Opened 15 years ago
Closed 10 years ago
#5773 closed bug (duplicate)
vm_page_fault after adding graph
| Reported by: | andreasf | Owned by: | axeld |
|---|---|---|---|
| Priority: | normal | Milestone: | R1 |
| Component: | Applications/ActivityMonitor | Version: | R1/Development |
| Keywords: | Cc: | ||
| Blocked By: | #6481 | Blocking: | |
| Platform: | x86 |
Description
Occasionally, File > Add graph leads to the app not properly redrawing and unable to close, followed by a vm_page_fault.
Last seen at hrev36382.
Change History (9)
follow-up: 2 comment:1 by , 15 years ago
comment:2 by , 15 years ago
It's a non-recoverable KDL. Could reproduce the hang now but not the KDL, here's the serial log, still at hrev36382:
virtual void AddOnManager::MessageReceived(BMessage*) what: vdwI virtual void AddOnManager::MessageReceived(BMessage*) what: vdwI Last message repeated 5 times. Last message repeated 10 times. Last message repeated 11 times. PageWriteWrapper: Failed to write page 0x83a0dfb0: General system error
The window fails to redraw beneath the menu, unless where I move the mouse. The graphs are not updated and the system seems sluggish. The window can be moved but not resized or closed, while other windows are okay. The CPU goes crazy, and after killing ActivityMonitor everything seemed normal again this time.
The PageWriteWrapper error was not reproducible next time.
follow-up: 4 comment:3 by , 15 years ago
Forgot to mention that this is on real SMP hardware. But the hang happens with second CPU disabled, too.
Sometimes the third (network) graph is sufficient to trigger this, sometimes it requires over twenty adds, but the hang is very much reproducible here. Not for you?
More PageWriteWrapper errors on third try:
slab memory manager: deleting area 0xf6000000 (4452) PageWriteWrapper: Failed to write page 0x83d02510: General system error PageWriteWrapper: Failed to write page 0x83d0253c: General system error PageWriteWrapper: Failed to write page 0x83d02568: General system error
comment:4 by , 15 years ago
Replying to andreasf:
Forgot to mention that this is on real SMP hardware. But the hang happens with second CPU disabled, too.
Sometimes the third (network) graph is sufficient to trigger this, sometimes it requires over twenty adds, but the hang is very much reproducible here. Not for you?
At least on real hardware I have never seen it. Though I've never tried to add 20 graphs either. I just tested a bit under qemu with hrev36408 and haven't encountered any problem either.
More PageWriteWrapper errors on third try:
slab memory manager: deleting area 0xf6000000 (4452) PageWriteWrapper: Failed to write page 0x83d02510: General system error PageWriteWrapper: Failed to write page 0x83d0253c: General system error PageWriteWrapper: Failed to write page 0x83d02568: General system error
When you encounter it again please try the page command in KDL for one of the pages that could not be written and then cache for its cache (if any). Also interesting would be avail and page_stats (no parameters). Note that you can inspect the syslog from KDL via syslog | tail. Furthermore please check which thread or threads eat the CPU and get a back trace for them (bt <thread ID> in KDL).
comment:5 by , 15 years ago
0x8281fe00 89216 0xccd24200 ActivityMonitor
kdebug> threads 89216
thread id state wait for object cpu pri stack team name
0x83342800 89311 waiting sem 326321 - 20 0xdebdb00089216 source refresh
0x83347000 89280 waiting sem 326321 - 20 0xdebc300089216 source refresh
0x83305800 89406 waiting sem 326321 - 20 0xdebd300089216 source refresh
0x83302000 89251 zzz - 20 0xdeb5b00089216 source refresh
0x83341800 89220 waiting cvar 0x81163fdc - 20 0xdeb2b00089216 _BMediaRoster_
0xce12b000 89376 waiting sem 326321 - 20 0xdebc700089216 source refresh
0x83343000 89221 waiting sem 326321 - 20 0xdeb2f00089216 source refresh
0x832cc800 89222 waiting sem 326321 - 20 0xdeb3300089216 source refresh
0xce19e000 89223 running - 1 15 0xdeb5300089216 w>ActivityMonitor
0x83321000 89349 waiting sem 326321 - 20 0xdeb6300089216 source refresh
0xce1a0800 89289 waiting sem 326321 - 20 0xdebcb00089216 source refresh
0x83345000 89258 waiting sem 326321 - 20 0xdebab00089216 source refresh
0xce19a000 89322 waiting sem 326321 - 20 0xdebe300089216 source refresh
0xce2d3000 89229 waiting sem 326321 - 20 0xdeb5f00089216 source refresh
0x83303800 89417 waiting sem 329283 - 20 0x8133000089216 source refresh
0x83346000 89264 waiting sem 326321 - 20 0xdebb300089216 source refresh
0x83345800 89390 waiting sem 326321 - 20 0xdebaf00089216 source refresh
0xce2d6800 89235 waiting sem 326321 - 20 0xdeb9300089216 source refresh
0x83346800 89363 zzz - 20 0xdebe700089216 source refresh
0xce19c800 89301 waiting sem 326321 - 20 0xdeb9700089216 source refresh
0xce19b800 89270 waiting sem 326321 - 20 0xdebbb00089216 source refresh
0xce19b000 89243 zzz - 20 0xdeb9b00089216 source refresh
0x83327800 89337 waiting sem 326321 - 20 0xdebeb00089216 source refresh
0x83344000 89216 waiting cvar 0x81163c58 - 10 0xdeb4b00089216 ActivityMonitor
kdebug> bt 89223
stack trace for thread 89223 "w>ActivityMonitor"
kernel stack: 0xdeb53000 to 0xdeb57000
user stack: 0x700c3000 to 0x70103000
frame caller <image>:function + offset
0 deb56c78 (+ 48) 8005bfa4 <kernel_x86> process_pending_ici(int32: 204) + 0x0094
1 deb56ca8 (+ 16) 8005c217 <kernel_x86>:smp_intercpu_int_handler + 0x0017
2 deb56cb8 (+ 48) 8004db28 <kernel_x86>:int_io_interrupt_handler + 0x0058
3 deb56ce8 (+ 32) 800f4244 <kernel_x86> hardware_interrupt(iframe*: 0xdeb56d14) + 0x0114
4 deb56d08 (+ 12) 800f9f8d <kernel_x86>:int_bottom + 0x003d
kernel iframe at 0xdeb56d14 (end = 0xdeb56d64)
eax 0x0 ebx 0xcd200d70 ecx 0x0 edx 0x80000000
esi 0x3 edi 0x82fe4258 ebp 0xdeb56d64 esp 0xdeb56d48
eip 0x8004c786 eflags 0x10296
vector: 0xfd, error code: 0x0
5 deb56d14 (+ 80) 8004c786 <kernel_x86>:free_etc + 0x0006
6 deb56d64 (+ 32) 800e7b85 <kernel_x86> VMArea::~VMArea() + 0x0015
7 deb56d84 (+ 32) 800ec004 <kernel_x86> VMUserAddressSpace<0x82fe4258>::DeleteArea(VMArea*: 0xcd200d70, uint32: 0x3 (3)) + 0x0014
8 deb56da4 (+ 96) 800d7109 <kernel_x86> map_backing_store(VMAddressSpace*: NULL, VMCache*: NULL, void**: 0x19790000, int64: 2, uint32: 0x33 (51), uint32: 0x0 (09
9 deb56e04 (+ 192) 800d8c43 <kernel_x86>:vm_create_anonymous_area + 0x02b3
10 deb56ec4 (+ 128) 800d92b1 <kernel_x86>:_user_create_area + 0x00c1
11 deb56f44 (+ 100) 800fa1c2 <kernel_x86>:handle_syscall + 0x00af
user iframe at 0xdeb56fa8 (end = 0xdeb57000)
eax 0xb3 ebx 0x817f68 ecx 0x70102630 edx 0xffff0114
esi 0x18000000 edi 0x19786d10 ebp 0x7010265c esp 0xdeb56fdc
eip 0xffff0114 eflags 0x212 user esp 0x70102630
vector: 0x63, error code: 0x0
12 deb56fa8 (+ 0) ffff0114 <commpage>:commpage_syscall + 0x0004
13 7010265c (+ 112) 007e44c8 </boot/system/lib/libroot.so@0x00744000>:unknown + 0xa04c8
14 701026cc (+ 48) 007e935f </boot/system/lib/libroot.so@0x00744000>:unknown + 0xa535f
15 701026fc (+ 80) 007e9da8 </boot/system/lib/libroot.so@0x00744000>:unknown + 0xa5da8
16 7010274c (+ 48) 007eb348 </boot/system/lib/libroot.so@0x00744000>:unknown + 0xa7348
17 7010277c (+ 48) 007eb563 </boot/system/lib/libroot.so@0x00744000>:unknown + 0xa7563
18 701027ac (+ 48) 0035058f </boot/system/lib/libbe.so@0x00227000>:unknown + 0x12958f
19 701027dc (+ 304) 002147f0 </boot/system/apps/ActivityMonitor@0x00200000>:unknown + 0x147f0
20 7010290c (+ 464) 002149ce </boot/system/apps/ActivityMonitor@0x00200000>:unknown + 0x149ce
21 70102adc (+ 304) 0030cbf8 </boot/system/lib/libbe.so@0x00227000>:unknown + 0xe5bf8
22 70102c0c (+ 48) 00385ca8 </boot/system/lib/libbe.so@0x00227000>:unknown + 0x15eca8
23 70102c3c (+ 752) 0038fc63 </boot/system/lib/libbe.so@0x00227000>:unknown + 0x168c63
24 70102f2c (+ 112) 0038c5d7 </boot/system/lib/libbe.so@0x00227000>:unknown + 0x1655d7
25 70102f9c (+ 32) 002cfe4c </boot/system/lib/libbe.so@0x00227000>:unknown + 0xa8e4c
26 70102fbc (+ 32) 0076e64e </boot/system/lib/libroot.so@0x00744000>:unknown + 0x2a64e
27 70102fdc (+ 0) 70102fec 2388290:w>ActivityMonitor_89223_stack@0x700c3000 + 0x3ffec
kdebug> avail
Available memory: 901455872/1072496640 bytes
kdebug> page_stats
page stats:
total: 261840
active: 14244 (busy: 0)
inactive: 201358 (busy: 0)
cached: 781 (busy: 0)
unused: 42 (busy: 0)
wired: 40969 (busy: 0)
modified: 0 (busy: 0)
free: 0
clear: 4446
unreserved free pages: 4446
unsatisfied page reservations: 0
mapped pages: 58183
waiting threads:
free queue: 0x801654f0, count = 0
clear queue: 0x80165504, count = 4446
modified queue: 0x801654c8, count = 0 (0 temporary, 0 swappable, inactive: 0)
active queue: 0x801654a0, count = 14244
inactive queue: 0x801654b4, count = 201358
cached queue: 0x801654dc, count = 781
kdebug>
follow-up: 7 comment:6 by , 15 years ago
Normally with three graphs ActivityMonitor memory usage is around 1.4 MB according to ProcessController. With a dozen graphs and the hang it skyrocketed over 800 MB and kept climbing.
Overall used memory according to AboutSystem went from maybe 120 MB to only around 390 MB and stagnated there. RAM is 1 GB, swap 2 GB.
The page_writer(?) thread in the kernel_team got pretty active at first but calmed down again.
This is at hrev36408.
comment:7 by , 15 years ago
Replying to andreasf:
Normally with three graphs ActivityMonitor memory usage is around 1.4 MB according to ProcessController. With a dozen graphs and the hang it skyrocketed over 800 MB and kept climbing.
Indeed. The stack trace seems to indicate that it just tried to allocate another area and failed (probably due to lack of resources -- though avail still shows almost 900 MB available).
Overall used memory according to AboutSystem went from maybe 120 MB to only around 390 MB and stagnated there. RAM is 1 GB, swap 2 GB.
I'm not quite sure what stats AboutSystem's info is based on, but page_stats and avail (well, I forgot swap) give a detailed picture. There are only 4446 free pages and 781 cached ones. There are 40969 wired ones, which is quite a lot, unless you have kernel tracing enabled with a sizable buffer. And we have 201358 inactive pages (i.e. actually (but not very recently) used ones), which were probably mostly allocated by ActivityMonitor. An areas <team id> for ActivityMonitor would be interesting.
The page_writer(?) thread in the kernel_team got pretty active at first but calmed down again.
This is expected when a team is allocating obscene amounts of memory, as it's the one swapping out pages. The page daemon should probably be even busier in the beginning. At least the low memory situation explains why the whole system slows down. That doesn't explain the crash, but without a stack trace there's little that can be done about it.
The "PageWriteWrapper: Failed to write page" output is probably related to http://dev.haiku-os.org/browser/haiku/trunk/src/system/kernel/port.cpp#L667 and basically harmless.
comment:8 by , 14 years ago
FYI on hrev39551 GCC2Hybrid, if I add graphs to ActivityMonitor it never reaches the debugger, but the program stops responding completely after I add around 8 or so graphs. This is under VirtualBox with 4 CPUs.
comment:9 by , 10 years ago
| Blocked By: | 6481 added |
|---|---|
| Resolution: | → duplicate |
| Status: | new → closed |
Closing this since the KDL is gone (at least here I can't reproduce that part). The freeze is tracked in #6481.
Can you be more specific please. vm_page_fault() is just the general page fault handler function. It is called very often during normal operation. If this is a kernel panic, please add a stack trace. If this is an application crash, please add a gdb stack trace and the relevant part of the syslog (or the complete thing).