| To: | Jacob Avraham <jacoba@xxxxxxxxx> |
|---|---|
| Subject: | Re: conflicting alignment requirements |
| From: | Andi Kleen <ak@xxxxxx> |
| Date: | Wed, 25 Jul 2001 21:15:52 +0200 |
| Cc: | Andi Kleen <ak@xxxxxx>, Network Development List <netdev@xxxxxxxxxxx> |
| In-reply-to: | <EJEHILNJPONOHGEOJKICAECOCAAA.jacoba@cisco.com>; from Jacob Avraham on Wed, Jul 25, 2001 at 06:40:27PM +0200 |
| References: | <20010725151707.44709@colin.muc.de> <EJEHILNJPONOHGEOJKICAECOCAAA.jacoba@cisco.com> |
| Sender: | owner-netdev@xxxxxxxxxxx |
On Wed, Jul 25, 2001 at 06:40:27PM +0200, Jacob Avraham wrote: > #if !defined(__i386__) && !defined(__mc68000__) > if ((unsigned long)ptr & 3) > return -1; > #endif As far as I can see this code is wrong: all architectures should be able to handle unaligned accesses in kernel, otherwise they're remotely exploitable anyways. I guess you can just drop the ifdef, and if it should break anything complain to the maintainer of the architecture. -Andi > |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: static routes and dead gateway detection, Julian Anastasov |
|---|---|
| Next by Date: | Re: InterScan NT Alert, Rik van Riel |
| Previous by Thread: | RE: conflicting alignment requirements, Jacob Avraham |
| Next by Thread: | Re: conflicting alignment requirements, Alexey Kuznetsov |
| Indexes: | [Date] [Thread] [Top] [All Lists] |