2007-07-26 17:55:47

by Michael Wu

Subject: [PATCH] mac80211: revert some GFP_ATOMIC -> GFP_KERNEL changes

From: Michael Wu <[email protected]>

Some of the allocations made with GFP_ATOMIC really were necessary.

Signed-off-by: Michael Wu <[email protected]>
---

net/mac80211/ieee80211_sta.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/mac80211/ieee80211_sta.c b/net/mac80211/ieee80211_sta.c
index 8296c0c..c9fc9a3 100644
--- a/net/mac80211/ieee80211_sta.c
+++ b/net/mac80211/ieee80211_sta.c
@@ -3911,7 +3911,7 @@ ieee80211_sta_scan_result(struct net_device *dev,

if (bss) {
char *buf;
- buf = kmalloc(30, GFP_KERNEL);
+ buf = kmalloc(30, GFP_ATOMIC);
if (buf) {
memset(&iwe, 0, sizeof(iwe));
iwe.cmd = IWEVCUSTOM;
@@ -3929,7 +3929,7 @@ ieee80211_sta_scan_result(struct net_device *dev,
if (!(local->scan_flags & IEEE80211_SCAN_EXTRA_INFO))
break;

- buf = kmalloc(100, GFP_KERNEL);
+ buf = kmalloc(100, GFP_ATOMIC);
if (!buf)
break;

2007-07-26 20:58:23

by Ulrich Kunitz

Subject: Re: [PATCH] mac80211: revert some GFP_ATOMIC -> GFP_KERNEL changes

Andy Green wrote:

> Somebody in the thread at some point said:
>
> > Some of the allocations made with GFP_ATOMIC really were necessary.
>
> I tested it: this makes the BUG stuff in dmesg go away, thanks :-)

Confirmed, patch should definitely go into wireless-dev.

--
Uli Kunitz

2007-07-26 18:20:54

by Andy Green

Subject: Re: [PATCH] mac80211: revert some GFP_ATOMIC -> GFP_KERNEL changes

Somebody in the thread at some point said:

> Some of the allocations made with GFP_ATOMIC really were necessary.

I tested it: this makes the BUG stuff in dmesg go away, thanks :-)

-Andy