lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d2a09bf3-9bd3-588a-99a1-598281d08678@kernel.org>
Date:   Wed, 9 Mar 2022 06:48:49 +0100
From:   Jiri Slaby <jirislaby@...nel.org>
To:     Helge Deller <deller@....de>,
        "James E.J. Bottomley" <James.Bottomley@...senPartnership.com>,
        linux-parisc@...r.kernel.org,
        Linux kernel mailing list <linux-kernel@...r.kernel.org>,
        Martin Liška <mliska@...e.cz>,
        Andreas Schwab <schwab@...ux-m68k.org>
Subject: Re: hppa vDSO and compiler (non-)support

Hi,

On 08. 03. 22, 15:51, Helge Deller wrote:
> Hi Jiri,
> 
> Thanks for testing on parisc!
> 
> On 3/8/22 12:06, Jiri Slaby wrote:
>> since the "parisc: Add vDSO support" commit, I can no longer cross-build a hppa kernel. I see two issues:
>>
>> 1) CROSS32_COMPILE detection doesn't work here, as openSUSE provides hppa-suse-linux-* binaries. It's easy to overcome by "CROSS32_COMPILE=hppa-suse-linux-"
> 
> How is it handled for other platforms like s390x?

s390 simply uses CC for vdso32:
cmd_vdso32cc = $(CC) $(c_flags) -c -o $@ $<

> Would it make sense to add the detection for SUSE too?

Maybe.

>> 2) openSUSE doesn't provide any libc for hppa. So gcc doesn't provide libgcc.a and the build of vDSO fails.
> 
> libgcc.a comes with the compiler, I don't think you need libc for that.

I was told glibc is needed to build libgcc.a.

> I'm currently installing opensuse to try myself though...
> 
>> So could vDSO be optional on hppa via KConfig?
> The vDSO is one of the first things which is built during kernel build process.
> This is why you fail.
> Making it optional doesn't make sense, because then the kernel wouldn't be able
> to start the user space processes.
> 
>> I used to use the cross compiler to at least compile-check the following  tty drivers:
>> arch/parisc/kernel/pdc_cons.o
>> drivers/tty/serial/mux.o
>> drivers/tty/serial/8250/8250_gsc.o
> 
> I assume you never built a full kernel, but stopped when building those modules?
> Without libgcc.a the kernel itself wouldn't have linked before either.

Correct, I am only build-testing parisc-only drivers when doing global 
tty changes.

thanks,
-- 
js
suse labs

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ