[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <202001301405.D90AF02@keescook>
Date: Thu, 30 Jan 2020 14:06:42 -0800
From: Kees Cook <keescook@...omium.org>
To: "H.J. Lu" <hjl.tools@...il.com>
Cc: LKML <linux-kernel@...r.kernel.org>,
Andy Lutomirski <luto@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Thomas Lendacky <Thomas.Lendacky@....com>,
Sami Tolvanen <samitolvanen@...gle.com>,
Heiko Carstens <heiko.carstens@...ibm.com>,
Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
"H . Peter Anvin" <hpa@...or.com>,
the arch/x86 maintainers <x86@...nel.org>,
Yu-cheng Yu <yu-cheng.yu@...el.com>
Subject: Re: [PATCH 2/2] x86: Discard .note.gnu.property sections in vmlinux
On Thu, Jan 30, 2020 at 12:20:30PM -0800, H.J. Lu wrote:
> On Thu, Jan 30, 2020 at 12:08 PM Kees Cook <keescook@...omium.org> wrote:
> >
> > On Thu, Jan 30, 2020 at 12:04:54PM -0800, H.J. Lu wrote:
> > > > I don't understand this. "may not be incompatible"? Is there an error
> > > > generated? If so, what does it look like?
> > >
> > > When -mx86-used-note=yes is passed to assembler, with my patch, I got
> > >
> > > [hjl@...-skx-1 linux]$ readelf -n vmlinux
> > >
> > > Displaying notes found in: .notes
> > > Owner Data size Description
> > > Xen 0x00000006 Unknown note type: (0x00000006)
> > > description data: 6c 69 6e 75 78 00
> > > Xen 0x00000004 Unknown note type: (0x00000007)
> > > description data: 32 2e 36 00
> > > Xen 0x00000008 Unknown note type: (0x00000005)
> > > description data: 78 65 6e 2d 33 2e 30 00
> > > Xen 0x00000008 Unknown note type: (0x00000003)
> > > description data: 00 00 00 80 ff ff ff ff
> > > Xen 0x00000008 Unknown note type: (0x0000000f)
> > > description data: 00 00 00 00 80 00 00 00
> > > Xen 0x00000008 NT_VERSION (version)
> > > description data: 80 a1 ba 82 ff ff ff ff
> > > Xen 0x00000008 NT_ARCH (architecture)
> > > description data: 00 10 00 81 ff ff ff ff
> > > Xen 0x00000029 Unknown note type: (0x0000000a)
> > > description data: 21 77 72 69 74 61 62 6c 65 5f 70 61 67 65 5f 74
> > > 61 62 6c 65 73 7c 70 61 65 5f 70 67 64 69 72 5f 61 62 6f 76 65 5f 34
> > > 67 62
> > > Xen 0x00000004 Unknown note type: (0x00000011)
> > > description data: 01 88 00 00
> > > Xen 0x00000004 Unknown note type: (0x00000009)
> > > description data: 79 65 73 00
> > > Xen 0x00000008 Unknown note type: (0x00000008)
> > > description data: 67 65 6e 65 72 69 63 00
> > > Xen 0x00000010 Unknown note type: (0x0000000d)
> > > description data: 01 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00
> > > Xen 0x00000004 Unknown note type: (0x0000000e)
> > > description data: 01 00 00 00
> > > Xen 0x00000004 Unknown note type: (0x00000010)
> > > description data: 01 00 00 00
> > > Xen 0x00000008 Unknown note type: (0x0000000c)
> > > description data: 00 00 00 00 00 80 ff ff
> > > Xen 0x00000008 Unknown note type: (0x00000004)
> > > description data: 00 00 00 00 00 00 00 00
> > > GNU 0x00000014 NT_GNU_BUILD_ID (unique build ID bitstring)
> > > Build ID: 11c73de2922f593e1b35b92ab3c70eaa1a80fa83
> > > Linux 0x00000018 OPEN
> > > description data: 35 2e 33 2e 39 2d 32 30 30 2e 30 2e 66 63 33 30
> > > 2e 78 38 36 5f 36 34 00
> > > Xen 0x00000008 Unknown note type: (0x00000012)
> > > description data: 70 04 00 01 00 00 00 00
> > > [hjl@...-skx-1 linux]$
> > >
> > > Without my patch,
> > >
> > > [hjl@...-skx-1 linux]$ readelf -n vmlinux
> > >
> > > Displaying notes found in: .notes
> > > Owner Data size Description
> > > Xen 0x00000006 Unknown note type: (0x00000006)
> > > description data: 6c 69 6e 75 78 00
> > > Xen 0x00000004 Unknown note type: (0x00000007)
> > > description data: 32 2e 36 00
> > > xen-3.0 0x00000005 Unknown note type: (0x006e6558)
> > > description data: 08 00 00 00 03
> > > readelf: Warning: note with invalid namesz and/or descsz found at offset 0x50
> > > readelf: Warning: type: 0xffffffff, namesize: 0x006e6558, descsize:
> > > 0x80000000, alignment: 8
> > > [hjl@...-skx-1 linux]$
> >
> > What is the source of this failure? Does readelf need updating instead?
> > Is the linking step producing an invalid section? It seems like
> > discarding the properties isn't the right solution here?
>
> With the command-line option, -mx86-used-note=yes, the x86 assembler
> in binutils 2.32 and above generates a program property note in a note
> section, .note.gnu.property, to encode used x86 ISAs and features.
> But x86 kernel linker script only contains a signle NOTE segment:
typo: signle -> single
>
> PHDRS {
> text PT_LOAD FLAGS(5);
> data PT_LOAD FLAGS(6);
> percpu PT_LOAD FLAGS(6);
> init PT_LOAD FLAGS(7);
> note PT_NOTE FLAGS(0);
> }
> SECTIONS
> {
> ...
> .notes : AT(ADDR(.notes) - 0xffffffff80000000) { __start_notes = .; KEEP(*(.not
> e.*)) __stop_notes = .; } :text :note
> ...
> }
>
> But .note.gnu.property must be 8-byte aligned. Linker deals with it
> by generating
> 2 PT_NOTE segments, one has 4-byte alignment and the other has 8-byte alignment:
Ah-ha! Okay, thanks for this detail.
>
> [hjl@...-cfl-1 ~]$ readelf -l /usr/local/bin/ld
>
> Elf file type is EXEC (Executable file)
> Entry point 0x404530
> There are 13 program headers, starting at offset 64
>
> Program Headers:
> Type Offset VirtAddr PhysAddr
> FileSiz MemSiz Flags Align
> PHDR 0x0000000000000040 0x0000000000400040 0x0000000000400040
> 0x00000000000002d8 0x00000000000002d8 R 0x8
> INTERP 0x0000000000000318 0x0000000000400318 0x0000000000400318
> 0x000000000000001c 0x000000000000001c R 0x1
> [Requesting program interpreter: /lib64/ld-linux-x86-64.so.2]
> LOAD 0x0000000000000000 0x0000000000400000 0x0000000000400000
> 0x0000000000002a30 0x0000000000002a30 R 0x1000
> LOAD 0x0000000000003000 0x0000000000403000 0x0000000000403000
> 0x00000000000d7b35 0x00000000000d7b35 R E 0x1000
> LOAD 0x00000000000db000 0x00000000004db000 0x00000000004db000
> 0x0000000000179248 0x0000000000179248 R 0x1000
> LOAD 0x0000000000254de0 0x0000000000655de0 0x0000000000655de0
> 0x00000000000062e8 0x000000000000ba68 RW 0x1000
> DYNAMIC 0x0000000000254df0 0x0000000000655df0 0x0000000000655df0
> 0x0000000000000200 0x0000000000000200 RW 0x8
> NOTE 0x0000000000000338 0x0000000000400338 0x0000000000400338
> 0x0000000000000030 0x0000000000000030 R 0x8
> NOTE 0x0000000000000368 0x0000000000400368 0x0000000000400368
> 0x0000000000000044 0x0000000000000044 R 0x4
> GNU_PROPERTY 0x0000000000000338 0x0000000000400338 0x0000000000400338
> 0x0000000000000030 0x0000000000000030 R 0x8
> GNU_EH_FRAME 0x0000000000233efc 0x0000000000633efc 0x0000000000633efc
> 0x000000000000478c 0x000000000000478c R 0x4
> GNU_STACK 0x0000000000000000 0x0000000000000000 0x0000000000000000
> 0x0000000000000000 0x0000000000000000 RW 0x10
> GNU_RELRO 0x0000000000254de0 0x0000000000655de0 0x0000000000655de0
> 0x0000000000000220 0x0000000000000220 R 0x1
>
> Section to Segment mapping:
> Segment Sections...
> 00
> 01 .interp
> 02 .interp .note.gnu.property .note.gnu.build-id .note.ABI-tag
> .hash .gnu.hash .dynsym .dynstr .gnu.version .gnu.version_r .rela.dyn
> .rela.plt
> 03 .init .plt .text .fini
> 04 .rodata .eh_frame_hdr .eh_frame
> 05 .init_array .fini_array .dynamic .got .got.plt .data .bss
> 06 .dynamic
> 07 .note.gnu.property
> 08 .note.gnu.build-id .note.ABI-tag
> 09 .note.gnu.property
> 10 .eh_frame_hdr
> 11
> 12 .init_array .fini_array .dynamic .got
> [hjl@...-cfl-1 ~]$
>
> Since .note.gnu.property in vmlinux is unused, it can be discarded.
Okay, I'm convinced. If we ever DO need it, it seems the solution would
be to make the regular .note 8 byte aligned...
Thanks for the extra details!
-Kees
--
Kees Cook
Powered by blists - more mailing lists