Dmitry Vyukov

I don't remember that it's wired into dashboard/emails, so users won't see anything in either case.

Or... maybe they will. But maybe we save config on failures as well? `kconfig.Minimize` should also yield the exact list of configs that are required to reproduce the bug (on...

I can reproduce this. Using config https://syzkaller.appspot.com/text?tag=KernelConfig&x=7dfdbc03b40f61a2 I created scripts/gcc-plugins/randomize_layout_seed.h and certs/signing_key.pem" files as pkg/build/linux.go would. And then built as: ``` KBUILD_BUILD_VERSION=0 KBUILD_BUILD_TIMESTAMP=now KBUILD_BUILD_USER=syzkaller KBUILD_BUILD_HOST=syzkaller KERNELVERSION=syzkaller LOCALVERSION=-syzkaller make ``` Build...

`.rodata` diff is 1 byte: ``` < 01959d40 32 34 00 31 20 34 38 20 34 38 20 34 20 64 61 74 |24.1 48 48 4 dat| ---...

So far I can't understand what .rodata symbol is this... Section headers: ``` [Nr] Name Type Address Offset Size EntSize Flags Link Info Align [ 2] .rodata PROGBITS ffffffff89400000 08600000...

for `.data` section I also can't figure out the changed symbol: ```` ffffffff8c001300 0000000000000040 d channel_2ghz ffffffff8c0045a0 0000000000000008 D d_level_size_i2400m ... ffffffff8cc44be0 0000000000000010 d rate_ctrl_algs ffffffff8cc52080 0000000000000060 d _rs.1 ```...

Trying with LD=ld.lld (LLD 12.0.0) is no better: ``` 1,2c1,2 < .text 3f12ce9c2574c43b16e4181f2481639a9a372edc8467990ce3722743ca8314a5 < .rodata dd98517c3ef09c5088a91b2e2f6237eb95e5d10a977ea0033dbdf935b3377740 --- > .text bc79a929983aedaa3f742ae5886abf85b62b48b363eab438ba7c1957252a78ea > .rodata ebe30ad0a9b1e5b083ed7f5394988c65428b469b82fa5214de5e4c2d4104990e 23,24c23,24 < .data 48b2b5d589172cf3f543f606d2ba5b789f50c60d358d69cc2f65913ec20d81cf < __bug_table 0ac21582eb0d90b77cf2f8958922240c477cd0c568deb192248b80f00f29e119...

This all makes me sad. One of these things we should not be spending time on at all...

Another example of non-stable build: https://syzkaller.appspot.com/bug?extid=830c6dbfc71edc4f0b8f "WARNING in dev_change_net_namespace" issue is detected on x86 instances, but bisected to "[arm64: cpufeature: Remove redundant call to id_aa64pfr0_32bit_el0()](https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=13dc4d836179444f0ca90188cfccd23f9cd9ff05)" which only touches arch/arm64.

I can't reproduce the previous case locally. Running: ``` $ syz-build -config .config -kernel_src /src/linux -trace -userspace ../.config -vm=qemu -compiler=gcc-11 ``` Gives me on both commits: ``` signature: d088522a56053ecbe7df3496a2dab349e7cc16f3f46964a888288f7aa1dfaf8f compiler:...