* Re: Linux 5.14-rc2
[not found] ` <20210719140437.GA2178213@roeck-us.net>
@ 2021-07-20 17:24 ` Linus Torvalds
2021-07-20 17:27 ` Linus Torvalds
2021-07-21 0:26 ` Akira Tsukamoto
0 siblings, 2 replies; 6+ messages in thread
From: Linus Torvalds @ 2021-07-20 17:24 UTC (permalink / raw)
To: Palmer Dabbelt; +Cc: linux-riscv
Don't know if you guys saw these mainline -rc reports:
https://lore.kernel.org/lkml/2c03b355-86dd-1fa6-dbaf-5857f4a45a66@roeck-us.net/
https://lore.kernel.org/lkml/20210719140437.GA2178213@roeck-us.net/
On Mon, Jul 19, 2021 at 7:05 AM Guenter Roeck <linux@roeck-us.net> wrote:
>
> Failed builds:
> riscv32:allmodconfig
> riscv:allmodconfig
>
> No major change here, except that arcv2 images now build.
> For reference, the riscv build problem is:
>
> cc1: error: '4496' is not a valid offset in '-mstack-protector-guard-offset='
and from the -rc1 one from Guenter:
> The riscv:allmodconfig build failure is not new. It is seen if both
> STACKPROTECTOR_PER_TASK and GCC_PLUGIN_RANDSTRUCT are enabled.
> See
> https://patchwork.kernel.org/project/linux-riscv/patch/20210706162621.940924-1-linux@roeck-us.net/
> for details and a proposed fix.
>
> riscv32 images fail to boot due to commit ca6eaaa210de ("riscv:
> __asm_copy_to-from_user: Optimize unaligned memory access and pipeline
> stall"). I reported this a couple of days ago, but have not seen a reply.
it would be nice to have these basic tests all passing again.
Linus
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Linux 5.14-rc2
2021-07-20 17:24 ` Linux 5.14-rc2 Linus Torvalds
@ 2021-07-20 17:27 ` Linus Torvalds
2021-07-21 0:26 ` Akira Tsukamoto
1 sibling, 0 replies; 6+ messages in thread
From: Linus Torvalds @ 2021-07-20 17:27 UTC (permalink / raw)
To: Palmer Dabbelt; +Cc: linux-riscv
On Tue, Jul 20, 2021 at 10:24 AM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> Don't know if you guys saw these mainline -rc reports:
>
> https://lore.kernel.org/lkml/2c03b355-86dd-1fa6-dbaf-5857f4a45a66@roeck-us.net/
> https://lore.kernel.org/lkml/20210719140437.GA2178213@roeck-us.net/
Duh. Wrong link for that first one. That was a reply in the same thread.
So the two links for rc1/rc2 test results should have been
https://lore.kernel.org/lkml/20210712015619.GA3547193@roeck-us.net/
https://lore.kernel.org/lkml/20210719140437.GA2178213@roeck-us.net/
mea culpa.
Linus
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Linux 5.14-rc2
2021-07-20 17:24 ` Linux 5.14-rc2 Linus Torvalds
2021-07-20 17:27 ` Linus Torvalds
@ 2021-07-21 0:26 ` Akira Tsukamoto
2021-07-24 0:58 ` Palmer Dabbelt
1 sibling, 1 reply; 6+ messages in thread
From: Akira Tsukamoto @ 2021-07-21 0:26 UTC (permalink / raw)
To: Linus Torvalds, Palmer Dabbelt; +Cc: linux-riscv
On 7/21/2021 2:24 AM, Linus Torvalds wrote:
> Don't know if you guys saw these mainline -rc reports:
>
> https://lore.kernel.org/lkml/2c03b355-86dd-1fa6-dbaf-5857f4a45a66@roeck-us.net/
> https://lore.kernel.org/lkml/20210719140437.GA2178213@roeck-us.net/
>
> On Mon, Jul 19, 2021 at 7:05 AM Guenter Roeck <linux@roeck-us.net> wrote:
>>
>> Failed builds:
>> riscv32:allmodconfig
>> riscv:allmodconfig
>>
>> No major change here, except that arcv2 images now build.
>> For reference, the riscv build problem is:
>>
>> cc1: error: '4496' is not a valid offset in '-mstack-protector-guard-offset='
>
> and from the -rc1 one from Guenter:
>
>> The riscv:allmodconfig build failure is not new. It is seen if both
>> STACKPROTECTOR_PER_TASK and GCC_PLUGIN_RANDSTRUCT are enabled.
>> See
>> https://patchwork.kernel.org/project/linux-riscv/patch/20210706162621.940924-1-linux@roeck-us.net/
>> for details and a proposed fix.
>>
>> riscv32 images fail to boot due to commit ca6eaaa210de ("riscv:
>> __asm_copy_to-from_user: Optimize unaligned memory access and pipeline
>> stall"). I reported this a couple of days ago, but have not seen a reply.
>
> it would be nice to have these basic tests all passing again.
I have sent the fixes yesterday in my time.
https://lists.infradead.org/pipermail/linux-riscv/2021-July/007621.html
Guenter and Geert have confirmed the fix.
https://lists.infradead.org/pipermail/linux-riscv/2021-July/007634.html
https://lists.infradead.org/pipermail/linux-riscv/2021-July/007626.html
Akira
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Linux 5.14-rc2
2021-07-21 0:26 ` Akira Tsukamoto
@ 2021-07-24 0:58 ` Palmer Dabbelt
2021-07-25 9:56 ` Akira Tsukamoto
0 siblings, 1 reply; 6+ messages in thread
From: Palmer Dabbelt @ 2021-07-24 0:58 UTC (permalink / raw)
To: akira.tsukamoto; +Cc: Linus Torvalds, linux-riscv
On Tue, 20 Jul 2021 17:26:23 PDT (-0700), akira.tsukamoto@gmail.com wrote:
> On 7/21/2021 2:24 AM, Linus Torvalds wrote:
>> Don't know if you guys saw these mainline -rc reports:
>>
>> https://lore.kernel.org/lkml/2c03b355-86dd-1fa6-dbaf-5857f4a45a66@roeck-us.net/
>> https://lore.kernel.org/lkml/20210719140437.GA2178213@roeck-us.net/
>>
>> On Mon, Jul 19, 2021 at 7:05 AM Guenter Roeck <linux@roeck-us.net> wrote:
>>>
>>> Failed builds:
>>> riscv32:allmodconfig
>>> riscv:allmodconfig
>>>
>>> No major change here, except that arcv2 images now build.
>>> For reference, the riscv build problem is:
>>>
>>> cc1: error: '4496' is not a valid offset in '-mstack-protector-guard-offset='
>>
>> and from the -rc1 one from Guenter:
>>
>>> The riscv:allmodconfig build failure is not new. It is seen if both
>>> STACKPROTECTOR_PER_TASK and GCC_PLUGIN_RANDSTRUCT are enabled.
>>> See
>>> https://patchwork.kernel.org/project/linux-riscv/patch/20210706162621.940924-1-linux@roeck-us.net/
>>> for details and a proposed fix.
>>>
>>> riscv32 images fail to boot due to commit ca6eaaa210de ("riscv:
>>> __asm_copy_to-from_user: Optimize unaligned memory access and pipeline
>>> stall"). I reported this a couple of days ago, but have not seen a reply.
>>
>> it would be nice to have these basic tests all passing again.
>
> I have sent the fixes yesterday in my time.
>
> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007621.html
>
> Guenter and Geert have confirmed the fix.
>
> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007634.html
> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007626.html
>
> Akira
Thanks. I've got these on fixes now, I'll send them up tomorrow morning
assuming the autobuilders don't find any issues. Like I said in the
other thread I wasn't getting these allmodconfig failures, so if there's
an easy way to reproduce them I'd be happy to add it to my setup -- I
don't really know anything about GCC plugins, I'm just using Debian's
toolchain for my Linux test setup (haven't gotten around to merging it
with my GCC test setup, and it's been a decade so I guess I should stop
pretending that's going to happen ;)).
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Linux 5.14-rc2
2021-07-24 0:58 ` Palmer Dabbelt
@ 2021-07-25 9:56 ` Akira Tsukamoto
2021-08-09 16:19 ` Palmer Dabbelt
0 siblings, 1 reply; 6+ messages in thread
From: Akira Tsukamoto @ 2021-07-25 9:56 UTC (permalink / raw)
To: Palmer Dabbelt; +Cc: Linus Torvalds, linux-riscv
On 7/24/21 09:58, Palmer Dabbelt wrote:
>> I have sent the fixes yesterday in my time.
>>
>> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007621.html
>>
>> Guenter and Geert have confirmed the fix.
>>
>> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007634.html
>> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007626.html
>>
>> Akira
>
> Thanks. I've got these on fixes now, I'll send them up tomorrow morning assuming the autobuilders don't find any issues. Like I said in the other thread I wasn't getting these allmodconfig failures, so if there's an easy way to reproduce them I'd be happy to add it to my setup -- I don't really know anything about GCC plugins, I'm just using Debian's toolchain for my Linux test setup (haven't gotten around to merging it with my GCC test setup, and it's been a decade so I guess I should stop pretending that's going to happen ;)).
Sure, my procedure is based on Guenter’s instructions.
I built qemu v6.0.0, buildroot 2021.05 and the kernel with the toochain came with buildroot.
# qemu
git clone https://gitlab.com/qemu-project/qemu.git
cd qemu
git checkout -b v6.0.0
git submodule update --init --recursive
git submodule sync --recursive
./configure --target-list=riscv32-softmmu --disable-user --disable-gnutls --disable-docs --disable-nettle --disable-gcrypt --disable-vnc-png --disable-xen --disable-xen-pci-passthrough --disable-libssh
make -j $(nproc)
# buildroot
git clone git://git.buildroot.net/buildroot
cd buildroot/
git checkout 2021.05 -b 2021.05
make qemu_riscv32_virt_defconfig
make
# kernel
export PATH=locaion_of_buildroot/output/host/bin:${PATH}"
Built kernel.
# prepare booting rv32
Downloaded rootfs.cpio and run.sh from Guenter’s link to inside buildroot/output/images/.
http://server.roeck-us.net/qemu/riscv32/
cd output/images
Used revised run.sh.
```
export PATH="location_of_qemu/build:${PATH}"
export KERNEL_DIR=location_of_kernel/arch/riscv/boot/
qemu-system-riscv32 -M virt -m 512M -no-reboot \
-bios default -kernel ${KERNEL_DIR}Image \
-initrd rootfs.cpio \
-append "rdinit=/sbin/init console=ttyS0,115200 earlycon=uart8250,mmio,0x10000000,115200" \
-nographic -monitor none
```
The rootfs.cpio checks the network after booting and poweroff immediately to return to shell.
Akira
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Linux 5.14-rc2
2021-07-25 9:56 ` Akira Tsukamoto
@ 2021-08-09 16:19 ` Palmer Dabbelt
0 siblings, 0 replies; 6+ messages in thread
From: Palmer Dabbelt @ 2021-08-09 16:19 UTC (permalink / raw)
To: akira.tsukamoto; +Cc: Linus Torvalds, linux-riscv
On Sun, 25 Jul 2021 02:56:51 PDT (-0700), akira.tsukamoto@gmail.com wrote:
> On 7/24/21 09:58, Palmer Dabbelt wrote:
>>> I have sent the fixes yesterday in my time.
>>>
>>> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007621.html
>>>
>>> Guenter and Geert have confirmed the fix.
>>>
>>> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007634.html
>>> https://lists.infradead.org/pipermail/linux-riscv/2021-July/007626.html
>>>
>>> Akira
>>
>> Thanks. I've got these on fixes now, I'll send them up tomorrow morning assuming the autobuilders don't find any issues. Like I said in the other thread I wasn't getting these allmodconfig failures, so if there's an easy way to reproduce them I'd be happy to add it to my setup -- I don't really know anything about GCC plugins, I'm just using Debian's toolchain for my Linux test setup (haven't gotten around to merging it with my GCC test setup, and it's been a decade so I guess I should stop pretending that's going to happen ;)).
>
> Sure, my procedure is based on Guenter’s instructions.
> I built qemu v6.0.0, buildroot 2021.05 and the kernel with the toochain came with buildroot.
>
> # qemu
> git clone https://gitlab.com/qemu-project/qemu.git
> cd qemu
> git checkout -b v6.0.0
> git submodule update --init --recursive
> git submodule sync --recursive
> ./configure --target-list=riscv32-softmmu --disable-user --disable-gnutls --disable-docs --disable-nettle --disable-gcrypt --disable-vnc-png --disable-xen --disable-xen-pci-passthrough --disable-libssh
> make -j $(nproc)
>
> # buildroot
> git clone git://git.buildroot.net/buildroot
> cd buildroot/
> git checkout 2021.05 -b 2021.05
> make qemu_riscv32_virt_defconfig
> make
>
> # kernel
> export PATH=locaion_of_buildroot/output/host/bin:${PATH}"
> Built kernel.
>
> # prepare booting rv32
> Downloaded rootfs.cpio and run.sh from Guenter’s link to inside buildroot/output/images/.
> http://server.roeck-us.net/qemu/riscv32/
>
> cd output/images
> Used revised run.sh.
> ```
> export PATH="location_of_qemu/build:${PATH}"
> export KERNEL_DIR=location_of_kernel/arch/riscv/boot/
> qemu-system-riscv32 -M virt -m 512M -no-reboot \
> -bios default -kernel ${KERNEL_DIR}Image \
> -initrd rootfs.cpio \
> -append "rdinit=/sbin/init console=ttyS0,115200 earlycon=uart8250,mmio,0x10000000,115200" \
> -nographic -monitor none
> ```
>
> The rootfs.cpio checks the network after booting and poweroff immediately to return to shell.
Thanks. I was actaully asking for the randstruct build instructions,
but Geert already provided those in the other thread. Given how easy
the 32-bit boot is now, though, I figured I'd just get it up and
running. It's in now.
Thanks!
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2021-08-09 16:20 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <CAHk-=wjYghA5O+ZsWKVNApMK3j4V1EibO5P+c4MrHnqp8YhY=g@mail.gmail.com>
[not found] ` <20210719140437.GA2178213@roeck-us.net>
2021-07-20 17:24 ` Linux 5.14-rc2 Linus Torvalds
2021-07-20 17:27 ` Linus Torvalds
2021-07-21 0:26 ` Akira Tsukamoto
2021-07-24 0:58 ` Palmer Dabbelt
2021-07-25 9:56 ` Akira Tsukamoto
2021-08-09 16:19 ` Palmer Dabbelt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).