On 7/16/06, Eric Tanguy <eric.tanguy@xxxxxxxxxxxxxx> wrote:
Le samedi 15 juillet 2006 à 21:35 -0700, Lonni J Friedman a écrit : > On 7/15/06, Ric Moore <wayward4now@xxxxxxxxx> wrote: > > On Sat, 2006-07-15 at 16:27 -0700, Lonni J Friedman wrote: > > > You're lucky. All I get are kernel panics when using my webcam with > > > the spca5xx module in FC5. The author seems completely uninsterested > > > in investigating since it "works for him". > > > > God Lonnie, you're slipping! You let him get away with that?? I cannot imagine. Ric > > He's in France, I'm not. I can't force him to fix his bugs, and quite > honestly, I think he's not capable of debugging it. > I think you don't have to say this without any proof. There is some drivers which have problems to support fedora mainly due to kernel versioning (see for example http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=1618 ). I don't say this is the problem with spca5xx driver i only say it is not so simple as you said ...
No, that's not the issue I'm experiencing. The spca5xx module builds just fine on my FC5 system. The problem is that as soon as its used, I encounter an immediate kernel panic. I have proof that the developer has taken zero interest in attempting to debug the problem. Initially he was claiming that FC kernels are just buggy, so I tried a vanillla kernel.org kernel, and that, too, exhibited the same problem (kernel panic). His response was that it had to be a gcc-4.x issue. This turned out to be partly true, as rebuilding the kernel with gcc-3.x did eliminate the problem. However seeing as how every other in kernel module isn't having this problem in a kernel built with gcc-4.x, i'm more inclined to believe that spca5xx is broken with gcc-4.x, not that gcc-4.x is broken with spca5xx. At any rate, I pushed back on the developer noting that at some point he needs to fix this problem or work with the gcc folks, as he can't just rely on gcc-3.x forever. His response was that it must be a FC problem, since he had upgraded to SuSE-10 which had gcc-4.x and he wasn't having the problem there, and its likely a lack of 8k stacks in the FC x86 kernels. That is complete nonsense since all kernel.org x86 kernels build with 4k stacks, and nothing greater is even possible without alot of hacking. I pointed that out to the developer, and he never replied. I sent another email asking him for assistance, and there was no reply. The bottom line is that this guy clearly has no clue what is wrong, and as long as it "works for him" he has no interest in investigating. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ L. Friedman netllama@xxxxxxxxx LlamaLand http://netllama.linux-sxs.org
