2003-06-13 10:45:04

by Frank Cusack

[permalink] [raw]
Subject: cryptoapi 2.5->2.4 backport

Any problem with just changing crypto/cipher.c:

enum km_type crypto_km_types[] = {
KM_USER0,
KM_USER1,
KM_SOFTIRQ0,
KM_SOFTIRQ1,
};

to

enum km_type crypto_km_types[] = {
KM_USER0,
KM_USER1,
KM_USER0,
KM_USER1,
};

?

(or the equivalent change in crypto/internal.h)

static inline enum km_type crypto_kmap_type(int out)
{
return crypto_km_types[(in_softirq() ? 2 : 0) + out];
}

static inline void *crypto_kmap(struct page *page, int out)
{
return kmap_atomic(page, crypto_kmap_type(out));
}

thanks
/fc


2003-06-13 15:29:19

by James Morris

[permalink] [raw]
Subject: Re: cryptoapi 2.5->2.4 backport

On Fri, 13 Jun 2003, Frank Cusack wrote:

> Any problem with just changing crypto/cipher.c:
>
> enum km_type crypto_km_types[] = {
> KM_USER0,
> KM_USER1,
> KM_SOFTIRQ0,
> KM_SOFTIRQ1,
> };
>
> to
>
> enum km_type crypto_km_types[] = {
> KM_USER0,
> KM_USER1,
> KM_USER0,
> KM_USER1,
> };
>
> ?

Yes, the crypto calls can run in user and softirq context.

Which backport is this?


- James
--
James Morris
<[email protected]>