> >Maybe you can apply it?
>
> Uploaded, thanks.
Thank you!
I noticed it failed to build for you.
That's my fault and sorry for that.
I missed the hunk for select.
I locally had patched select, but the temporary original file was named
"select.S,org" (note the comma!) instead of "select.S.org", s
Helge Deller dixit:
>* Thorsten Glaser :
>>
>> Maybe other 5/6-argument syscalls also need review…
>
>Yes, and basically I tink it's stupid to try to distinguish syscalls
>with 1-4 arguments from those which use 5 or 6 arguments.
>With every added syscall someone needs to check again.
Agreed.
>T
* Thorsten Glaser :
> > So, instead of adding and ifdef for hppa/syscall6_weak, simply
> > drop the weak function and call the normal __unified_syscall6.
>
> Looking at the nm output of libc.a, it seems to use the 6 version
> for sendto but not recvfrom. Huh.
>
> Maybe other 5/6-argument syscalls
Helge Deller dixit:
> The attached patch fixes the dietlibc testsuite crash on hppa
> in the socketfns testcase.
Thanks, applied.
> So, instead of adding and ifdef for hppa/syscall6_weak, simply
> drop the weak function and call the normal __unified_syscall6.
Looking at the nm output of libc.a,
The attached patch fixes the dietlibc testsuite crash on hppa
in the socketfns testcase.
The problem is with recvfrom(), which takes 6 arguments.
The 5th and 6th argument is pushed on the stack by gcc, and
on hppa the argument is moved from the stack into a register
in __unified_syscall5 and __un
Helge Deller dixit:
> Beside the kernel & glibc patches some qemu-user patches will probably
> be necessary as well.
I think so, given how it basically interposes for both.
> I think disabling such tests is probably best right now.
OK. I’ll try debootstrapping an hppa chroot in qemu and see
if
On 7/5/24 17:28, Thorsten Glaser wrote:
Helge Deller dixit:
I just did a give-back, and this time the build succeeded (on the
"c8000" buildd server). The last build failed on the "pasta" buildd
server, which is a qemu-user based build machine, while "c8000" is
a physical server which runs the *
Helge Deller dixit:
> I just did a give-back, and this time the build succeeded (on the
> "c8000" buildd server). The last build failed on the "pasta" buildd
> server, which is a qemu-user based build machine, while "c8000" is
> a physical server which runs the *very latest* stable native
> Linux
On 7/5/24 15:36, Thorsten Glaser wrote:
Helge Deller dixit:
Ok, so we have to wait until at least the oldstable kernels are "new" enough.
Given hppa is a ports architecture and this had some time to percolate,
I applied the patches now.
Thank you!
However, I get a (possibly unrelated) bui
Helge Deller dixit:
> Ok, so we have to wait until at least the oldstable kernels are "new" enough.
Given hppa is a ports architecture and this had some time to percolate,
I applied the patches now.
However, I get a (possibly unrelated) build failure on hppa now
(dietlibc in experimental): debia
Helge Deller dixit:
> Is it ok to keep this bug report open for - let's say -
> a year or so, and then apply it.
Sure, why not?
bye,
//mirabilos
--
15:41⎜ Somebody write a testsuite for helloworld :-)
On 2/10/23 16:16, Thorsten Glaser wrote:
Helge Deller dixit:
No, it needs to have been in the release.
Please help me to understand the dependencies you mention.
That patch is targetted for dietlibc in unstable.
Is dietlibc/unstable shipped backwards into oldstable (in which
No, but Debian
Helge Deller dixit:
>> No, it needs to have been in the release.
>
> Please help me to understand the dependencies you mention.
> That patch is targetted for dietlibc in unstable.
> Is dietlibc/unstable shipped backwards into oldstable (in which
No, but Debian supports partial upgrades in both di
On 2/10/23 15:57, Thorsten Glaser wrote:
Helge Deller dixit:
You know that new kernel features have to be in oldstable before
they can be depended on, right?
The kernel patch was backported and is in upstream Linux kernel version 6.2,
alternatively stable kernels from versions 6.1.6, 5.15.87,
Helge Deller dixit:
>> You know that new kernel features have to be in oldstable before
>> they can be depended on, right?
>
> The kernel patch was backported and is in upstream Linux kernel version 6.2,
> alternatively stable kernels from versions 6.1.6, 5.15.87, 5.10.163,
> 5.4.228, 4.19.270 or
On 2/10/23 15:27, Thorsten Glaser wrote:
Helge Deller dixit:
This upstream patch has been downported to all stable kernel trees as
well.
Please apply for next upload.
You know that new kernel features have to be in oldstable before
they can be depended on, right?
The kernel patch was bac
Helge Deller dixit:
>This upstream patch has been downported to all stable kernel trees as
>well.
>Please apply for next upload.
You know that new kernel features have to be in oldstable before
they can be depended on, right?
bye,
//mirabilos
--
(gnutls can also be used, but if you are compili
Package: dietlibc
Tags: hppa, patch
Version: 0.34~cvs20160606-14
The Linux kernel upstream commit 71bdea6f798b ("parisc: Align parisc
MADV_XXX constants with all other architectures") dropped the
parisc-specific MADV_* values in favour of the same constants as
other architectures. In the same com
18 matches
Mail list logo