[CRIU] [PATCHv3 0/7] Large pages support for aarch64/ppc64

Dmitry Safonov 0x7f454c46 at gmail.com
Mon Apr 23 18:16:33 MSK 2018


2018-04-20 7:56 GMT+01:00 Adrian Reber <areber at redhat.com>:
> On Fri, Apr 20, 2018 at 12:42:34AM +0100, Dmitry Safonov wrote:
>> 2018-04-12 15:52 GMT+01:00 Dmitry Safonov <0x7f454c46 at gmail.com>:
>> > 2018-04-12 14:57 GMT+01:00 Adrian Reber <areber at redhat.com>:
>> >> On Mon, Apr 09, 2018 at 05:46:29PM +0100, Dmitry Safonov wrote:
>> >>> > ppc64le:
>> >>> >
>> >>> > ################### 3 TEST(S) FAILED (TOTAL 353/SKIPPED 35) ####################
>> >>> >  * zdtm/static/aio00(h)
>> >>> >  * zdtm/static/aio01(h)
>> >>> >  * zdtm/static/maps06(h)
>> >>> > ##################################### FAIL #####################################
>> >>>
>> >>> Are these regressions? Were they failing before the patches set?
>> >>
>> >> Without your patches only zdtm/static/maps06 fails. So the aio0?
>> >> failures can be related to your patches. And maps06 is one of the test
>> >> cases you identified as hard coding 4096.
>> >
>> > Sure, I'll look into the fails..
>>
>> Hi Adrian,
>>
>> Could you send me your aarch64 kernel's .config?
>> I've tried to reproduce aio00/aio01 failure with 4K/16K/64K pages
>> on qemu-system-aarch64, but no luck, tests are passing in (h) :-/
>
> My kernel config must have been pretty close to that:
>
> https://git.centos.org/blob/rpms!kernel-alt.git/4c79b75ab8381328e6ea3712b3430fe6699412e6/SOURCES!kernel-alt-4.14.0-aarch64.config

Heh, I've tried to reproduce this with that config (enabling every config option
also mentioned https://criu.org/Linux_kernel ), with mainline v4.16 kernel, but
still aio* tests passes in qemu-system-aarch64 =/

Could you run the test with inserted printing?:
--->8---
--- a/criu/pie/restorer.c
+++ b/criu/pie/restorer.c
@@ -700,6 +700,7 @@ static int restore_aio_ring(struct rst_aio_ring *raio)
        ret = sys_io_setup(raio->nr_req, &ctx);
        if (ret < 0) {
                pr_err("Ring setup failed with %d\n", ret);
+               pr_err("nr_req=%u\n", raio->nr_req);
                return -1;
        }

-- 
             Dmitry


More information about the CRIU mailing list