff8020be98>] show_stack+0x80/0x98
> [    2.036000] [<ffffffff805f2f6c>] __schedule_bug+0x44/0x6c
> [    2.040000] [<ffffffff805fac58>] __schedule+0x518/0x5b0
> [    2.044000] [<ffffffff805f8a58>] schedule_timeout+0x128/0x1f0
> [    2.048000] [<ffffffff80240314>] msleep+0x3c/0x60
> [    2.052000] [<ffffffff80495400>] do_probe+0x238/0x3a8
> [    2.056000] [<ffffffff804958b0>] ide_probe_port+0x340/0x7e8
> [    2.060000] [<ffffffff80496028>] ide_host_register+0x2d0/0x7a8
> [    2.064000] [<ffffffff8049c65c>] ide_pci_init_two+0x4e4/0x790
> [    2.068000] [<ffffffff8049f9b8>] amd74xx_probe+0x148/0x2c8
> [    2.072000] [<ffffffff803f571c>] pci_device_probe+0xc4/0x130
> [    2.076000] [<ffffffff80478f60>] driver_probe_device+0x98/0x270
> [    2.080000] [<ffffffff80479298>] __driver_attach+0xe0/0xe8
> [    2.084000] [<ffffffff80476ab0>] bus_for_each_dev+0x78/0xe0
> [    2.088000] [<ffffffff80478468>] bus_add_driver+0x230/0x310
> [    2.092000] [<ffffffff80479b44>] driver_register+0x84/0x158
> [    2.096000] [<ffffffff80200504>] do_one_initcall+0x104/0x160
> 
> Signed-off-by: Yoichi Yuasa <yuasa@linux-mips.org>
> Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi>
> Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi>
> Cc: linux-mips@linux-mips.org
> Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
> Patchwork: https://patchwork.linux-mips.org/patch/5941/
> Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
> ---
>  arch/mips/mm/c-r4k.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c
> index 627883b..bc6f96f 100644
> --- a/arch/mips/mm/c-r4k.c
> +++ b/arch/mips/mm/c-r4k.c
> @@ -609,6 +609,7 @@ static void r4k_dma_cache_wback_inv(unsigned long addr, unsigned long size)
>  			r4k_blast_scache();
>  		else
>  			blast_scache_range(addr, addr + size);
> +		preempt_enable();
>  		__sync();
>  		return;
>  	}
> @@ -650,6 +651,7 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size)
>  			 */
>  			blast_inv_scache_range(addr, addr + size);
>  		}
> +		preempt_enable();
>  		__sync();
>  		return;
>  	}
> -- 
> 1.9.3
> 

> 
> -- 
> Alexandre Oliva, freedom fighter    http://FSFLA.org/~lxoliva/
> You must be the change you wish to see in the world. -- Gandhi
> Be Free! -- http://FSFLA.org/   FSF Latin America board member
> Free Software Evangelist|Red Hat Brasil GNU Toolchain Engineer


  Ralf

Re: incomplete mips patch made 3.10.55, remains broken in 3.10.58

[Greg Kroah-Hartman]

On Thu, Oct 16, 2014 at 03:57:33AM -0300, Alexandre Oliva wrote:
> Greg,
> 
> Commit ff522058bd717506b2fa066fa564657f2b86477e was merged into 3.10.55
> stable as commit 4f91cb537d2f7fa700a2b6d86a2cc77d20ee2616.
> 
> Without the complement, commit 5596b0b245fb9d2cefb5023b11061050351c1398,
> included below, cache invalidation functions modified by the former
> patch may return between preempt_disable() and preempt_enable(), causing
> such machines as yeeloongs to go down in flames early in the boot.
> 
> The complement patch had already made v3.12-rc4, and it's quite
> obviously needed and correct.  I've also tested that it fixes the
> regression on the yeeloong.
> 
> So, would you please merge it into the 3.10 stable series, at your
> earlier convenience, so as to fix this regression?

Applied, now, thanks.

greg k-h