NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: port-amd64/58366: KASLR broken



The following reply was made to PR port-amd64/58366; it has been noted by GNATS.

From: RVP <rvp%SDF.ORG@localhost>
To: Taylor R Campbell <campbell%mumble.net@localhost>
Cc: Harold Gutch <logix%foobar.franken.de@localhost>, gnats-bugs%NetBSD.org@localhost,
        port-amd64-maintainer%NetBSD.org@localhost
Subject: Re: port-amd64/58366: KASLR broken
Date: Wed, 18 Nov 2026 00:14:06 +0000 (UTC)

 On Sun, 30 Jun 2024, Taylor R Campbell wrote:
 
 > [...]
 > It looks like when ld splits sections, it sometimes chooses
 > non-aligned splitting points and then reduces the alignment of the
 > next section accordingly:
 >
 > section		address		size		align
 > .rodata		0x22c0		0x5cfe0		64
 > .rodata.0	0x5f2a0		0xe2c80		32
 >
 > The starting address of .rodata is 64-byte-aligned, but its size is
 > only 32-byte-aligned.  The starting address of .rodata.0, which starts
 > contiguously after .rodata in the virtual address space of the ELF
 > file, is only 32-byte-aligned.  And when we get to .rodata.4, it's
 > gone down to only 8-byte alignment.
 >
 > So when the KASLR bootloader (`prekern') randomizes the address space,
 > if it respects the requested alignment but roughly uniformly
 > randomizes everything else, there's a roughly 1/2 probability that the
 > .rodata.4 section will come out misaligned for PADDQ and the kernel
 > will crash at boot.
 > [...]
 >
 
 That's a great analysis!
 
 > But the right fix is probably to convince ld to insert appropriate
 > padding in the split sections so that the alignment can be maintained
 > [...]
 >
 
 You can use SUBALIGN[1] to force an alignment on input sections:
 
 ```
 --- src/sys/arch/amd64/conf/kern.ldscript.kaslr.orig	2018-01-21 11:21:40.000000000 +0000
 +++ src/sys/arch/amd64/conf/kern.ldscript.kaslr	2026-11-08 04:26:50.878950027 +0000
 @@ -31,7 +31,7 @@
   		__rodata_hotpatch_end = . ;
   	}
 
 -	.rodata :
 +	.rodata : SUBALIGN(COHERENCY_UNIT)
   	{
   		*(.rodata)
   		*(.rodata.*)
 ```
 
 I've done this to the .rodata.* sections and it's been working for me so far:
 
 ```
 $ readelf -Wa /mnt/netbsd.KASLR | fgrep .rodata
    [18] .rodata.hotpatch  PROGBITS        0000000000000011 10e268 002382 00   A  0   0  1
    [19] .rela.rodata.hotpatch RELA            0000000000000000 1af8228 005538 18   I 151  18  8
    [20] .rodata           PROGBITS        00000000000023c0 1105f0 05f640 00   A  0   0 64
    [21] .rela.rodata      RELA            0000000000000000 1afd760 01be10 18   I 151  20  8
    [137] .rodata.0         PROGBITS        0000000000061a00 10ea928 0e2fc0 00   A  0   0 64
    [138] .rela.rodata.0    RELA            0000000000000000 2620e88 003bd0 18   I 151 137  8
    [139] .rodata.1         PROGBITS        00000000001449c0 11cd8e8 0f5bc0 00   A  0   0 64
    [140] .rela.rodata.1    RELA            0000000000000000 2624a58 012390 18   I 151 139  8
    [141] .rodata.2         PROGBITS        000000000023a580 12c34a8 108b40 00   A  0   0 64
    [142] .rela.rodata.2    RELA            0000000000000000 2636de8 07a868 18   I 151 141  8
    [143] .rodata.3         PROGBITS        00000000003430c0 13cbfe8 10cf40 00   A  0   0 64
    [144] .rela.rodata.3    RELA            0000000000000000 26b1650 040560 18   I 151 143  8
    [145] .rodata.4         PROGBITS        0000000000450000 14d8f28 1154c0 00   A  0   0 64
    [146] .rodata.5         PROGBITS        00000000005654c0 15ee3e8 041b02 00   A  0   0 64
 [...]
 $
 ```
 
 -RVP
 
 [1]: https://sourceware.org/binutils/docs/ld/Forced-Input-Alignment.html
 



Home | Main Index | Thread Index | Old Index