linux-kernel - Re: intel-ixp42x-welltech-epbx100 no longer boot due to 463dbba4d189 ("ARM: 9104/2: Fix Keystone 2 kernel mapping regression")

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <YVL+zd33DwmLXo6B@Red>
Date:   Tue, 28 Sep 2021 13:38:53 +0200
From:   Corentin Labbe <clabbe.montjoie@...il.com>
To:     Krzysztof Hałasa <khalasa@...p.pl>
Cc:     linusw@...nel.org, kaloz@...nwrt.org, linux@...linux.org.uk,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: intel-ixp42x-welltech-epbx100 no longer boot due to 463dbba4d189
 ("ARM: 9104/2: Fix Keystone 2 kernel mapping regression")

Le Tue, Sep 28, 2021 at 12:32:13PM +0200, Krzysztof Hałasa a écrit :
> Hello Corentin,
> 
> Corentin Labbe <clabbe.montjoie@...il.com> writes:
> 
> > I am working on puting my ixp4xx board on kernelCI.
> > But it no longer boot since 4.14.
> > It only prints "Starting kernelUncompressing Linux... done, booting the kernel." then nothing.
> >
> > I bisected the break to 463dbba4d189 ("ARM: 9104/2: Fix Keystone 2 kernel mapping regression")
> 
> I guess you build your kernel as big-endian.

Hello

Yes, IXP4XX is by default BE.

> 
> These parts seem suspicious to me:
> 
> -    str r8, [r5]            @ Save physical start of kernel
> +#ifdef CONFIG_CPU_ENDIAN_BE8
> +    str r8, [r5, #4]            @ Save physical start of kernel (BE)
> +#else
> +    str r8, [r5]            @ Save physical start of kernel (LE)
> +#endif
> 
> and
> 
> -    str r3, [r5]            @ Save physical end of kernel
> +#ifdef CONFIG_CPU_ENDIAN_BE8
> +    str r3, [r5, #4]            @ Save physical end of kernel (BE)
> +#else
> +    str r3, [r5]            @ Save physical end of kernel (LE)
> +#endif
> 
> IXP4xx, being ARMv5, uses CONFIG_CPU_ENDIAN_BE32 (when configured as
> BE), but it needs the ", #4" as well.
> 
> I'd test with: #if defined CONFIG_CPU_ENDIAN_BE8 || defined CONFIG_CPU_ENDIAN_BE32

Thanks, with the following patch, my boards boots

diff --git a/arch/arm/kernel/head.S b/arch/arm/kernel/head.S
index 29070eb8df7d..3fc7f9750ce4 100644
--- a/arch/arm/kernel/head.S
+++ b/arch/arm/kernel/head.S
@@ -253,7 +253,7 @@ __create_page_tables:
        add     r0, r4, #KERNEL_OFFSET >> (SECTION_SHIFT - PMD_ORDER)
        ldr     r6, =(_end - 1)
        adr_l   r5, kernel_sec_start            @ _pa(kernel_sec_start)
-#ifdef CONFIG_CPU_ENDIAN_BE8
+#if defined CONFIG_CPU_ENDIAN_BE8 || defined CONFIG_CPU_ENDIAN_BE32
        str     r8, [r5, #4]                    @ Save physical start of kernel (BE)
 #else
        str     r8, [r5]                        @ Save physical start of kernel (LE)
@@ -266,7 +266,7 @@ __create_page_tables:
        bls     1b
        eor     r3, r3, r7                      @ Remove the MMU flags
        adr_l   r5, kernel_sec_end              @ _pa(kernel_sec_end)
-#ifdef CONFIG_CPU_ENDIAN_BE8
+#if defined CONFIG_CPU_ENDIAN_BE8 || defined CONFIG_CPU_ENDIAN_BE32
        str     r3, [r5, #4]                    @ Save physical end of kernel (BE)
 #else
        str     r3, [r5]                        @ Save physical end of kernel (LE)

Regards