Total jobs: 205
Total errors: 58 (28.29%)
LAVA errors: 0 (0.00%)
Test errors: 31 (15.12%)
Job errors: 22 (10.73%)
Infra errors: 5 (2.44%)
Canceled jobs: 0 (0.00%)
Device type: sm8550-hdk
Total jobs: 21
Total errors: 21 (100.00%)
Error type: Job
Error count: 21 (100.00%)
Error: No match for error type 'Job', message 'login-action timed out after 875 seconds'
Count: 12 (57.14%)
IDs:
sm8550-hdk-01:
8017416 8017418 8017422 8017423 8017427
8017453 8017454 8019771 8019776 8020465
8020508 8020511
Error: No match for error type 'Job', message 'login-action timed out after 876 seconds'
Count: 1 (4.76%)
IDs:
sm8550-hdk-01:
8020468
Error: wait for prompt timed out
Count: 7 (33.33%)
IDs:
sm8550-hdk-01:
8017421 8017431 8017442 8017452 8017455
8019773 8019774
Error: No match for error type 'Job', message 'login-action timed out after 870 seconds'
Count: 1 (4.76%)
IDs:
sm8550-hdk-01:
8019772
Device type: hi6220-hikey-r2
Total jobs: 29
Total errors: 16 (55.17%)
Error type: Test
Error count: 13 (44.83%)
Error: No match for error type 'Test', message 'tradefed - adb device lost[235D989C003B0752]'
Count: 5 (17.24%)
IDs:
hikey-6220-r2-03:
8017751 8019520 8019576 8019947 8020467
Error: No match for error type 'Test', message 'tradefed - adb device lost[7B1579B5003E507A]'
Count: 1 (3.45%)
IDs:
hikey-6220-r2-05:
8019961
Error: No match for error type 'Test', message 'The network seems not available, as the ping command failed'
Count: 2 (6.90%)
IDs:
hikey-6220-r2-05:
8019584 8019946
Error: No match for error type 'Test', message 'tradefed - adb device lost[191FC75C003312EE]'
Count: 3 (10.34%)
IDs:
hikey-6220-r2-06:
8017400 8017517 8019579
Error: No match for error type 'Test', message 'tradefed - adb device lost[4595FE84003F72FC]'
Count: 1 (3.45%)
IDs:
hikey-6220-r2-04:
8017590
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 28798 seconds'
Count: 1 (3.45%)
IDs:
hikey-6220-r2-05:
8017456
Error type: Infrastructure
Error count: 2 (6.90%)
Error: No match for error type 'Infrastructure', message 'wait-device-boardid timed out after 888 seconds'
Count: 1 (3.45%)
IDs:
hikey-6220-r2-06:
8020458
Error: No match for error type 'Infrastructure', message 'wait-device-boardid timed out after 1171 seconds'
Count: 1 (3.45%)
IDs:
hikey-6220-r2-06:
8019577
Error type: Job
Error count: 1 (3.45%)
Error: No match for error type 'Job', message 'login-action timed out after 880 seconds'
Count: 1 (3.45%)
IDs:
hikey-6220-r2-01:
8017457
Device type: rk3399-rock-pi-4b
Total jobs: 2
Total errors: 0 (0.00%)
Device type: qrb5165-rb5
Total jobs: 73
Total errors: 10 (13.70%)
Error type: Test
Error count: 8 (10.96%)
Error: Device NOT found!
Count: 2 (2.74%)
IDs:
rb5-03:
8017738 8019745
Error: No match for error type 'Test', message 'tradefed - adb device lost[d8a1879f]'
Count: 1 (1.37%)
IDs:
rb5-07:
8019714
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 598 seconds'
Count: 4 (5.48%)
IDs:
rb5-01:
8017623
rb5-03:
8017529 8019677 8019680
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 581 seconds'
Count: 1 (1.37%)
IDs:
rb5-07:
8019565
Error type: Infrastructure
Error count: 2 (2.74%)
Error: fastboot-flash-action timed out
Count: 2 (2.74%)
IDs:
rb5-07:
8017688 8017703
Device type: x15
Total jobs: 1
Total errors: 1 (100.00%)
Error type: Test
Error count: 1 (100.00%)
Error: lava-test-shell timed out
Count: 1 (100.00%)
IDs:
x15-06:
8018361
Device type: dragonboard-845c
Total jobs: 57
Total errors: 10 (17.54%)
Error type: Test
Error count: 9 (15.79%)
Error: No match for error type 'Test', message 'tradefed - adb device lost[c7cf1849]'
Count: 1 (1.75%)
IDs:
db845c-07:
8017549
Error: No match for error type 'Test', message 'tradefed - adb device lost[b6b742b9]'
Count: 2 (3.51%)
IDs:
db845c-01:
8017481 8017547
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 594 seconds'
Count: 1 (1.75%)
IDs:
db845c-07:
8017510
Error: No match for error type 'Test', message 'The network seems not available, as the ping command failed'
Count: 3 (5.26%)
IDs:
db845c-04:
8017505 8017508
db845c-08:
8017507
Error: No match for error type 'Test', message 'lava-test-interactive timed out after 600 seconds'
Count: 1 (1.75%)
IDs:
db845c-03:
8017488
Error: Device NOT found!
Count: 1 (1.75%)
IDs:
db845c-09:
8017168
Error type: Infrastructure
Error count: 1 (1.75%)
Error: fastboot-flash-action timed out
Count: 1 (1.75%)
IDs:
db845c-01:
8017525
Device type: hi960-hikey
Total jobs: 22
Total errors: 0 (0.00%)
Hello,
On Fri, 6 Dec 2024, David Laight wrote:
> From: Julian Anastasov
> > Sent: 06 December 2024 16:23
> ...
> > I'm not sure how much memory we can see in small system,
> > IMHO, problem should not be possible in practice:
> >
> > - nobody expects 0 from totalram_pages() in the code
> >
> > - order_base_2(sizeof(struct ip_vs_conn)) is probably 8 on 32-bit
>
> It is 0x120 bytes on 64bit, so 8 could well be right.
That is already for 9 :)
> > > Is all stems from order_base_2(totalram_pages()).
> > > order_base_2(n) is 'n > 1 ? ilog2(n - 1) + 1 : 0'.
> > > And the compiler generates two copies of the code that follows
> > > for the 'constant zero' and ilog2() values.
> > > And the 'zero' case compiles clamp(20, 8, 0) which is errored.
> > > Note that it is only executed if totalram_pages() is zero,
> > > but it is always compiled 'just in case'.
> >
> > I'm confused with these compiler issues,
>
> The compiler is just doing its job.
> Consider this expression:
> (x >= 1 ? 2 * x : 1) - 1
> It is likely to get converted to:
> (x >= 1 ? 2 * x - 1 : 0)
> to avoid the subtract when x < 1.
>
> The same thing is happening here.
> order_base_2() has a (condition ? fn() : 0) in it.
> All the +/- constants get moved inside, on 64bit that is +12 -2 -1 -9 = 0.
> Then the clamp() with constants gets moved inside:
> (condition ? clamp(27, 8, fn() + 0) : clamp(27, 8, 0 + 0))
> Now, at runtime, we know that 'condition' is true and (fn() >= 8)
> so the first clamp() is valid and the second one never used.
> But this isn't known by the compiler and clamp() detects the invalid
> call and generates a warning.
I see, such optimizations are beyond my expectations,
I used max_avail var to separate the operations between
different macro calls but in the end they are mixed together...
> > if you
> > think we should go with the patch just decide if it is a
> > net or net-next material. Your change is safer for bad
> > max_avail values but I don't expect to see problem while
> > running without the change, except the building bugs.
> >
> > Also, please use nf/nf-next tag to avoid any
> > confusion with upstreaming...
>
> I've copied Andrew M - he's taken the minmax.h change into his mm tree.
> This is one of the build breakages.
>
> It probably only needs to go into next for now (via some route).
> But I can image the minmax.h changes getting backported a bit.
OK, then can you send v2 with Fixes header, precised comments
and nf tag, it fixes a recent commit, so it can be backported easily...
Regards
--
Julian Anastasov <ja(a)ssi.bg>
Hello,
On Fri, 6 Dec 2024, David Laight wrote:
> From: Julian Anastasov
> > Sent: 06 December 2024 12:19
> >
> > On Fri, 6 Dec 2024, David Laight wrote:
> >
> > > The intention of the code seems to be that the minimum table
> > > size should be 256 (1 << min).
> > > However the code uses max = clamp(20, 5, max_avail) which implies
> >
> > Actually, it tries to reduce max=20 (max possible) below
> > max_avail: [8 .. max_avail]. Not sure what 5 is here...
>
> Me mistyping values between two windows :-)
>
> Well min(max, max_avail) would be the reduced upper limit.
> But you'd still fall foul of the compiler propagating the 'n > 1'
> check in order_base_2() further down the function.
>
> > > the author thought max_avail could be less than 5.
> > > But clamp(val, min, max) is only well defined for max >= min.
> > > If max < min whether is returns min or max depends on the order of
> > > the comparisons.
> >
> > Looks like max_avail goes below 8 ? What value you see
> > for such small system?
>
> I'm not, but clearly you thought the value could be small otherwise
> the code would only have a 'max' limit.
> (Apart from a 'sanity' min of maybe 2 to stop the code breaking.)
I'm not sure how much memory we can see in small system,
IMHO, problem should not be possible in practice:
- nobody expects 0 from totalram_pages() in the code
- order_base_2(sizeof(struct ip_vs_conn)) is probably 8 on 32-bit
- PAGE_SHIFT: 12 (for 4KB) or more?
So, if totalram_pages() returns below 128 pages (4KB each)
max_avail will be below 19 (7 + 12), then 19 is reduced with 2 + 1
and becomes 16, finally with 8 (from the 2nd order_base_2) to reach
16-8=8. You need a system with less than 512KB (19 bits) to trigger
problem in clamp() that will lead to max below 8. Further, without
checks, for ip_vs_conn_tab_bits=1 we need totalram_pages() to return 0
pages.
> > > Detected by compile time checks added to clamp(), specifically:
> > > minmax.h: use BUILD_BUG_ON_MSG() for the lo < hi test in clamp()
> >
> > Existing or new check? Does it happen that max_avail
> > is a constant, so that a compile check triggers?
>
> Is all stems from order_base_2(totalram_pages()).
> order_base_2(n) is 'n > 1 ? ilog2(n - 1) + 1 : 0'.
> And the compiler generates two copies of the code that follows
> for the 'constant zero' and ilog2() values.
> And the 'zero' case compiles clamp(20, 8, 0) which is errored.
> Note that it is only executed if totalram_pages() is zero,
> but it is always compiled 'just in case'.
I'm confused with these compiler issues, if you
think we should go with the patch just decide if it is a
net or net-next material. Your change is safer for bad
max_avail values but I don't expect to see problem while
running without the change, except the building bugs.
Also, please use nf/nf-next tag to avoid any
confusion with upstreaming...
Regards
--
Julian Anastasov <ja(a)ssi.bg>
Hello,
On Fri, 6 Dec 2024, David Laight wrote:
> The intention of the code seems to be that the minimum table
> size should be 256 (1 << min).
> However the code uses max = clamp(20, 5, max_avail) which implies
Actually, it tries to reduce max=20 (max possible) below
max_avail: [8 .. max_avail]. Not sure what 5 is here...
> the author thought max_avail could be less than 5.
> But clamp(val, min, max) is only well defined for max >= min.
> If max < min whether is returns min or max depends on the order of
> the comparisons.
Looks like max_avail goes below 8 ? What value you see
for such small system?
> Change to clamp(max_avail, 5, 20) which has the expected behaviour.
It should be clamp(max_avail, 8, 20)
>
> Replace the clamp_val() on the line below with clamp().
> clamp_val() is just 'an accident waiting to happen' and not needed here.
OK
> Fixes: 4f325e26277b6
> (Although I actually doubt the code is used on small memory systems.)
>
> Detected by compile time checks added to clamp(), specifically:
> minmax.h: use BUILD_BUG_ON_MSG() for the lo < hi test in clamp()
Existing or new check? Does it happen that max_avail
is a constant, so that a compile check triggers?
>
> Signed-off-by: David Laight <david.laight(a)aculab.com>
The code below looks ok to me but can you change the
comments above to more correctly specify the values and if the
problem is that max_avail goes below 8 (min).
> ---
> net/netfilter/ipvs/ip_vs_conn.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/net/netfilter/ipvs/ip_vs_conn.c b/net/netfilter/ipvs/ip_vs_conn.c
> index 98d7dbe3d787..c0289f83f96d 100644
> --- a/net/netfilter/ipvs/ip_vs_conn.c
> +++ b/net/netfilter/ipvs/ip_vs_conn.c
> @@ -1495,8 +1495,8 @@ int __init ip_vs_conn_init(void)
> max_avail -= 2; /* ~4 in hash row */
> max_avail -= 1; /* IPVS up to 1/2 of mem */
> max_avail -= order_base_2(sizeof(struct ip_vs_conn));
More likely we can additionally clamp max_avail here:
max_avail = max(min, max_avail);
But your solution solves the problem with less lines.
> - max = clamp(max, min, max_avail);
> - ip_vs_conn_tab_bits = clamp_val(ip_vs_conn_tab_bits, min, max);
> + max = clamp(max_avail, min, max);
> + ip_vs_conn_tab_bits = clamp(ip_vs_conn_tab_bits, min, max);
> ip_vs_conn_tab_size = 1 << ip_vs_conn_tab_bits;
> ip_vs_conn_tab_mask = ip_vs_conn_tab_size - 1;
>
> --
> 2.17.1
Regards
--
Julian Anastasov <ja(a)ssi.bg>
Total jobs: 281
Total errors: 101 (35.94%)
LAVA errors: 1 (0.36%)
Test errors: 73 (25.98%)
Job errors: 19 (6.76%)
Infra errors: 3 (1.07%)
Canceled jobs: 5 (1.78%)
Device type: sm8550-hdk
Total jobs: 19
Total errors: 19 (100.00%)
Error type: Job
Error count: 19 (100.00%)
Error: No match for error type 'Job', message 'login-action timed out after 875 seconds'
Count: 8 (42.11%)
IDs:
sm8550-hdk-01:
8015974 8015975 8015984 8016008 8016034
8017224 8017227 8017414
Error: wait for prompt timed out
Count: 5 (26.32%)
IDs:
sm8550-hdk-01:
8015983 8016023 8016324 8017223 8017402
Error: No match for error type 'Job', message 'login-action timed out after 868 seconds'
Count: 1 (5.26%)
IDs:
sm8550-hdk-01:
8017399
Error: No match for error type 'Job', message 'login-action timed out after 861 seconds'
Count: 1 (5.26%)
IDs:
sm8550-hdk-01:
8017226
Error: No match for error type 'Job', message 'login-action timed out after 860 seconds'
Count: 1 (5.26%)
IDs:
sm8550-hdk-01:
8017225
Error: No match for error type 'Job', message 'login-action timed out after 874 seconds'
Count: 2 (10.53%)
IDs:
sm8550-hdk-01:
8016004 8016022
Error: No match for error type 'Job', message 'login-action timed out after 851 seconds'
Count: 1 (5.26%)
IDs:
sm8550-hdk-01:
8016021
Device type: qrb5165-rb5
Total jobs: 78
Total errors: 17 (21.79%)
Error type: Test
Error count: 12 (15.38%)
Error: No match for error type 'Test', message 'tradefed - adb device lost[f11b68eb]'
Count: 1 (1.28%)
IDs:
rb5-05:
8017408
Error: Device NOT found!
Count: 1 (1.28%)
IDs:
rb5-03:
8017353
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 590 seconds'
Count: 1 (1.28%)
IDs:
rb5-03:
8017173
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 587 seconds'
Count: 1 (1.28%)
IDs:
rb5-07:
8016373
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 297 seconds'
Count: 1 (1.28%)
IDs:
rb5-03:
8016369
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 598 seconds'
Count: 4 (5.13%)
IDs:
rb5-03:
8016340 8016343 8016356 8016367
Error: No match for error type 'Test', message 'tradefed - adb device lost[4fc7b22]'
Count: 1 (1.28%)
IDs:
rb5-06:
8015982
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 597 seconds'
Count: 1 (1.28%)
IDs:
rb5-03:
8015968
Error: No match for error type 'Test', message 'tradefed - adb device lost[74d67c95]'
Count: 1 (1.28%)
IDs:
rb5-03:
8015920
Error type: Infrastructure
Error count: 2 (2.56%)
Error: Connection closed
Count: 1 (1.28%)
IDs:
rb5-05:
8016436
Error: No match for error type 'Infrastructure', message 'Unable to run: ['docker', 'exec', 'lava-8015998-3.1-63c4cc77-7251-45ac-a948-b6ee68e2f18e', 'fastboot', '-s', 'f11b68eb', 'oem', 'select-display-panel', 'hdmi']'
Count: 1 (1.28%)
IDs:
rb5-05:
8015998
Error type: Canceled
Error count: 2 (2.56%)
Error: The job was canceled
Count: 2 (2.56%)
IDs:
rb5-06:
8016001 8016017
Error type: Bug
Error count: 1 (1.28%)
Error: No match for error type 'Bug', message 'Failed to share device '/dev/bus/usb/002/033' to docker container lava-8015995-3.1-869b7f70-7ed4-4d42-85f8-ca9bb226c356'
Count: 1 (1.28%)
IDs:
rb5-05:
8015995
Device type: hi6220-hikey-r2
Total jobs: 71
Total errors: 47 (66.20%)
Error type: Test
Error count: 47 (66.20%)
Error: No match for error type 'Test', message 'The network seems not available, as the ping command failed'
Count: 36 (50.70%)
IDs:
hikey-6220-r2-01:
8016722 8016728 8016730 8016731 8016735
8016736 8016738 8017041 8017049 8017059
8017063 8017080 8017403 8017405
hikey-6220-r2-05:
8016737 8016743 8016746 8016748 8016749
8016750 8016752 8016755 8016757 8016785
8016786 8016801 8017043 8017050 8017062
8017067 8017086 8017087 8017367 8017372
8017379 8017383
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 597 seconds'
Count: 1 (1.41%)
IDs:
hikey-6220-r2-06:
8017375
Error: No match for error type 'Test', message 'tradefed - adb device lost[7B1579B5003E507A]'
Count: 3 (4.23%)
IDs:
hikey-6220-r2-05:
8016726 8017070 8017374
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 598 seconds'
Count: 1 (1.41%)
IDs:
hikey-6220-r2-03:
8017363
Error: No match for error type 'Test', message 'tradefed - adb device lost[4595FE84003F72FC]'
Count: 2 (2.82%)
IDs:
hikey-6220-r2-04:
8016800 8016804
Error: Device NOT found!
Count: 2 (2.82%)
IDs:
hikey-6220-r2-03:
8016751
hikey-6220-r2-04:
8016803
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 586 seconds'
Count: 1 (1.41%)
IDs:
hikey-6220-r2-01:
8016733
Error: No match for error type 'Test', message 'tradefed - adb device lost[191FC75C003312EE]'
Count: 1 (1.41%)
IDs:
hikey-6220-r2-06:
8016723
Device type: dragonboard-845c
Total jobs: 101
Total errors: 18 (17.82%)
Error type: Test
Error count: 14 (13.86%)
Error: Device NOT found!
Count: 4 (3.96%)
IDs:
db845c-09:
8015810 8016350 8017079 8017167
Error: No match for error type 'Test', message 'tradefed - adb device lost[d966d3f5]'
Count: 1 (0.99%)
IDs:
db845c-03:
8016401
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 298 seconds'
Count: 2 (1.98%)
IDs:
db845c-01:
8015934
db845c-07:
8016393
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 598 seconds'
Count: 4 (3.96%)
IDs:
db845c-09:
8015842 8015846 8015867 8016376
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 294 seconds'
Count: 1 (0.99%)
IDs:
db845c-07:
8015935
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 296 seconds'
Count: 1 (0.99%)
IDs:
db845c-03:
8015897
Error: No match for error type 'Test', message 'lava-docker-test-shell timed out after 595 seconds'
Count: 1 (0.99%)
IDs:
db845c-09:
8015816
Error type: Infrastructure
Error count: 1 (0.99%)
Error: fastboot-flash-action timed out
Count: 1 (0.99%)
IDs:
db845c-07:
8016336
Error type: Canceled
Error count: 3 (2.97%)
Error: The job was canceled
Count: 3 (2.97%)
IDs:
db845c-03:
8015994
db845c-07:
8015997
db845c-09:
8016000
Device type: hi960-hikey
Total jobs: 11
Total errors: 0 (0.00%)
Device type: rk3399-rock-pi-4b
Total jobs: 1
Total errors: 0 (0.00%)