[ros-bugs] [Bug 703] ReactOS is a fat pig. 0.2.7-RC2 mem reqs exceeded.

[ReactOS.Bugzilla at reactos.org]

http://www.reactos.org/bugzilla/show_bug.cgi?id=703





------- Additional Comments From waxdragon at gmail.com  2005-10-06 05:01 CET -------
Doing some more testing, it seems with 32MB that I get a NO_PAGES_AVAILABLE
crash on the first boot under qemu:

(drivers/net/afd/afd/select.c:256)(AfdSelect) Returning 103
(ntoskrnl/mm/freelist.c:927) MmAllocPage(): Out of memory
KeBugCheck at ntoskrnl/mm/balance.c:209
A problem has been detected and ReactOS has been shut down to prevent damage to
your computer.

Technical information:

*** STOP: 0x0000004D (0x00000000,0x00000000,0x00000000,0x00000000)

Frames:
<ntoskrnl.exe:2615 (ntoskrnl/ke/bug.c:479 (KeBugCheckEx))>
<ntoskrnl.exe:264d (ntoskrnl/ke/bug.c:500 (KeBugCheck))>
<ntoskrnl.exe:5a7f4 (ntoskrnl/mm/balance.c:209 (MmRequestPageMemoryConsumer))>
<ntoskrnl.exe:648f5 (ntoskrnl/mm/npool.c:1224 (grow_block))>
<ntoskrnl.exe:64d86 (ntoskrnl/mm/npool.c:1376 (get_block))>
<ntoskrnl.exe:651b7 (ntoskrnl/mm/npool.c:1559 (ExAllocateNonPagedPoolWithTag))>
<ntoskrnl.exe:69641 (ntoskrnl/mm/pool.c:40 (EiAllocatePool))>
<ntoskrnl.exe:696e5 (ntoskrnl/mm/pool.c:109 (ExAllocatePoolWithTag))>
<ntoskrnl.exe:c4921 (lib/rossym/initkm.c:24 (RosSymAllocMemKM))>
<ntoskrnl.exe:c51e3 (lib/rossym/fromfile.c:120 (RosSymCreateFromFile))>
<ntoskrnl.exe:a97cd (ntoskrnl/kdbg/kdb_symbols.c:452 (KdbpSymLoadModuleSymbols))>
<ntoskrnl.exe:a98cb (ntoskrnl/kdbg/kdb_symbols.c:512 (KdbSymLoadUserModuleSymbols))>
<ntoskrnl.exe:1cebb (ntoskrnl/ex/dbgctrl.c:45 (NtSystemDebugControl))>
<ntoskrnl.exe:99afc ({standard input}:373 (KiSystemService))>
<ntdll.dll:35b9 (lib/ntdll/ldr/utils.c:76 (LdrpLoadUserModuleSymbols))>

Entered debugger on embedded INT3 at 0x0008:0x800af766.
kdb:>

This is unrelated and probably due to the fact I'm running a debug build.


I checked vmware again, and I think the reason the first stage installer works
under qemu, but not under VMWare is "wall time".  QEmu putters along copying
files taking nearly 40 seconds to copy all the files, and I assume that gives
the memory/cache manager time to free pages. OTOH, VMWare takes less than a
second to to hit 35-40% and I don't think enough pages get freed in that time.

That means this bug's behaviour has changed. I think the process that frees
memory actually works now, but there is a timing bug exposed now.

-- 
Configure bugmail: http://www.reactos.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
You are the QA contact for the bug, or are watching the QA contact.