make things happen, follow the white rabbit

sata controller for non-x86

Workstations, servers, PDAs, ... MIPS, PPC, HPPA, ARM, X86, ...

C3600, Adaptec 1210SA

Postby ivelegacy on Mon May 13, 2019 7:03 pm

burnin test, 10 hours

script for the test
Code: Select all
while [ 1 ]
do
for item in `ls *.bin`
    do
        rm -f $copy.out
        echo -n "$item ... "
        mycp $item $copy.out
        echo "done"
    done
done


test1
Code: Select all
S1: PCI-64/33, pci0, 5 V
S2: PCI-64/66, pci1, 3.3 V
S3: PCI-64/33, pci0, 5 V
S4: PCI-64/33, pci2, 5 V
S5: PCI-32/33, pci3, 5 V
S6: PCI-32/33, pci3, 5 V <--------- tested here

kernel 4.16-softfail: working

test2
Code: Select all
S1: PCI-64/33, pci0, 5 V
S2: PCI-64/66, pci1, 3.3 V
S3: PCI-64/33, pci0, 5 V
S4: PCI-64/33, pci2, 5 V <--------- tested here
S5: PCI-32/33, pci3, 5 V
S6: PCI-32/33, pci3, 5 V

kernel 4.16-softfail: working
like being a monkey coloring within the lines

show post TAG


C3600, SYBA-SY-PCX40009

Postby ivelegacy on Wed May 15, 2019 11:26 am

burnin test, 10 hours

test1

script for the test
Code: Select all
while [ 1 ]
do
for item in `ls *.bin`
    do
        rm -f $copy.out
        echo -n "$item ... "
        mycp $item $copy.out
        echo "done"
    done
done


Code: Select all
S1: PCI-64/33, pci0, 5 V
S2: PCI-64/66, pci1, 3.3 V
S3: PCI-64/33, pci0, 5 V
S4: PCI-64/33, pci2, 5 V
S5: PCI-32/33, pci3, 5 V
S6: PCI-32/33, pci3, 5 V <--------- tested here PCI-X card forced to 32bit

kernel 4.16-softfail: panics, HPMC PCI timeout when the controller is stressed out
The PCI-X variant (3124) on 32 bit PCI @ 33MHz can reach the PCI bus limit.

log
Code: Select all
./do_test_adv
data_01GB.bin ...
real    0m28.304s
user    0m0.055s
sys     0m20.541s
done
data_02GB.bin ...
real    0m54.184s
user    0m0.048s
sys     0m40.351s
done
data_04GB.bin ...
real    1m34.203s
user    0m0.100s
sys     1m16.753s
done
data_08GB.bin ...
real    4m19.789s
user    0m0.368s
sys     2m43.647s
done
data_16GB.bin ...
real    8m43.602s
user    0m0.666s
sys     5m28.698s
done
data_32GB.bin ...
real    17m47.451s
user    0m1.027s
sys     11m3.431s
done
data_01GB.bin ...
real    0m26.390s
user    0m0.064s
sys     0m20.635s
done
data_02GB.bin ...
real    0m54.257s
user    0m0.063s
sys     0m40.589s
done
data_04GB.bin ...
real    1m33.727s
user    0m0.110s
sys     1m17.770s
done
data_08GB.bin ...
real    3m57.131s
user    0m0.266s
sys     2m29.096s

data_16GB.bin ...
panics, HPMC PCI timeout



test2

script for the test
Code: Select all
while [ 1 ]
do
for item in `ls *.bin`
    do
        rm -f $copy.out
        echo -n "$item ... "
        mycp $item $copy.out
        echo "done"
    done
done


Code: Select all
S1: PCI-64/33, pci0, 5 V
S2: PCI-64/66, pci1, 3.3 V  <--------- tested here
S3: PCI-64/33, pci0, 5 V
S4: PCI-64/33, pci2, 5 V
S5: PCI-32/33, pci3, 5 V
S6: PCI-32/33, pci3, 5 V

kernel 4.16-softfail: panics, HPMC PCI timeout when the controller is stressed out

test3

script for the test (without md5sucm, and with a delay)
Code: Select all
while [ 1 ]
do
for item in `ls *.bin`
    do
        rm -f $copy.out
        echo -n "$item ... "
        cp $item $copy.out
        echo "done"
        sync
        sleep 10
    done
done



Code: Select all
S1: PCI-64/33, pci0, 5 V
S2: PCI-64/66, pci1, 3.3 V  <--------- tested here
S3: PCI-64/33, pci0, 5 V
S4: PCI-64/33, pci2, 5 V
S5: PCI-32/33, pci3, 5 V
S6: PCI-32/33, pci3, 5 V

kernel 4.16-softfail: panics, HPMC PCI timeout when the controller is stressed out

log
data_01GB.bin ...
real 0m28.224s
user 0m0.048s
sys 0m19.895s
done
data_02GB.bin ...
real 0m58.382s
user 0m0.064s
sys 0m40.622s
done
data_04GB.bin ...
real 2m2.109s
user 0m0.167s
sys 1m20.512s
done
data_08GB.bin ...
real 4m13.158s
user 0m0.227s
sys 2m41.224s
done
data_16GB.bin ...
real 8m43.251s
user 0m0.494s
sys 5m22.696s
done
data_32GB.bin ...
real 17m28.304s
user 0m1.031s
sys 10m50.234s
done
data_01GB.bin ...
real 0m29.201s
user 0m0.032s
sys 0m20.273s
done
data_02GB.bin ...
real 0m55.108s
user 0m0.032s
sys 0m39.741s
done
data_04GB.bin ...
real 1m46.139s
user 0m0.115s
sys 1m18.823s
done
data_08GB.bin ...
real 4m9.451s
user 0m0.270s
sys 2m41.259s
done
data_16GB.bin ...
real 8m36.120s
user 0m0.549s
sys 5m24.918s
done
data_32GB.bin ...
real 17m25.319s
user 0m1.146s
sys 10m54.813s
done
data_01GB.bin ...
real 0m25.629s
user 0m0.016s
sys 0m20.244s
done
data_02GB.bin ...
real 0m55.077s
user 0m0.059s
sys 0m39.600s
done
data_04GB.bin ...
panics, HPMC PCI timeout

like being a monkey coloring within the lines

show post TAG


PowerMacG4, SYBA-SY-PCX40009

Postby madame on Fri May 17, 2019 8:38 pm

burnin test, 10 hours

test1

script for the test
Code: Select all
while [ 1 ]
do
for item in `ls *.bin`
    do
        rm -f $copy.out
        echo -n "$item ... "
        mycp $item $copy.out
        echo "done"
    done
done


Code: Select all
S4: PCI-64/33, pci3, 5 V <--------- tested here
S3: PCI-64/33, pci2, 5 V
S2: PCI-64/33, pci1, 5 V
S1: PCI-64/33, pci0, 5 V
S0: AGP 4x

kernel 4.16: no issue, no problem

log, 1 cycle
...
data_01GB.bin ...
real 0m34.809s
user 0m9.983s
sys 0m14.099s
done
data_02GB.bin ...
real 1m13.730s
user 0m20.158s
sys 0m29.842s
done
data_04GB.bin ...
real 2m31.833s
user 0m39.816s
sys 1m0.220s
done
data_08GB.bin ...
real 4m57.093s
user 1m20.328s
sys 1m59.137s
done
data_16GB.bin ...
real 9m57.283s
user 2m41.205s
sys 3m59.818s
done
data_32GB.bin ...
real 20m2.895s
user 5m21.326s
sys 8m4.992s
done
...

youse guys have got to turn your world around. chinese stuff is deadly and crap.

show post TAG


SYBA-SY-PCX40009

Postby LordCrimson on Sat May 18, 2019 10:53 am

PowerMacG4 MDD PCI64@33Mhz 5V burn-in test success
HPPA C3600 PCI64@33Mhz 5V burn-in test failed!!, HPMC PCI timeout
HPPA C3600 PCI64@66Mhz 3.3V burn-in test failed!!, HPMC PCI timeout
HPPA C3600 PCI32@33Mhz 5V burn-in test failed!!, HPMC PCI timeout


According to Apple's specs, in the default configuration, the PowerMacG4 MDD has four open 33 MHz 64-bit PCI slots, and a 4X AGP slot occupied by the graphics card.
Mystères of the invisible, Loa are intermediaries between Bondye the Supreme Creator, who is distant from the world—and humanity.

show post TAG


HP RocketRAID 2224, PowerMacG4

Postby madame on Sun May 19, 2019 9:14 am

burnin test, 10 hours

test1

script for the test
Code: Select all
while [ 1 ]
do
for item in `ls *.bin`
    do
        rm -f $copy.out
        echo -n "$item ... "
        mycp $item $copy.out
        echo "done"
    done
done


Code: Select all
S4: PCI-64/33, pci3, 5 V <--------- tested here
S3: PCI-64/33, pci2, 5 V
S2: PCI-64/33, pci1, 5 V
S1: PCI-64/33, pci0, 5 V
S0: AGP 4x

kernel 4.16: no issue, no problem

log, 1 cycle
...
data_01GB.bin ...
real 0m32.805s
user 0m9.627s
sys 0m14.282s
done
data_02GB.bin ...
real 1m9.836s
user 0m19.408s
sys 0m30.725s
done
data_04GB.bin ...
real 2m21.577s
user 0m38.571s
sys 1m0.949s
done
data_08GB.bin ...
real 4m45.360s
user 1m18.418s
sys 2m1.852s
done
data_16GB.bin ...
real 9m26.800s
user 2m34.980s
sys 4m7.087s
done
data_32GB.bin ...

...



Note:
It's not clear if the card required to be forced into canonical PCI mode rather than PCI-X mode.
youse guys have got to turn your world around. chinese stuff is deadly and crap.

show post TAG


sata_sil24, sata_mv, a few notes

Postby madame on Sun May 19, 2019 9:24 am

sata_sil24
It seems likely to me that the sil24 driver depends on the card operating in PCI-X mode. PCI-X differs from PCI in interrupt handling and the driver has this flag SIL24_FLAG_PCIX_IRQ_WOC.

a comment in sata_mv.c
if PCIX_IRQ_WOC, there's an inherent race window between clearing IRQ pending status and reading PORT_SLOT_STAT which may cause spurious interrupts afterwards. This is unavoidable and much better than losing interrupts which happens if IRQ pending is cleared after reading PORT_SLOT_STAT


sata_mv
There is a register on the chip, which software could use to override the normal auto-detected PCI mode and bus speed for the chip. This could be used to, say, select 100Mhz or 66Mhz, or even 33Mhz operation, but the register is autodetected from the bus at power-on, and so if software wants to override that by rewriting the reg, it will also need to reset the PCI bus afterward. Which requires knowing how to reset a PCI bridge.
Code: Select all
sata_mv 0000:03:06.0: PCI ERROR; PCI IRQ cause=0x30000040

This error message might be reported for a similar reason, but it's not clear to me.
youse guys have got to turn your world around. chinese stuff is deadly and crap.

show post TAG


Previous

Return to Computers with a UNIX-like OS

cron