[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v3 1/4] x86/asm: move inline string functions to <asm/string_inline.h>
- To: Mauricio Faria de Oliveira <mfo@xxxxxxxxxx>, Thomas Gleixner <tglx@xxxxxxxxxx>, Ingo Molnar <mingo@xxxxxxxxxx>, Borislav Petkov <bp@xxxxxxxxx>, Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>, x86@xxxxxxxxxx, "H. Peter Anvin" <hpa@xxxxxxxxx>, Alexey Dobriyan <adobriyan@xxxxxxxxx>, Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>
- From: Juergen Gross <jgross@xxxxxxxx>
- Date: Tue, 26 May 2026 11:51:46 +0200
- Authentication-results: eu.smtp.expurgate.cloud; dkim=pass header.s=google header.d=suse.com header.i="@suse.com" header.h="In-Reply-To:Autocrypt:From:Content-Language:References:Cc:To:Subject:User-Agent:MIME-Version:Date:Message-ID"
- Autocrypt: addr=jgross@xxxxxxxx; keydata= xsBNBFOMcBYBCACgGjqjoGvbEouQZw/ToiBg9W98AlM2QHV+iNHsEs7kxWhKMjrioyspZKOB ycWxw3ie3j9uvg9EOB3aN4xiTv4qbnGiTr3oJhkB1gsb6ToJQZ8uxGq2kaV2KL9650I1SJve dYm8Of8Zd621lSmoKOwlNClALZNew72NjJLEzTalU1OdT7/i1TXkH09XSSI8mEQ/ouNcMvIJ NwQpd369y9bfIhWUiVXEK7MlRgUG6MvIj6Y3Am/BBLUVbDa4+gmzDC9ezlZkTZG2t14zWPvx XP3FAp2pkW0xqG7/377qptDmrk42GlSKN4z76ELnLxussxc7I2hx18NUcbP8+uty4bMxABEB AAHNH0p1ZXJnZW4gR3Jvc3MgPGpncm9zc0BzdXNlLmNvbT7CwHkEEwECACMFAlOMcK8CGwMH CwkIBwMCAQYVCAIJCgsEFgIDAQIeAQIXgAAKCRCw3p3WKL8TL8eZB/9G0juS/kDY9LhEXseh mE9U+iA1VsLhgDqVbsOtZ/S14LRFHczNd/Lqkn7souCSoyWsBs3/wO+OjPvxf7m+Ef+sMtr0 G5lCWEWa9wa0IXx5HRPW/ScL+e4AVUbL7rurYMfwCzco+7TfjhMEOkC+va5gzi1KrErgNRHH kg3PhlnRY0Udyqx++UYkAsN4TQuEhNN32MvN0Np3WlBJOgKcuXpIElmMM5f1BBzJSKBkW0Jc Wy3h2Wy912vHKpPV/Xv7ZwVJ27v7KcuZcErtptDevAljxJtE7aJG6WiBzm+v9EswyWxwMCIO RoVBYuiocc51872tRGywc03xaQydB+9R7BHPzsBNBFOMcBYBCADLMfoA44MwGOB9YT1V4KCy vAfd7E0BTfaAurbG+Olacciz3yd09QOmejFZC6AnoykydyvTFLAWYcSCdISMr88COmmCbJzn sHAogjexXiif6ANUUlHpjxlHCCcELmZUzomNDnEOTxZFeWMTFF9Rf2k2F0Tl4E5kmsNGgtSa aMO0rNZoOEiD/7UfPP3dfh8JCQ1VtUUsQtT1sxos8Eb/HmriJhnaTZ7Hp3jtgTVkV0ybpgFg w6WMaRkrBh17mV0z2ajjmabB7SJxcouSkR0hcpNl4oM74d2/VqoW4BxxxOD1FcNCObCELfIS auZx+XT6s+CE7Qi/c44ibBMR7hyjdzWbABEBAAHCwF8EGAECAAkFAlOMcBYCGwwACgkQsN6d 1ii/Ey9D+Af/WFr3q+bg/8v5tCknCtn92d5lyYTBNt7xgWzDZX8G6/pngzKyWfedArllp0Pn fgIXtMNV+3t8Li1Tg843EXkP7+2+CQ98MB8XvvPLYAfW8nNDV85TyVgWlldNcgdv7nn1Sq8g HwB2BHdIAkYce3hEoDQXt/mKlgEGsLpzJcnLKimtPXQQy9TxUaLBe9PInPd+Ohix0XOlY+Uk QFEx50Ki3rSDl2Zt2tnkNYKUCvTJq7jvOlaPd6d/W0tZqpyy7KVay+K4aMobDsodB3dvEAs6 ScCnh03dDAFgIq5nsB11j3KPKdVoPlfucX2c7kGNH+LUMbzqV6beIENfNexkOfxHfw==
- Cc: kernel-dev@xxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, xen-devel@xxxxxxxxxxxxxxxxxxxx, kernel test robot <lkp@xxxxxxxxx>
- Delivery-date: Tue, 26 May 2026 09:51:55 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
On 20.05.26 23:12, Mauricio Faria de Oliveira wrote:
In next patch, inline string functions are included from 'boot/string.c'.
Please don't use "In next patch". You can't be sure the patches of a
series are going to be committed all together.
A better alternative wording would be "In a future patch".
Using the header <asm/string.h> is problematic for a couple of reasons (*)
(i.e., build errors), which can be addressed, but introduce unnecessary
complexity and regression risk (beyond these _found_ couple of reasons).
Using a new header <asm/string_inline.h> is simpler and transparent to
existing users of <asm/string.h>, with less changes to 'boot/string.c'
and its users (eg 'boot/compressed/string.c' and 'purgatory/purgatory.ro'),
which minimize regression risk.
No functional change intended.
I'd rather put the footnote below after the "---" line, as it is more
interesting for the reviewer than the consumer of "git log" or "git show".
Note that others might disagree, though.
(*) Reasons not to include <asm/string.h> in 'boot/string.c':
1) 'boot/string.c' is built for 16-bit/real mode thus some type and word
size errors happen when <asm/string.h> include, e.g., <asm/string_64.h>.
This can be addressed with '#ifndef _SETUP' (defined by 'boot/Makefile').
2) 'boot/string.c' is included by 'boot/compressed/string.c' which is
the source of 'purgatory/string.o', linked by 'purgatory/purgatory.ro'
(CONFIG_KEXEC_FILE).
In 64BIT, <asm/string.h> includes <asm/string_64.h>, which references
__memset() and __memmove() with KCFI_REFERENCE(), ie, __ADDRESSABLE();
however, 'purgatory/purgatory.ro' is not linked with implementations.
So, CONFIG_KEXEC_FILE and CONFIG_CFI without CONFIG_KASAN hit errors:
>> ld.lld: error: undefined symbol: __memset
>>> referenced by string.c
>>> arch/x86/purgatory/purgatory.ro:\
(__UNIQUE_ID_addressable___memset_0)
--
>> ld.lld: error: undefined symbol: __memmove
>>> referenced by string.c
>>> arch/x86/purgatory/purgatory.ro:\
(__UNIQUE_ID_addressable___memmove_1)
(Note: this is not hit with CONFIG_KASAN because 'boot/compressed/string.c'
adds aliases __memset()/__memmove() to memset()/memmove() in that case.)
This can be addressed with 'CFLAGS_string.o := -D__DISABLE_EXPORTS' so to
disable KCFI_REFERENCE() in 'purgatory/Makefile' (it removes CC_FLAGS_CFI
anyway).
...
However, since a change in this series would need more changes to address
errors it causes, it is reasonable to change the series not to cause them,
by using a separate header with _just_ inline string functions.
Reported-by: kernel test robot <lkp@xxxxxxxxx>
Closes:
https://lore.kernel.org/oe-kbuild-all/202605140922.q7IlUv7o-lkp@xxxxxxxxx/
Signed-off-by: Mauricio Faria de Oliveira <mfo@xxxxxxxxxx>
With (at least) my first remark addressed:
Reviewed-by: Juergen Gross <jgross@xxxxxxxx>
Juergen
Attachment:
OpenPGP_0xB0DE9DD628BF132F.asc
Description: OpenPGP public key
Attachment:
OpenPGP_signature.asc
Description: OpenPGP digital signature
|