Re: Have my PA8800 back online...
On 2017-12-03, at 1:31 PM, Carlos O'Donell wrote:
> Helge, Dave,
>
> After a disk failure, and some memory corruption (and reseating
> memory), I have my hppa back online and building glibc again. Though
> with 8GiB of RAM, and I'll have to test more sticks to see if I can
> get this back to 12GiB. All 4 CPUs are operating fine at a claimed
> 1GHz.
Excellent!
>
> I've been away from the port for so long, but I'd like to contribute
> again. I'll have a look at the glibc master test results. Anything
> else that's critical to the port that needs help?
That is wonderful news.
>
> My usual process for this is:
> * Play with core tools and linux kernel.
> * Fix anything that looks wrong.
The fails from my last glibc build were:
FAIL: debug/tst-backtrace2
FAIL: debug/tst-backtrace3
FAIL: debug/tst-backtrace4
FAIL: debug/tst-backtrace5
FAIL: debug/tst-backtrace6
FAIL: elf/check-execstack
FAIL: elf/tst-audit2
FAIL: intl/tst-gettext
FAIL: nptl/tst-cleanupx4
FAIL: stdlib/tst-setcontext2
Summary of test results:
10 FAIL
3755 PASS
18 XFAIL
The tst-audit2 failure is a segv. I thought that would be easy to fix but the plugin messes up
gdb, etc.
At the the moment, gcc-8 is in stage 3 and I'm trying to go through the test failures.
Regarding the kernel, the v4.13 series was pretty good on my rp3440. I've only run the v4.14.2 and
v4.14.3 for about 3 days. It seems the number of random segvs has increased again and I had one
stall:
INFO: task updatedb.mlocat:5610 blocked for more than 120 seconds.
Not tainted 4.14.3+ #1
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
updatedb.mlocat D 0 5610 5607 0x00000010
Backtrace:
[<00000000401636f0>] __schedule+0x318/0x650
[<0000000040163a84>] schedule+0x5c/0xe8
[<0000000040168f68>] schedule_timeout+0x1b8/0x218
[<0000000040164c48>] wait_for_common+0x140/0x228
[<0000000040164d54>] wait_for_completion+0x24/0x38
[<000000001cd27628>] xfs_buf_submit_wait+0x130/0x1e0 [xfs]
[<000000001cd2771c>] _xfs_buf_read+0x44/0xc8 [xfs]
[<000000001cd278a0>] xfs_buf_read_map+0x100/0x158 [xfs]
[<000000001cd7c890>] xfs_trans_read_buf_map+0xf8/0x2c8 [xfs]
[<000000001cd07918>] xfs_imap_to_bp+0x78/0x140 [xfs]
[<000000001cd08338>] xfs_iread+0xa8/0x2f8 [xfs]
[<000000001cd372b4>] xfs_iget+0x374/0xa28 [xfs]
[<000000001cd4603c>] xfs_lookup+0xac/0x160 [xfs]
[<000000001cd40708>] xfs_vn_lookup+0x70/0x110 [xfs]
[<0000000040304f90>] lookup_slow+0x108/0x248
Helge and I track the Debian buildd status here:
https://buildd.debian.org/status/architecture.php?a=hppa&suite=sid
When we get time, we file packages bugs and some fixes, particularly for core tools.
Helge looks after the phantom, panama and parisc buildds. I look after mx3210.
We have accounts at wuiet.debian.org which allow us to give-back packages that fail due to random
build failures, etc.
We communicate offline frequently on kernel issues, etc.
Alan Modra has been very helpful in maintaining binutils on hppa.
Dave
--
John David Anglin dave.anglin@bell.net
Reply to: