Evgeniy Polyakov wrote:
Does this patch (on top of unpatched tree) helps?
I can not test it with real hardware, since I'm sicking at home with
laptop only, but I already know how to make a test case for this problem
without complex setup, so if it does not work, I will investigate it
further as
Hi Patrick.
On Thu, Feb 21, 2008 at 04:29:54PM +0100, Patrick McHardy ([EMAIL PROTECTED])
wrote:
> Unfortunately still no luck. I got an error from ablkcipher_add()
> because of this condition:
...
> - further down it uses dst->length in the last while-loop,
> which seems to need a similar ch
Patrick McHardy wrote:
It also looks like at least two more changes are needed:
> ...
One more thing: I got lots of "printk: 4 messages suppressed"
without any real messages. The reason is a printk_ratelimit()
before a dprintk(). This probably wants either an #ifdef or
a real printk.
-
To unsub
Patrick McHardy wrote:
Evgeniy Polyakov wrote:
On Thu, Feb 21, 2008 at 03:20:45PM +0100, Patrick McHardy
([EMAIL PROTECTED]) wrote:
Almost I guess :) There are similar loops in hifn_setup_session().
Additionally we need to check that the return value of ablkcipher_walk()
is not a negative errno
Evgeniy Polyakov wrote:
On Thu, Feb 21, 2008 at 03:20:45PM +0100, Patrick McHardy ([EMAIL PROTECTED])
wrote:
Almost I guess :) There are similar loops in hifn_setup_session().
Additionally we need to check that the return value of ablkcipher_walk()
is not a negative errno code.
Yep. Kind of t
On Thu, Feb 21, 2008 at 03:20:45PM +0100, Patrick McHardy ([EMAIL PROTECTED])
wrote:
> Almost I guess :) There are similar loops in hifn_setup_session().
> Additionally we need to check that the return value of ablkcipher_walk()
> is not a negative errno code.
Yep. Kind of this one:
diff --git a
Evgeniy Polyakov wrote:
Hi Herbert.
On Thu, Feb 21, 2008 at 10:10:13PM +0800, Herbert Xu ([EMAIL PROTECTED]) wrote:
Yes by the time the request gets to hifn req->nbytes would have
been decremented by 12 bytes which means that it is now the exact
amount of data that needs to be encrypted.
Howev
Hi Herbert.
On Thu, Feb 21, 2008 at 10:10:13PM +0800, Herbert Xu ([EMAIL PROTECTED]) wrote:
> Yes by the time the request gets to hifn req->nbytes would have
> been decremented by 12 bytes which means that it is now the exact
> amount of data that needs to be encrypted.
>
> However, if you follow
On Thu, Feb 21, 2008 at 12:10:12PM +0300, Evgeniy Polyakov wrote:
>
> But req->nbytes should or should not correspond to number of bytes to be
> encrypted? HIFN driver assumes so, it can be changed to run through
> scatterlists (btw, how will it determine how many of them were
> provided?) and get
On Thu, Feb 21, 2008 at 01:26:20AM +0800, Herbert Xu ([EMAIL PROTECTED]) wrote:
> On Wed, Feb 20, 2008 at 01:33:22PM +0100, Patrick McHardy wrote:
> >
> > I saw the discrepancy between the elen value used
> > for aead_request_set_crypt() in esp_input() and the
> > req->nbytes value seen in hifn_set
On Wed, Feb 20, 2008 at 04:19:03PM +0300, Evgeniy Polyakov wrote:
>
> What iv generation scheme do you use? It looks like only gcm and ccm add
> 16 bytes to cryptlen and set nbytes to them. Although they both install
> two scatterlists for crypto operation: one for data and one for above
> addition
On Wed, Feb 20, 2008 at 01:33:22PM +0100, Patrick McHardy wrote:
>
> I saw the discrepancy between the elen value used
> for aead_request_set_crypt() in esp_input() and the
> req->nbytes value seen in hifn_setup_session().
OK that is normal. esp_input is now using the AEAD interface
so the size f
Hi Patrick.
On Wed, Feb 20, 2008 at 01:33:22PM +0100, Patrick McHardy ([EMAIL PROTECTED])
wrote:
> I saw the discrepancy between the elen value used
> for aead_request_set_crypt() in esp_input() and the
> req->nbytes value seen in hifn_setup_session().
What iv generation scheme do you use? It lo
Herbert Xu wrote:
On Tue, Feb 19, 2008 at 05:27:25PM +0100, Patrick McHardy wrote:
Yes, probably. So I take it the assumption in HIFN that the
sg-list length matches req->nbytes is incorrect?
Where were you seeing the discrepancy? If it's at the point of
entry into the HIFN code then the bug i
On Tue, Feb 19, 2008 at 05:27:25PM +0100, Patrick McHardy wrote:
>
> Yes, probably. So I take it the assumption in HIFN that the
> sg-list length matches req->nbytes is incorrect?
Where were you seeing the discrepancy? If it's at the point of
entry into the HIFN code then the bug is further up. I
Herbert Xu wrote:
Patrick McHardy <[EMAIL PROTECTED]> wrote:
I couldn't figure out where in the crypto code the
nbytes decrement by 12 bytes compared to the length
seen when setting up the crypto operation happens
or I might have tried to properly fix it myself.
I'll happily test patches in case
Patrick McHardy wrote:
Evgeniy Polyakov wrote:
Hi Patrick.
On Wed, Feb 13, 2008 at 05:44:42PM +0300, Evgeniy Polyakov
([EMAIL PROTECTED]) wrote:
Any chance you can apply following patch and check output for correct
and broken cases (it will produce 2 or 3 debug strings for each crypto
operati
Evgeniy Polyakov wrote:
Hi Patrick.
On Wed, Feb 13, 2008 at 05:44:42PM +0300, Evgeniy Polyakov ([EMAIL PROTECTED])
wrote:
Any chance you can apply following patch and check output for correct
and broken cases (it will produce 2 or 3 debug strings for each crypto
operation)?
diff --git a/drive
Patrick McHardy <[EMAIL PROTECTED]> wrote:
>
> I couldn't figure out where in the crypto code the
> nbytes decrement by 12 bytes compared to the length
> seen when setting up the crypto operation happens
> or I might have tried to properly fix it myself.
> I'll happily test patches in case someone
Hi Patrick.
On Wed, Feb 13, 2008 at 05:44:42PM +0300, Evgeniy Polyakov ([EMAIL PROTECTED])
wrote:
> Any chance you can apply following patch and check output for correct
> and broken cases (it will produce 2 or 3 debug strings for each crypto
> operation)?
>
> diff --git a/drivers/crypto/hifn_79
Hi Patric.
On Wed, Feb 13, 2008 at 02:17:45PM +0100, Patrick McHardy ([EMAIL PROTECTED])
wrote:
> I'm getting crashes when using HIFN and IPsec (ESP with
> AES + MD5) in the current -git tree. I didn't capture the
> Oops, but there seem to be a number of problems:
>
> - hifn_setup_session walks
I'm getting crashes when using HIFN and IPsec (ESP with
AES + MD5) in the current -git tree. I didn't capture the
Oops, but there seem to be a number of problems:
- hifn_setup_session walks over the scatterlist, subtracting
the scatterlist element size from nbytes until nbytes
reaches zero. I
22 matches
Mail list logo
