Re: [PATCH AUTOSEL 5.3 169/203] x86/platform/uv: Fix kmalloc() NULL check routine

From: Mike Travis
Date: Wed Oct 02 2019 - 10:37:12 EST




On 10/2/2019 1:34 AM, David Laight wrote:
From: Sasha Levin
Sent: 01 October 2019 17:06
Subject: Re: [PATCH AUTOSEL 5.3 169/203] x86/platform/uv: Fix kmalloc() NULL check routine

On Sun, Sep 22, 2019 at 10:25:44PM +0200, Greg KH wrote:
On Sun, Sep 22, 2019 at 02:43:15PM -0400, Sasha Levin wrote:
From: Austin Kim <austindh.kim@xxxxxxxxx>

[ Upstream commit 864b23f0169d5bff677e8443a7a90dfd6b090afc ]

The result of kmalloc() should have been checked ahead of below statement:

pqp = (struct bau_pq_entry *)vp;

Move BUG_ON(!vp) before above statement.

Signed-off-by: Austin Kim <austindh.kim@xxxxxxxxx>
...
---
arch/x86/platform/uv/tlb_uv.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/platform/uv/tlb_uv.c b/arch/x86/platform/uv/tlb_uv.c
index 20c389a91b803..5f0a96bf27a1f 100644
--- a/arch/x86/platform/uv/tlb_uv.c
+++ b/arch/x86/platform/uv/tlb_uv.c
@@ -1804,9 +1804,9 @@ static void pq_init(int node, int pnode)

plsize = (DEST_Q_SIZE + 1) * sizeof(struct bau_pq_entry);
vp = kmalloc_node(plsize, GFP_KERNEL, node);
- pqp = (struct bau_pq_entry *)vp;
- BUG_ON(!pqp);
+ BUG_ON(!vp);

+ pqp = (struct bau_pq_entry *)vp;
cp = (char *)pqp + 31;
pqp = (struct bau_pq_entry *)(((unsigned long)cp >> 5) << 5);


How did this even get merged in the first place? I thought a number of
us complained about it.

This isn't any change in code, and the original is just fine, the author
didn't realize how C works :(

Mind you, the code itself if pretty horrid.
Looks like it is aligning to 32 bytes, easier done by:
pqp = (void *)((unsigned long)vp + 31 & ~31);
(and there's a roundup macro to obfuscate it somewhere.)
But I'd also expect to see a matching '+ 31' in the size passed to kmalloc().
Not to mention a comment!

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)


Thanks, I will put all of these comments in my notes. This whole function is slated to move to a specialized UV APIC driver since it uses a unique scaling feature available in the UV hardware. (The original author has long retired.)

-Mike