Skip to contentSkip to conversation

BUG at mm/memory.c:1501 in 2.6.0-test5

DISCUSSION Petr Vandrovec 4 participants 4 messages Sep 18, 2003 – Sep 20, 2003

lore
Share & tools
lore
#0 lore raw
Quoted 10 lines · William Lee Irwin III
Quoted 1 line · Petr Vandrovec

EIP: 0060:[<c015be10>] Tainted: PF

snprintf(buf, sizeof(buf), "Tainted: %c%c%c",
tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G', tainted & TAINT_FORCED_MODULE ? 'F' : ' ', tainted & TAINT_UNSAFE_SMP ? 'S' : ' ');

This is probably the reason you're not getting much in the way of a response.

I explicitly stated that it happened shortly after I shut down VMware UI, and that I spent whole day trying to find what's going on, finally politely asking for help, hoping that someone could have a clue what went wrong.
Petr Vandrovec

#1 lore raw
Quoted 9 lines · Petr Vandrovec
Quoted 8 lines · William Lee Irwin III
Quoted 1 line · Petr Vandrovec

EIP: 0060:[<c015be10>] Tainted: PF

snprintf(buf, sizeof(buf), "Tainted: %c%c%c",
tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G', tainted & TAINT_FORCED_MODULE ? 'F' : ' ', tainted & TAINT_UNSAFE_SMP ? 'S' : ' ');

Quoted 9 lines
Quoted 2 lines

This is probably the reason you're not getting much in the way of a response.

I explicitly stated that it happened shortly after I shut down VMware UI, and that I spent whole day trying to find what's going on, finally politely asking for help, hoping that someone could have a clue what went wrong.
Petr Vandrovec

Okay. I'll be more specific. The "Tainted PF" shown above is because you have installed a module that is [P]roprietary and it was [F]orced to load.

Any module running inside the kernel can destroy anything. There is no protection inside the kernel. A simple bug in any module can not only cause your machine to die, but it can, in principle, destroy everything on your hard disk as well as shutting down your LAN, causing millions of dollars of damages (seriously). It is possible.

Therefore, If you report a bug, and your system is tainted with a proprietary module, nobody can help you because the source-code of the module isn't available for review. Therefore nobody will waste time reviewing kernel code when a single out-of-bounds pointer, or a single failure to free a resource or a lock on some seldom-used path in that proprietary module could kill the kernel.

That said, the proprietary module may be faultless. However, nobody is going to be able to find the problem without being able to review the source-code of that module. So, don't install that module, wait for the crash again, then file another report. If the machine doesn't crash again, then you've probably found the problem, that "secret" module code that the vendor doesn't want you to see because you might vomit or might reverse- engineer, resulting in a better product.

Cheers,
Dick Johnson
Penguin : Linux version 2.4.22 on an i686 machine (794.73 BogoMips).
Note 96.31% of all statistics are fiction.

#2 lore raw
Quoted 36 lines · Richard B. Johnson
Quoted 11 lines · Petr Vandrovec
Quoted 9 lines · William Lee Irwin III
Quoted 2 lines · Petr Vandrovec

EIP: 0060:[<c015be10>] Tainted: PF

snprintf(buf, sizeof(buf), "Tainted: %c%c%c",
tainted & TAINT_PROPRIETARY_MODULE ? 'P' : 'G', tainted & TAINT_FORCED_MODULE ? 'F' : ' ', tainted & TAINT_UNSAFE_SMP ? 'S' : ' ');

Quoted 10 lines
Quoted 3 lines

This is probably the reason you're not getting much in the way of a response.

I explicitly stated that it happened shortly after I shut down VMware UI, and that I spent whole day trying to find what's going on, finally politely asking for help, hoping that someone could have a clue what went wrong.
Petr Vandrovec

Okay. I'll be more specific. The "Tainted PF" shown above is because you have installed a module that is [P]roprietary and it was [F]orced to load.

Any module running inside the kernel can destroy anything. There is no protection inside the kernel. A simple bug in any module can not only cause your machine to die, but it can, in principle, destroy everything on your hard disk as well as shutting down your LAN, causing millions of dollars of damages (seriously). It is possible.

Quoted 3 lines

Therefore, If you report a bug, and your system is tainted with a proprietary module, nobody can help you because the

You should send your report to the vendor. They might not be supporting 2.6 yet though. Alternatively, try to reproduce the problem having never loaded closed source modules since booting: it is now much less frustrating for developers to track down and fix.

#3 lore raw
In reply to Petr Vandrovec
Quoted 2 lines · William Lee Irwin III
Quoted 2 lines

This is probably the reason you're not getting much in the way of a response.

Quoted 4 lines · Petr Vandrovec

I explicitly stated that it happened shortly after I shut down VMware UI, and that I spent whole day trying to find what's going on, finally politely asking for help, hoping that someone could have a clue what went wrong.

Much better. I presumed something like nvidia. Now that that's cleared up, I'll have to find time to take a breather and peek at the vmware bits to see if they're doing anything that might get misunderstood by the VM.

One thing to look for in vmware if you have the time/motivation yourself is to check for dirty bits getting left in (or put into) non- present ptes; this clashes with _PTE_FILE on i386.

-- wli