On 2026/2/4 下午4:57, John Paul Adrian Glaubitz wrote:
Hi Sergio,
On Fri, 2026-01-30 at 16:27 -0500, Sergio Durigan Junior wrote:
Tiezhu Yang <[email protected]> writes:
On 2026/1/24 下午2:57, Tiezhu Yang wrote:
On 2026/1/24 下午2:11, Miao Wang wrote:
...
I do not know the details of the builder environment.
Anyway, as I described in the previous reply [1], tested with the
official Debian ISO [2] on a physical machine, it should support
the hardware breakpoints, the testcases should pass. If tested on
a virtual machine, it needs to adjust the testcases to make them
happy.
Let us wait for the loongarch64 builders to be redeployed and
see if that solves the issue [3].
I discussed offline with Dandan Zhang, she confirmed that the builder
environment is a virtual machine, so the failures is because there is
no support for hardware breakpoints on a virtual machine of LoongArch.
I discussed offline with Miao Wang to reach an agreement, and then I
sent a formal patch [1] to upstream gdb maillist to fix the failures,
it will be merged into upstream master branch and gdb-17-branch ASAP,
then this patch can be backported to debian gdb.
[1]
https://inbox.sourceware.org/gdb-patches/[email protected]/
Thanks, that's great. I'm following the upstream discussions and will
backport the patch as soon as it's accepted there.
This bug is blocking a lot of packages on loong64 and also makes creating a
fresh
chroot on the porterbox impossible.
Is there a way that this fix could be expedited?
Cc: Upstream GDB Global Maintainer Tom Tromey <[email protected]>
The upstream patch is still under review, one way is to disable the
failed testcase in debian GDB for now if possible.
We can enable the testcase once the patch is merged into mainline
and backported to debian GDB.
Here is the v2 patch:
https://inbox.sourceware.org/gdb-patches/[email protected]/
Thanks,
Tiezhu
