Date:   Fri, 10 Apr 2020 16:16:58 +0800
From:   Like Xu <like.xu@...ux.intel.com>
To:     Mikhail Gavrilov <mikhail.v.gavrilov@...il.com>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        linux-x86_64@...r.kernel.org, nivedita@...m.mit.edu
Subject: Re: [Bug report] Kernel 5.7 become unbootable

On 2020/4/10 3:12, Mikhail Gavrilov wrote:
> Hi folks.
> I use Fedora Rawhide and today's update came with the new kernel
> 5.7.0-0.rc0.git6.1.
> As it turned out this kernel won't boot. Instead, I see the black
> screen and some times see messages:
> exit_boot() failed!
> efi_main() failed!
> 
> The last working kernel is kernel-core-5.7.0-0.rc0.git2.1
> 
> I looked at the spec file and found out that the last bootable kernel
> was related to the commit cad18da0afb1
> (https://src.fedoraproject.org/rpms/kernel/c/9949db6070c6ca501d7235760ca8aba120db7448?branch=master),
> and the first non-bootable kernel related to the commit a10c9c710f9e
> (https://src.fedoraproject.org/rpms/kernel/c/0561e0bf0e459dd4b27a77670de838128b87bd0e?branch=master)
> 
> Of course, knowing these two commits, I made bisecting with the vanilla kernel:
> 
> $ git bisect log
> git bisect start
> # good: [cad18da0afb1bc7b37d73a74067ab7ff5974897c] Merge tag
> 'please-pull-ia64_for_5.7' of
> git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux
> git bisect good cad18da0afb1bc7b37d73a74067ab7ff5974897c
> # bad: [a10c9c710f9ecea87b9f4bbb837467893b4bef01] Merge tag 'for-v5.7'
> of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply
> git bisect bad a10c9c710f9ecea87b9f4bbb837467893b4bef01
> # bad: [f365ab31efacb70bed1e821f7435626e0b2528a6] Merge tag
> 'drm-next-2020-04-01' of git://anongit.freedesktop.org/drm/drm
> git bisect bad f365ab31efacb70bed1e821f7435626e0b2528a6
> # good: [7f80ccfe996871ca69648efee74a60ae7ad0dcd9] net: ipv6:
> rpl_iptunnel: Fix potential memory leak in rpl_do_srh_inline
> git bisect good 7f80ccfe996871ca69648efee74a60ae7ad0dcd9
> # good: [9001b17698d86f842e2b13e0cafe8021d43209e9] Merge tag
> 'drm-intel-next-2020-03-13' of
> git://anongit.freedesktop.org/drm/drm-intel into drm-next
> git bisect good 9001b17698d86f842e2b13e0cafe8021d43209e9
> # bad: [dfabb077d62552797ca0ae7756cb30d3e195ead5] Merge tag 'mmc-v5.7'
> of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
> git bisect bad dfabb077d62552797ca0ae7756cb30d3e195ead5
> # bad: [15c981d16d70e8a5be297fa4af07a64ab7e080ed] Merge tag
> 'for-5.7-tag' of
> git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
> git bisect bad 15c981d16d70e8a5be297fa4af07a64ab7e080ed
> # good: [dc9492c14c758639d7b2468d4ed3c77e785c1a35] btrfs: hold a ref
> on the root on the dead roots list
> git bisect good dc9492c14c758639d7b2468d4ed3c77e785c1a35
> # good: [3cd86a58f7734bf9cef38f6f899608ebcaa3da13] Merge tag
> 'arm64-upstream' of
> git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
> git bisect good 3cd86a58f7734bf9cef38f6f899608ebcaa3da13
> # bad: [fdf5563a720004199324371c08071b8ea27bd994] Merge branch
> 'x86-cleanups-for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> git bisect bad fdf5563a720004199324371c08071b8ea27bd994
> # skip: [cf122cfba5b1d9daf64009d143f51dfec4b1705a] kill uaccess_try()
> git bisect skip cf122cfba5b1d9daf64009d143f51dfec4b1705a
> # skip: [4dd2a1b92b91b5f2acf853ee1dc0df135054698f] x86: Replace
> setup_irq() by request_irq()
> git bisect skip 4dd2a1b92b91b5f2acf853ee1dc0df135054698f
> # skip: [5c1f178094631e8b9acc67e4a9b6e03abfbc2529] x86:
> __setup_frame(): consolidate uaccess areas
> git bisect skip 5c1f178094631e8b9acc67e4a9b6e03abfbc2529
> # skip: [3add42c29cebb1d5f83c6205c59466a06ccf8da1] x86: get rid of
> get_user_ex() in restore_sigcontext()
> git bisect skip 3add42c29cebb1d5f83c6205c59466a06ccf8da1
> # skip: [ead8e4e7e2c75ced6fcd9a53d3e9a2ecd7368553] x86:
> __setup_rt_frame(): consolidate uaccess areas
> git bisect skip ead8e4e7e2c75ced6fcd9a53d3e9a2ecd7368553
> # skip: [31a9122058bc5f042cb04bcdb8cd9e6c77fdae8d] x86/mm: Drop
> pud_mknotpresent()
> git bisect skip 31a9122058bc5f042cb04bcdb8cd9e6c77fdae8d
> # skip: [791612e9668cecbf5dd24d13400ac74e099f005c] x86:
> x32_setup_rt_frame(): consolidate uaccess areas
> git bisect skip 791612e9668cecbf5dd24d13400ac74e099f005c
> # skip: [77f3c6166ddc7567455b244074b3ebb63862b56f] x86: kill
> get_user_{try,catch,ex}
> git bisect skip 77f3c6166ddc7567455b244074b3ebb63862b56f
> # skip: [b00d8f8f0b2b39223c3fd6713d318aba95420264] x86:
> setup_sigcontext(): list user_access_{begin,end}() into callers
> git bisect skip b00d8f8f0b2b39223c3fd6713d318aba95420264
> # skip: [0e79ad863df43b01090ae18c97de5c3787f069c6] x86/cpu: Fix a
> -Wmissing-prototypes warning for init_ia32_feat_ctl()
> git bisect skip 0e79ad863df43b01090ae18c97de5c3787f069c6
> # skip: [b87df6594486626a9ae5944807307f2604cea3e2] x86:
> unsafe_put-style macro for sigmask
> git bisect skip b87df6594486626a9ae5944807307f2604cea3e2
> # skip: [a37d01ead405e3aa14d72d284721fe46422b3b63] x86: switch
> save_v86_state() to unsafe_put_user()
> git bisect skip a37d01ead405e3aa14d72d284721fe46422b3b63
> # skip: [119cd59fcfbe70fb3fcab4e64cd232bcc3807585] x86: get rid of
> put_user_try in __setup_rt_frame() (both 32bit and 64bit)
> git bisect skip 119cd59fcfbe70fb3fcab4e64cd232bcc3807585
> # skip: [119cd59fcfbe70fb3fcab4e64cd232bcc3807585] x86: get rid of
> put_user_try in __setup_rt_frame() (both 32bit and 64bit)
> git bisect skip 119cd59fcfbe70fb3fcab4e64cd232bcc3807585
> # bad: [9589351ccf47a85a75180a430627c16bc28da929] Merge branch
> 'x86-boot-for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> git bisect bad 9589351ccf47a85a75180a430627c16bc28da929
> # skip: [003602ad5516e59940de42e44c8d8033387bb363] x86/*/Makefile: Use
> -fno-asynchronous-unwind-tables to suppress .eh_frame sections
> git bisect skip 003602ad5516e59940de42e44c8d8033387bb363
> # skip: [0eea39a234dc52063d14541fabcb2c64516a2328]
> x86/boot/compressed: Remove .eh_frame section from bzImage
> git bisect skip 0eea39a234dc52063d14541fabcb2c64516a2328
> # skip: [bac59d18c7018a2fd5e800a1e72a8271bf404977] x86/setup: Fix
> static memory detection
> git bisect skip bac59d18c7018a2fd5e800a1e72a8271bf404977
> # skip: [3ee372ccce4d4e7c610748d0583979d3ed3a0cf4]
> x86/boot/compressed/64: Remove .bss/.pgtable from bzImage
> git bisect skip 3ee372ccce4d4e7c610748d0583979d3ed3a0cf4
> # skip: [a86255fe5258714e1f7c1bdfe95f08e4d098d450]
> x86/boot/compressed/64: Use 32-bit (zero-extended) MOV for
> z_output_len
> git bisect skip a86255fe5258714e1f7c1bdfe95f08e4d098d450
> # skip: [6f8f0dc980028e98ae339876a8403edae4d20e39] x86/vmlinux: Drop
> unneeded linker script discard of .eh_frame
> git bisect skip 6f8f0dc980028e98ae339876a8403edae4d20e39
> # skip: [48bfdb9deffdc6b683feb25e15f4f26aac503501]
> x86/boot/compressed/64: Use LEA to initialize boot stack pointer
> git bisect skip 48bfdb9deffdc6b683feb25e15f4f26aac503501
> # skip: [c90beea22a2bece4b0bbb39789bf835504421594]
> x86/boot/compressed: Fix debug_puthex() parameter type
> git bisect skip c90beea22a2bece4b0bbb39789bf835504421594
> # only skipped commits left to test
> # possible first bad commit:
> [9589351ccf47a85a75180a430627c16bc28da929] Merge branch
> 'x86-boot-for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> # possible first bad commit:
> [c90beea22a2bece4b0bbb39789bf835504421594] x86/boot/compressed: Fix
> debug_puthex() parameter type
> # possible first bad commit:
> [bac59d18c7018a2fd5e800a1e72a8271bf404977] x86/setup: Fix static
> memory detection
> # possible first bad commit:
> [6f8f0dc980028e98ae339876a8403edae4d20e39] x86/vmlinux: Drop unneeded
> linker script discard of .eh_frame
> # possible first bad commit:
> [003602ad5516e59940de42e44c8d8033387bb363] x86/*/Makefile: Use
> -fno-asynchronous-unwind-tables to suppress .eh_frame sections
> # possible first bad commit:
> [0eea39a234dc52063d14541fabcb2c64516a2328] x86/boot/compressed: Remove
> .eh_frame section from bzImage
> # possible first bad commit:
> [3ee372ccce4d4e7c610748d0583979d3ed3a0cf4] x86/boot/compressed/64:
> Remove .bss/.pgtable from bzImage
> # possible first bad commit:
> [a86255fe5258714e1f7c1bdfe95f08e4d098d450] x86/boot/compressed/64: Use
> 32-bit (zero-extended) MOV for z_output_len
> # possible first bad commit:
> [48bfdb9deffdc6b683feb25e15f4f26aac503501] x86/boot/compressed/64: Use
> LEA to initialize boot stack pointer
> 
> Unfortunately, kernel bisecting not telling exactly which commit is blame.
> In my bisecting a lot of skipping commits because of the unsuccessful builds:
> ld: arch/x86/boot/compressed/pgtable_64.o:(.bss+0x0): multiple
> definition of `__force_order';
> arch/x86/boot/compressed/kaslr_64.o:(.bss+0x0): first defined here
> make[2]: *** [arch/x86/boot/compressed/Makefile:120:
> arch/x86/boot/compressed/vmlinux] Error 1
> make[1]: *** [arch/x86/boot/Makefile:114:
> arch/x86/boot/compressed/vmlinux] Error 2
> make: *** [arch/x86/Makefile:283: bzImage] Error 2
> 
> Can anyone here help me with this issue?

I encountered the same issue as you:

exit_boot() failed!
efi_main() failed!

and the following patches are very suspicious:

x86/boot/compressed/64: Use LEA to initialize boot stack pointer
x86/boot/compressed/64: Use 32-bit (zero-extended) MOV for z_output_len
x86/boot/compressed/64: Remove .bss/.pgtable from bzImage
x86/boot/compressed: Remove .eh_frame section from bzImage

I am not sure if reverting these patches can help you.

Thanks,
Like Xu

> 
> Thanks.
> 
> Downstream bugreport:
> https://bugzilla.redhat.com/show_bug.cgi?id=1822731
> 
> --
> Best Regards,
> Mike Gavrilov.
> 

Powered by blists - more mailing lists