<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="/rss.xsl.xml"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
    <title>Changes in image.h</title>
    <description></description>
    <language>en</language>
    <copyright>Copyright 2015</copyright>
    <generator>Java</generator><item>
        <title>120dc60d - arm64: get rid of TEXT_OFFSET</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#120dc60d</link>
        <description>arm64: get rid of TEXT_OFFSETTEXT_OFFSET serves no purpose, and for this reason, it was redefinedas 0x0 in the v5.8 timeframe. Since this does not appear to have causedany issues that require us to revisit that decision, let&apos;s get rid of themacro entirely, along with any references to it.Signed-off-by: Ard Biesheuvel &lt;ardb@kernel.org&gt;Link: https://lore.kernel.org/r/20200825135440.11288-1-ardb@kernel.orgSigned-off-by: Will Deacon &lt;will@kernel.org&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Tue, 25 Aug 2020 13:54:40 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ardb@kernel.org&gt;</dc:creator>
    </item>
<item>
        <title>90776dd1 - arm64/efi: Move variable assignments after SECTIONS</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#90776dd1</link>
        <description>arm64/efi: Move variable assignments after SECTIONSIt seems that LLVM&apos;s linker does not correctly handle variable assignmentsinvolving section positions that are updated during the SECTIONSparsing. Commit aa69fb62bea1 (&quot;arm64/efi: Mark __efistub_stext_offset asan absolute symbol explicitly&quot;) ran into this too, but found a differentworkaround.However, this was not enough, as other variables were also miscalculatedwhich manifested as boot failures under UEFI where __efistub__end wasnot taking the correct _end value (they should be the same):$ ld.lld -EL -maarch64elf --no-undefined -X -shared \	-Bsymbolic -z notext -z norelro --no-apply-dynamic-relocs \	-o vmlinux.lld -T poc.lds --whole-archive vmlinux.o &amp;&amp; \  readelf -Ws vmlinux.lld | egrep &apos;\b(__efistub_|)_end\b&apos;368272: ffff000002218000     0 NOTYPE  LOCAL  HIDDEN    38 __efistub__end368322: ffff000012318000     0 NOTYPE  GLOBAL DEFAULT   38 _end$ aarch64-linux-gnu-ld.bfd -EL -maarch64elf --no-undefined -X -shared \	-Bsymbolic -z notext -z norelro --no-apply-dynamic-relocs \	-o vmlinux.bfd -T poc.lds --whole-archive vmlinux.o &amp;&amp; \  readelf -Ws vmlinux.bfd | egrep &apos;\b(__efistub_|)_end\b&apos;338124: ffff000012318000     0 NOTYPE  LOCAL  DEFAULT  ABS __efistub__end383812: ffff000012318000     0 NOTYPE  GLOBAL DEFAULT 15325 _endTo work around this, all of the __efistub_-prefixed variable assignmentsneed to be moved after the linker script&apos;s SECTIONS entry. As it turnsout, this also solves the problem fixed in commit aa69fb62bea1, so thosechanges are reverted here.Link: https://github.com/ClangBuiltLinux/linux/issues/634Link: https://bugs.llvm.org/show_bug.cgi?id=42990Acked-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Kees Cook &lt;keescook@chromium.org&gt;Signed-off-by: Will Deacon &lt;will@kernel.org&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Tue, 13 Aug 2019 23:04:50 +0000</pubDate>
        <dc:creator>Kees Cook &lt;keescook@chromium.org&gt;</dc:creator>
    </item>
<item>
        <title>aa69fb62 - arm64/efi: Mark __efistub_stext_offset as an absolute symbol explicitly</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#aa69fb62</link>
        <description>arm64/efi: Mark __efistub_stext_offset as an absolute symbol explicitlyAfter r363059 and r363928 in LLVM, a build using ld.lld as the linkerwith CONFIG_RANDOMIZE_BASE enabled fails like so:ld.lld: error: relocation R_AARCH64_ABS32 cannot be used against symbol__efistub_stext_offset; recompile with -fPICFangrui and Peter figured out that ld.lld is incorrectly considering__efistub_stext_offset as a relative symbol because of the order inwhich symbols are evaluated. _text is treated as an absolute symboland stext is a relative symbol, making __efistub_stext_offset arelative symbol.Adding ABSOLUTE will force ld.lld to evalute this expression in theright context and does not change ld.bfd&apos;s behavior. ld.lld willneed to be fixed but the developers do not see a quick or simple fixwithout some research (see the linked issue for further explanation).Add this simple workaround so that ld.lld can continue to link kernels.Link: https://github.com/ClangBuiltLinux/linux/issues/561Link: https://github.com/llvm/llvm-project/commit/025a815d75d2356f2944136269aa5874721ec236Link: https://github.com/llvm/llvm-project/commit/249fde85832c33f8b06c6b4ac65d1c4b96d23b83Acked-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Debugged-by: Fangrui Song &lt;maskray@google.com&gt;Debugged-by: Peter Smith &lt;peter.smith@linaro.org&gt;Suggested-by: Fangrui Song &lt;maskray@google.com&gt;Signed-off-by: Nathan Chancellor &lt;natechancellor@gmail.com&gt;[will: add comment]Signed-off-by: Will Deacon &lt;will@kernel.org&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Wed, 26 Jun 2019 04:20:17 +0000</pubDate>
        <dc:creator>Nathan Chancellor &lt;natechancellor@gmail.com&gt;</dc:creator>
    </item>
<item>
        <title>caab277b - treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 234</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#caab277b</link>
        <description>treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 234Based on 1 normalized pattern(s):  this program is free software you can redistribute it and or modify  it under the terms of the gnu general public license version 2 as  published by the free software foundation this program is  distributed in the hope that it will be useful but without any  warranty without even the implied warranty of merchantability or  fitness for a particular purpose see the gnu general public license  for more details you should have received a copy of the gnu general  public license along with this program if not see http www gnu org  licensesextracted by the scancode license scanner the SPDX license identifier  GPL-2.0-onlyhas been chosen to replace the boilerplate/reference in 503 file(s).Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;Reviewed-by: Alexios Zavras &lt;alexios.zavras@intel.com&gt;Reviewed-by: Allison Randal &lt;allison@lohutok.net&gt;Reviewed-by: Enrico Weigelt &lt;info@metux.net&gt;Cc: linux-spdx@vger.kernel.orgLink: https://lkml.kernel.org/r/20190602204653.811534538@linutronix.deSigned-off-by: Greg Kroah-Hartman &lt;gregkh@linuxfoundation.org&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Mon, 03 Jun 2019 05:44:50 +0000</pubDate>
        <dc:creator>Thomas Gleixner &lt;tglx@linutronix.de&gt;</dc:creator>
    </item>
<item>
        <title>f56063c5 - arm64: add image head flag definitions</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#f56063c5</link>
        <description>arm64: add image head flag definitionsThose image head&apos;s flags will be used later by kexec_file loader.Signed-off-by: AKASHI Takahiro &lt;takahiro.akashi@linaro.org&gt;Cc: Catalin Marinas &lt;catalin.marinas@arm.com&gt;Cc: Will Deacon &lt;will.deacon@arm.com&gt;Acked-by: James Morse &lt;james.morse@arm.com&gt;Signed-off-by: Will Deacon &lt;will.deacon@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Thu, 15 Nov 2018 05:52:46 +0000</pubDate>
        <dc:creator>AKASHI Takahiro &lt;takahiro.akashi@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>dd6846d7 - arm64: drop linker script hack to hide __efistub_ symbols</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#dd6846d7</link>
        <description>arm64: drop linker script hack to hide __efistub_ symbolsCommit 1212f7a16af4 (&quot;scripts/kallsyms: filter arm64&apos;s __efistub_symbols&quot;) updated the kallsyms code to filter out symbols withthe __efistub_ prefix explicitly, so we no longer require thehack in our linker script to emit them as absolute symbols.Cc: Nick Desaulniers &lt;ndesaulniers@google.com&gt;Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Will Deacon &lt;will.deacon@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Fri, 30 Nov 2018 11:35:58 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>fdfb69a7 - arm64/efi: Make strrchr() available to the EFI namespace</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#fdfb69a7</link>
        <description>arm64/efi: Make strrchr() available to the EFI namespacelibfdt gained a new dependency on strrchr, so make it available to theEFI namespace before we update libfdt.Thanks to Ard for providing this fix.Cc: Catalin Marinas &lt;catalin.marinas@arm.com&gt;Acked-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Acked-by: Will Deacon &lt;will.deacon@arm.com&gt;Signed-off-by: Rob Herring &lt;robh@kernel.org&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Thu, 01 Mar 2018 16:12:07 +0000</pubDate>
        <dc:creator>Rob Herring &lt;robh@kernel.org&gt;</dc:creator>
    </item>
<item>
        <title>57fdb89a - arm64/efi/libstub: Make screen_info accessible to the UEFI stub</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#57fdb89a</link>
        <description>arm64/efi/libstub: Make screen_info accessible to the UEFI stubUnlike on 32-bit ARM, where we need to pass the stub&apos;s version of structscreen_info to the kernel proper via a configuration table, on 64-bit ARMit simply involves making the core kernel&apos;s copy of struct screen_infovisible to the stub by exposing an __efistub_ alias for it.Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Matt Fleming &lt;matt@codeblueprint.co.uk&gt;Acked-by: Will Deacon &lt;will.deacon@arm.com&gt;Cc: Borislav Petkov &lt;bp@alien8.de&gt;Cc: David Herrmann &lt;dh.herrmann@gmail.com&gt;Cc: Mark Rutland &lt;mark.rutland@arm.com&gt;Cc: Peter Jones &lt;pjones@redhat.com&gt;Cc: Peter Zijlstra &lt;peterz@infradead.org&gt;Cc: Thomas Gleixner &lt;tglx@linutronix.de&gt;Cc: linux-efi@vger.kernel.orgLink: http://lkml.kernel.org/r/1461614832-17633-21-git-send-email-matt@codeblueprint.co.ukSigned-off-by: Ingo Molnar &lt;mingo@kernel.org&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Mon, 25 Apr 2016 20:06:52 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>18b9c0d6 - arm64: don&apos;t map TEXT_OFFSET bytes below the kernel if we can avoid it</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#18b9c0d6</link>
        <description>arm64: don&apos;t map TEXT_OFFSET bytes below the kernel if we can avoid itFor historical reasons, the kernel Image must be loaded into physicalmemory at a 512 KB offset above a 2 MB aligned base address. The regionbetween the base address and the start of the kernel Image has nosignificance to the kernel itself, but it is currently mapped explicitlyinto the early kernel VMA range for all translation granules.In some cases (i.e., 4 KB granule), this is unavoidable, due to the 2 MBgranularity of the early kernel mappings. However, in other cases, e.g.,when running with larger page sizes, or in the future, with more granularKASLR, there is no reason to map it explicitly like we do currently.So update the logic so that the region is mapped only if that happens asa side effect of rounding the start address of the kernel to swapper blocksize, and leave it unmapped otherwise.Since the symbol kernel_img_size now simply resolves to the memoryfootprint of the kernel Image, we can drop its definition from image.hand opencode its calculation.Acked-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Will Deacon &lt;will.deacon@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Mon, 18 Apr 2016 15:09:46 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>546c8c44 - arm64: move early boot code to the .init segment</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#546c8c44</link>
        <description>arm64: move early boot code to the .init segmentApart from the arm64/linux and EFI header data structures, there is nothingin the .head.text section that must reside at the beginning of the Image.So let&apos;s move it to the .init section where it belongs.Note that this involves some minor tweaking of the EFI header, primarilybecause the address of &apos;stext&apos; no longer coincides with the start of the.text section. It also requires a couple of relocated symbol referencesto be slightly rewritten or their definition moved to the linker script.Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Will Deacon &lt;will.deacon@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Wed, 30 Mar 2016 15:43:07 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>6ad1fe5d - arm64: avoid R_AARCH64_ABS64 relocations for Image header fields</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#6ad1fe5d</link>
        <description>arm64: avoid R_AARCH64_ABS64 relocations for Image header fieldsUnfortunately, the current way of using the linker to emit build timeconstants into the Image header will no longer work once we switch tothe use of PIE executables. The reason is that such constants are emittedinto the binary using R_AARCH64_ABS64 relocations, which are resolved atruntime, not at build time, and the places targeted by those relocationswill contain zeroes before that.So refactor the endian swapping linker script constant generation code sothat it emits the upper and lower 32-bit words separately.Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Sat, 26 Dec 2015 12:48:02 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>a7f8de16 - arm64: allow kernel Image to be loaded anywhere in physical memory</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#a7f8de16</link>
        <description>arm64: allow kernel Image to be loaded anywhere in physical memoryThis relaxes the kernel Image placement requirements, so that itmay be placed at any 2 MB aligned offset in physical memory.This is accomplished by ignoring PHYS_OFFSET when installingmemblocks, and accounting for the apparent virtual offset ofthe kernel Image. As a result, virtual address referencesbelow PAGE_OFFSET are correctly mapped onto physical referencesinto the kernel Image regardless of where it sits in memory.Special care needs to be taken for dealing with memory limits passedvia mem=, since the generic implementation clips memory top down, whichmay clip the kernel image itself if it is loaded high up in memory. Todeal with this case, we simply add back the memory covering the kernelimage, which may result in more memory to be retained than was passedas a mem= parameter.Since mem= should not be considered a production feature, a panic notifierhandler is installed that dumps the memory limit at panic time if one wasset.Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Tue, 16 Feb 2016 12:52:42 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>7f4e3462 - arm64/efi: Make strnlen() available to the EFI namespace</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#7f4e3462</link>
        <description>arm64/efi: Make strnlen() available to the EFI namespaceChanges introduced in the upstream version of libfdt pulled in by commit91feabc2e224 (&quot;scripts/dtc: Update to upstream commit b06e55c88b9b&quot;) usethe strnlen() function, which isn&apos;t currently available to the EFI name-space. Add it to the EFI namespace to avoid a linker error.Cc: Catalin Marinas &lt;catalin.marinas@arm.com&gt;Cc: Rob Herring &lt;robh@kernel.org&gt;Reviewed-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Thierry Reding &lt;treding@nvidia.com&gt;Signed-off-by: Will Deacon &lt;will.deacon@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Tue, 16 Feb 2016 10:16:31 +0000</pubDate>
        <dc:creator>Thierry Reding &lt;treding@nvidia.com&gt;</dc:creator>
    </item>
<item>
        <title>75feee3d - arm64: hide __efistub_ aliases from kallsyms</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#75feee3d</link>
        <description>arm64: hide __efistub_ aliases from kallsymsCommit e8f3010f7326 (&quot;arm64/efi: isolate EFI stub from the kernelproper&quot;) isolated the EFI stub code from the kernel proper by prefixingall of its symbols with __efistub_, and selectively allowing access tocore kernel symbols from the stub by emitting __efistub_ aliases forfunctions and variables that the stub can access legally.As an unintended side effect, these aliases are emitted into thekallsyms symbol table, which means they may turn up in backtraces,e.g.,  ...  PC is at __efistub_memset+0x108/0x200  LR is at fixup_init+0x3c/0x48  ...  [&lt;ffffff8008328608&gt;] __efistub_memset+0x108/0x200  [&lt;ffffff8008094dcc&gt;] free_initmem+0x2c/0x40  [&lt;ffffff8008645198&gt;] kernel_init+0x20/0xe0  [&lt;ffffff8008085cd0&gt;] ret_from_fork+0x10/0x40The backtrace in question has nothing to do with the EFI stub, butsimply returns one of the several aliases of memset() that have beenrecorded in the kallsyms table. This is undesirable, since it maysuggest to people who are not aware of this that the issue they areseeing is somehow EFI related.So hide the __efistub_ aliases from kallsyms, by emitting them asabsolute linker symbols explicitly. The distinction between thoseand section relative symbols is completely irrelevant to thesedefinitions, and to the final link we are performing when thesedefinitions are being taken into account (the distinction is onlyrelevant to symbols defined inside a section definition when performinga partial link), and so the resulting values are identical to theoriginal ones. Since absolute symbols are ignored by kallsyms, thiswill result in these values to be omitted from its symbol table.After this patch, the backtrace generated from the same address lookslike this:  ...  PC is at __memset+0x108/0x200  LR is at fixup_init+0x3c/0x48  ...  [&lt;ffffff8008328608&gt;] __memset+0x108/0x200  [&lt;ffffff8008094dcc&gt;] free_initmem+0x2c/0x40  [&lt;ffffff8008645198&gt;] kernel_init+0x20/0xe0  [&lt;ffffff8008085cd0&gt;] ret_from_fork+0x10/0x40Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Will Deacon &lt;will.deacon@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Fri, 15 Jan 2016 12:28:57 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>9d372c9f - arm64: Add page size to the kernel image header</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#9d372c9f</link>
        <description>arm64: Add page size to the kernel image headerThis patch adds the page size to the arm64 kernel image headerso that one can infer the PAGESIZE used by the kernel. This willbe helpful to diagnose failures to boot the kernel with page sizenot supported by the CPU.Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Signed-off-by: Suzuki K. Poulose &lt;suzuki.poulose@arm.com&gt;Acked-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;Reviewed-by: Christoffer Dall &lt;christoffer.dall@linaro.org&gt;Acked-by: Mark Rutland &lt;mark.rutland@arm.com&gt;Signed-off-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Mon, 19 Oct 2015 13:19:36 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>39d114dd - arm64: add KASAN support</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#39d114dd</link>
        <description>arm64: add KASAN supportThis patch adds arch specific code for kernel address sanitizer(see Documentation/kasan.txt).1/8 of kernel addresses reserved for shadow memory. There was nobig enough hole for this, so virtual addresses for shadow werestolen from vmalloc area.At early boot stage the whole shadow region populated with justone physical page (kasan_zero_page). Later, this page reusedas readonly zero shadow for some memory that KASan currentlydon&apos;t track (vmalloc).After mapping the physical memory, pages for shadow memory areallocated and mapped.Functions like memset/memmove/memcpy do a lot of memory accesses.If bad pointer passed to one of these function it is importantto catch this. Compiler&apos;s instrumentation cannot do this sincethese functions are written in assembly.KASan replaces memory functions with manually instrumented variants.Original functions declared as weak symbols so strong definitionsin mm/kasan/kasan.c could replace them. Original functions have aliaseswith &apos;__&apos; prefix in name, so we could call non-instrumented variantif needed.Some files built without kasan instrumentation (e.g. mm/slub.c).Original mem* function replaced (via #define) with prefixed variantsto disable memory access checks for such files.Signed-off-by: Andrey Ryabinin &lt;ryabinin.a.a@gmail.com&gt;Tested-by: Linus Walleij &lt;linus.walleij@linaro.org&gt;Reviewed-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;Signed-off-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Mon, 12 Oct 2015 15:52:58 +0000</pubDate>
        <dc:creator>Andrey Ryabinin &lt;ryabinin.a.a@gmail.com&gt;</dc:creator>
    </item>
<item>
        <title>e8f3010f - arm64/efi: isolate EFI stub from the kernel proper</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#e8f3010f</link>
        <description>arm64/efi: isolate EFI stub from the kernel properSince arm64 does not use a builtin decompressor, the EFI stub is builtinto the kernel proper. So far, this has been working fine, but actually,since the stub is in fact a PE/COFF relocatable binary that is executedat an unknown offset in the 1:1 mapping provided by the UEFI firmware, weshould not be seamlessly sharing code with the kernel proper, which is aposition dependent executable linked at a high virtual offset.So instead, separate the contents of libstub and its dependencies, byputting them into their own namespace by prefixing all of its symbolswith __efistub. This way, we have tight control over what parts of thekernel proper are referenced by the stub.Signed-off-by: Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;Reviewed-by: Matt Fleming &lt;matt.fleming@intel.com&gt;Signed-off-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Thu, 08 Oct 2015 19:02:04 +0000</pubDate>
        <dc:creator>Ard Biesheuvel &lt;ard.biesheuvel@linaro.org&gt;</dc:creator>
    </item>
<item>
        <title>a2c1d73b - arm64: Update the Image header</title>
        <link>http://172.16.0.5:8080/history/linux-6.15/arch/arm64/kernel/image.h#a2c1d73b</link>
        <description>arm64: Update the Image headerCurrently the kernel Image is stripped of everything past the initialstack, and at runtime the memory is initialised and used by the kernel.This makes the effective minimum memory footprint of the kernel largerthan the size of the loaded binary, though bootloaders have no mechanismto identify how large this minimum memory footprint is. This makes itdifficult to choose safe locations to place both the kernel and otherbinaries required at boot (DTB, initrd, etc), such that the kernel won&apos;tclobber said binaries or other reserved memory during initialisation.Additionally when big endian support was added the image load offset wasoverlooked, and is currently of an arbitrary endianness, which makes itdifficult for bootloaders to make use of it. It seems that bootloadersaren&apos;t respecting the image load offset at present anyway, and areassuming that offset 0x80000 will always be correct.This patch adds an effective image size to the kernel header whichdescribes the amount of memory from the start of the kernel Image binarywhich the kernel expects to use before detecting memory and handling anymemory reservations. This can be used by bootloaders to choose suitablelocations to load the kernel and/or other binaries such that the kernelwill not clobber any memory unexpectedly. As before, memory reservationsare required to prevent the kernel from clobbering these locationslater.Both the image load offset and the effective image size are forced to belittle-endian regardless of the native endianness of the kernel toenable bootloaders to load a kernel of arbitrary endianness. Bootloaderswhich wish to make use of the load offset can inspect the effectiveimage size field for a non-zero value to determine if the offset is of aknown endianness. To enable software to determine the endinanness of thekernel as may be required for certain use-cases, a new flags field (alsolittle-endian) is added to the kernel header to export this information.The documentation is updated to clarify these details. To discouragefuture assumptions regarding the value of text_offset, the value at thispoint in time is removed from the main flow of the documentation (thoughkept as a compatibility note). Some minor formatting issues in thedocumentation are also corrected.Signed-off-by: Mark Rutland &lt;mark.rutland@arm.com&gt;Acked-by: Tom Rini &lt;trini@ti.com&gt;Cc: Geoff Levand &lt;geoff@infradead.org&gt;Cc: Kevin Hilman &lt;kevin.hilman@linaro.org&gt;Acked-by: Will Deacon &lt;will.deacon@arm.com&gt;Signed-off-by: Catalin Marinas &lt;catalin.marinas@arm.com&gt;

            List of files:
            /linux-6.15/arch/arm64/kernel/image.h</description>
        <pubDate>Tue, 24 Jun 2014 15:51:36 +0000</pubDate>
        <dc:creator>Mark Rutland &lt;mark.rutland@arm.com&gt;</dc:creator>
    </item>
</channel>
</rss>
