Re: [PATCH 6.1 000/791] 6.1.113-rc2 review

[Guenter Roeck <linux@roeck-us.net>]

On 10/15/24 04:26, Greg Kroah-Hartman wrote:
> This is the start of the stable review cycle for the 6.1.113 release.
> There are 791 patches in this series, all will be posted as a response
> to this one.  If anyone has any issues with these being applied, please
> let me know.
> 
> Responses should be made by Thu, 17 Oct 2024 11:22:41 +0000.
> Anything received after that time might be too late.
> 
...

> Nicholas Piggin <npiggin@gmail.com>
>      powerpc/64: Option to build big-endian with ELFv2 ABI
> 

With this patch in the tree, trying to build ppc:allmodconfig
with gcc version 13.3 and binutils version 2.42 results in:

Error log:
/tmp/cccUl45i.s: Assembler messages:
/tmp/cccUl45i.s: Error: .size expression for gcm_init_p8 does not evaluate to a constant

and many other similar messages. The problem affects various drivers in
drivers/crypto/vmx/. The problem afects all configurations with CRYPTO_DEV_VMX_ENCRYPT
enabled.

Reverting the patch doesn't work because this patch is part of a larger series. However,
I found that applying upstream commit 505ea33089dc ("powerpc/64: Add big-endian ELFv2
flavour to crypto VMX asm generation") fixes the problem.

Guenter

---
# bad: [54d90d17e8cee20b163d395829162cec92b583f4] Linux 6.1.113
# good: [aa4cd140bba57b7064b4c7a7141bebd336d32087] Linux 6.1.112
git bisect start 'HEAD' 'v6.1.112'
# bad: [2bf4c101d7c99483b8b15a0c8f881e3f399f7e18] net: ethernet: lantiq_etop: fix memory disclosure
git bisect bad 2bf4c101d7c99483b8b15a0c8f881e3f399f7e18
# good: [f88f1145e134fe1da3966d86adb5d813ce2d7c1a] PCI/PM: Increase wait time after resume
git bisect good f88f1145e134fe1da3966d86adb5d813ce2d7c1a
# good: [12aea49495d99bebf185275e7ff33deee4d849a9] fs: Create a generic is_dot_dotdot() utility
git bisect good 12aea49495d99bebf185275e7ff33deee4d849a9
# bad: [f8a29300150e2b18405ff62cc4ed1554bc9c431d] usb: yurex: Replace snprintf() with the safer scnprintf() variant
git bisect bad f8a29300150e2b18405ff62cc4ed1554bc9c431d
# good: [5410d1529047dd32039a2796ea9bd955c6f38b1d] ACPI: resource: Add another DMI match for the TongFang GMxXGxx
git bisect good 5410d1529047dd32039a2796ea9bd955c6f38b1d
# good: [85868884298a81078f5be51be75ad46c14c6e831] hwrng: cctrng - Add missing clk_disable_unprepare in cctrng_resume
git bisect good 85868884298a81078f5be51be75ad46c14c6e831
# good: [41fbdd452460341399757b4db3cd12a67951fb83] EDAC/igen6: Fix conversion of system address to physical memory address
git bisect good 41fbdd452460341399757b4db3cd12a67951fb83
# good: [b986ec200f9fb5c4d863d789e28f45cd2f253656] soc: versatile: realview: fix soc_dev leak during device remove
git bisect good b986ec200f9fb5c4d863d789e28f45cd2f253656
# bad: [9eb76d5168c10a46647df5514acc863039a44885] powerpc/64: Add support to build with prefixed instructions
git bisect bad 9eb76d5168c10a46647df5514acc863039a44885
# bad: [8b9f7d8d71bf9b91ad4cb1ff589d7cdf4bc0673a] powerpc/64: Option to build big-endian with ELFv2 ABI
git bisect bad 8b9f7d8d71bf9b91ad4cb1ff589d7cdf4bc0673a
# first bad commit: [8b9f7d8d71bf9b91ad4cb1ff589d7cdf4bc0673a] powerpc/64: Option to build big-endian with ELFv2 ABI