Skip to contentSkip to conversation

Option to clear allocated kernel memory before freeing it?

DISCUSSION Michael Opdenacker 3 participants 4 messages Jun 19, 2006

lore
Share & tools
lore
#0 lore raw

Hello,

Would it make sense to implement a kernel option that would clear kernel memory before freeing it (by kfree or free_page(s))?

Unless I'm missing something, uncleared memory previously used for kernel allocations could later be recycled for user allocations, making it possible for a user program to access sensitive driver data if it's lucky.

Tough clearing memory should be efficient (thanks to the use of memset(), optimized for each platform), there would of course be a significant performance hit. However, this could be acceptable for systems with strong security requirements...

What do you think? If this idea makes sense, I'll be glad to help in implementing it.

Thanks in advance,

Cheers,

Michael.

Signature · 4 lines
Michael Opdenacker, Free Electrons
Free Embedded Linux Training Materials
on http://free-electrons.com/training
(More than 1000 pages!)
#1 lore raw
Quoted 5 lines · Michael Opdenacker

Hello,

Would it make sense to implement a kernel option that would clear kernel memory before freeing it (by kfree or free_page(s))?

Unless I'm missing something, uncleared memory previously used for kernel allocations could later be recycled for user allocations, making it possible for a user program to access sensitive driver data if it's lucky.

No sane operating system lets user-space have access to pages that haven't been cleared.

Quoted 1 line

Tough clearing memory should be efficient (thanks to the use of memset(), optimized for each platform), there would of course be a significant performance hit. However, this could be acceptable for systems with strong security requirements...

A better way (and I believe the way Linux does it) is to map new pages to the zero page, read-only. If the process attempts to write to the zero page, we take a page fault and set up a new zeroed page, change the mapping, and return control back to that process.

Quoted 8 lines

What do you think? If this idea makes sense, I'll be glad to help in implementing it.

Thanks in advance,

Cheers,

Michael.

Thanks,
Chase

#2 lore raw
In reply to Michael Opdenacker
Quoted 5 lines · Michael Opdenacker

Hello,

Would it make sense to implement a kernel option that would clear kernel memory before freeing it (by kfree or free_page(s))?

No. Memory is cleared before being mapped to user-space. Memory that is allocated for use by the kernel is never cleared by default. To do so would waste valuable time for nothing gained.

Quoted 4 lines

Unless I'm missing something, uncleared memory previously used for kernel allocations could later be recycled for user allocations, making it possible for a user program to access sensitive driver data if it's lucky.

Wrong. You are missing a lot.

Quoted 6 lines

Tough clearing memory should be efficient (thanks to the use of memset(), optimized for each platform), there would of course be a significant performance hit. However, this could be acceptable for systems with strong security requirements...

Clearing, using the CPU is never efficient. That's why demand-zero paging is used by many operating systems.

Quoted 11 lines

What do you think? If this idea makes sense, I'll be glad to help in implementing it.

Thanks in advance,
Cheers,
Michael.
--
Michael Opdenacker, Free Electrons
Free Embedded Linux Training Materials
on http://free-electrons.com/training
(More than 1000 pages!)

Cheers,
Dick Johnson
Penguin : Linux version 2.6.16.4 on an i686 machine (5592.72 BogoMips). New book: http://www.AbominableFirebug.com/
_


****************************************************************
The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [email protected] - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.

#3 lore raw

Chase, Dick,

Quoted 5 lines

No. Memory is cleared before being mapped to user-space. Memory that is allocated for use by the kernel is never cleared by default. To do so would waste valuable time for nothing gained.

Quoted 8 lines
Quoted 5 lines

Unless I'm missing something, uncleared memory previously used for kernel allocations could later be recycled for user allocations, making it possible for a user program to access sensitive driver data if it's lucky.

Wrong. You are missing a lot.

Oops, I realize I was really missing a lot! Thank you very much for leading me to the right path!

Cheers,

Michael.

Signature · 4 lines
Michael Opdenacker, Free Electrons
Free Embedded Linux Training Materials
on http://free-electrons.com/training
(More than 1000 pages!)..