Menu
▾
▴
acx100-devel
|
From: Oliver W. <oli...@ol...> - 2012-05-15 19:17:46
|
On 05/15/2012 09:00 PM, Jim Cromie wrote: >> >> Im attaching a test script Ive been using to bang on the driver >> Dot it into your shell, and go from there. >> How about putting the script into a 'scripts' sub-dir in the tree ? For me that's fine. Cheers, Oliver > oof. now attached. > > > ieee80211 phy0: device now idle |
|
From: Larry F. <Lar...@lw...> - 2012-05-15 20:02:31
|
On 05/15/2012 02:17 PM, Oliver Winker wrote: > On 05/15/2012 09:00 PM, Jim Cromie wrote: >>> >>> Im attaching a test script Ive been using to bang on the driver >>> Dot it into your shell, and go from there. >>> > > How about putting the script into a 'scripts' sub-dir in the tree ? For > me that's fine. Sounds good to me. Larry |
|
From: Jim C. <jim...@gm...> - 2012-05-15 20:47:36
|
hey, this one seems to have problems: Ive gotten a couple null-pointer oopses. common.c: set debug-flags by symbolic names I'll look into it. |
|
From: Oliver W. <oli...@ol...> - 2012-05-17 12:10:30
|
Hi Jim, Larry,
Ok, the series and pushed! Thanks a lot again for this heap of work :)!!
There are several changes regarding builds and some functional fixes.
You'll see in the commits messages.
I had also to revert b4661bf5 (CONFIG_PCI vs. CONFIG_ACX_MAC80211_PCI,
see commit message). Maybe a runtime check would still be better here.
The sparse checks are interesting. I'll have also a look at some one time.
There regarding debugfs.c:
There were warnings, which have been fixed in 740058 and then seem to
have been rolled back again due to sparse warnings.
I checked those and, at least from my perspective, could fix them for
both gcc and sparse - see patch below.
Is this ok to commit, or is there something, that missed maybe ?
Best Regards, Oliver
---
> CC [M] /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.o
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: In function ‘acx_dbgfs_open’:
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:61:13: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: In function ‘acx_dbgfs_write’:
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:90:13: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: At top level:
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:2: warning: initialization from incompatible pointer type [enabled by default]
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:2: warning: (near initialization for ‘acx_fops.write’) [enabled by default]
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: In function ‘acx_debugfs_add_adev’:
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:158:6: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
---
Then I get also still these sparse warning in debugfs.c:
---
> CHECK /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:27: warning: incorrect type in initializer (different type sizes)
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:27: expected long ( *write )( ... )
> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:27: got int ( static [toplevel] *<noident> )( ... )
---
This patch seems to satisfy both, gcc and sparse for x86_32 and x86_64:
---
> diff --git a/debugfs.c b/debugfs.c
> index cd403bc..d6bbe7e 100644
> --- a/debugfs.c
> +++ b/debugfs.c
> @@ -58,7 +58,7 @@ BUILD_BUG_DECL(dbgfs_files__VS__enum_REG_DOMAIN,
>
> static int acx_dbgfs_open(struct inode *inode, struct file *file)
> {
> - int fidx = (int) inode->i_private;
> + int fidx = (uintptr_t) inode->i_private;
> struct acx_device *adev = (struct acx_device *)
> file->f_path.dentry->d_parent->d_inode->i_private;
>
> @@ -83,11 +83,11 @@ static int acx_dbgfs_open(struct inode *inode, struct file *file)
> return single_open(file, acx_proc_show_funcs[fidx], adev);
> }
>
> -static int acx_dbgfs_write(struct file *file, const char __user *buf,
> +static ssize_t acx_dbgfs_write(struct file *file, const char __user *buf,
> size_t count, loff_t *ppos)
> {
> /* retrieve file-index and adev from private fields */
> - int fidx = (int) file->f_path.dentry->d_inode->i_private;
> + int fidx = (uintptr_t) file->f_path.dentry->d_inode->i_private;
> struct acx_device *adev = (struct acx_device *)
> file->f_path.dentry->d_parent->d_inode->i_private;
>
> @@ -155,7 +155,7 @@ int acx_debugfs_add_adev(struct acx_device *adev)
> /* save file-index in in file's private field */
> file = debugfs_create_file(dbgfs_files[i], fmode,
> acx_dbgfs_devdir,
> - (void*) i, &acx_fops);
> + (void *) (uintptr_t) i, &acx_fops);
> if (!file)
> goto fail;
> }
---
On 05/15/2012 09:50 AM, Jim Cromie wrote:
> hi Oliver, Larry,
>
> heres another batch of merges, sparse fixes, comment wraps and grammar
> fixes, and various other stuff. Another chance to amortize your testing efforts over more changes ;-)
>
>
> common.c | 188 +++++++++++++++++++----------
> debugfs.c | 26 ++--
> inlines.h | 75 +++++-------
> mem.c | 55 +++++----
> mem.h | 4 +-
> merge.c | 408 +++++++++++++++++++++++++++++++-------------------------------
> merge.h | 204 +++++++++++++++++++++++--------
> pci.c | 23 ++--
> 8 files changed, 569 insertions(+), 414 deletions(-)
>
|
|
From: Larry F. <Lar...@lw...> - 2012-05-17 14:43:55
|
On 05/17/2012 07:10 AM, Oliver Winker wrote:
> Hi Jim, Larry,
>
> Ok, the series and pushed! Thanks a lot again for this heap of work :)!!
>
> There are several changes regarding builds and some functional fixes.
> You'll see in the commits messages.
>
> I had also to revert b4661bf5 (CONFIG_PCI vs. CONFIG_ACX_MAC80211_PCI,
> see commit message). Maybe a runtime check would still be better here.
>
> The sparse checks are interesting. I'll have also a look at some one time.
>
> There regarding debugfs.c:
>
> There were warnings, which have been fixed in 740058 and then seem to
> have been rolled back again due to sparse warnings.
>
> I checked those and, at least from my perspective, could fix them for
> both gcc and sparse - see patch below.
>
> Is this ok to commit, or is there something, that missed maybe ?
>
> Best Regards, Oliver
>
> ---
>> CC [M] /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.o
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: In function ‘acx_dbgfs_open’:
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:61:13: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: In function ‘acx_dbgfs_write’:
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:90:13: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: At top level:
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:2: warning: initialization from incompatible pointer type [enabled by default]
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:2: warning: (near initialization for ‘acx_fops.write’) [enabled by default]
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c: In function ‘acx_debugfs_add_adev’:
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:158:6: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
> ---
>
> Then I get also still these sparse warning in debugfs.c:
> ---
>> CHECK /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:27: warning: incorrect type in initializer (different type sizes)
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:27: expected long ( *write )( ... )
>> /media/OE/Devel/Git-acx-mac80211/acx-mac80211/debugfs.c:117:27: got int ( static [toplevel] *<noident> )( ... )
> ---
>
> This patch seems to satisfy both, gcc and sparse for x86_32 and x86_64:
> ---
>> diff --git a/debugfs.c b/debugfs.c
>> index cd403bc..d6bbe7e 100644
>> --- a/debugfs.c
>> +++ b/debugfs.c
>> @@ -58,7 +58,7 @@ BUILD_BUG_DECL(dbgfs_files__VS__enum_REG_DOMAIN,
>>
>> static int acx_dbgfs_open(struct inode *inode, struct file *file)
>> {
>> - int fidx = (int) inode->i_private;
>> + int fidx = (uintptr_t) inode->i_private;
>> struct acx_device *adev = (struct acx_device *)
>> file->f_path.dentry->d_parent->d_inode->i_private;
>>
>> @@ -83,11 +83,11 @@ static int acx_dbgfs_open(struct inode *inode, struct file *file)
>> return single_open(file, acx_proc_show_funcs[fidx], adev);
>> }
>>
>> -static int acx_dbgfs_write(struct file *file, const char __user *buf,
>> +static ssize_t acx_dbgfs_write(struct file *file, const char __user *buf,
>> size_t count, loff_t *ppos)
>> {
>> /* retrieve file-index and adev from private fields */
>> - int fidx = (int) file->f_path.dentry->d_inode->i_private;
>> + int fidx = (uintptr_t) file->f_path.dentry->d_inode->i_private;
>> struct acx_device *adev = (struct acx_device *)
>> file->f_path.dentry->d_parent->d_inode->i_private;
>>
>> @@ -155,7 +155,7 @@ int acx_debugfs_add_adev(struct acx_device *adev)
>> /* save file-index in in file's private field */
>> file = debugfs_create_file(dbgfs_files[i], fmode,
>> acx_dbgfs_devdir,
>> - (void*) i,&acx_fops);
>> + (void *) (uintptr_t) i,&acx_fops);
>> if (!file)
>> goto fail;
>> }
> ---
>
I had noticed the warnings from debugfs on x86_64 and prepared a patch. As I
decided that you had enough to test and merge, I have been holding it back. Your
solution looks like mine with one exception. In acx-debugfs_add_adev(), I
changed "i" to a long rather than using the uintptr_t cast. AFAIK, the solutions
are equal.
I purchased a DWL-120+ on E-bay yesterday. When it arrives, I hope that turns
out to be an acx100. The one labeled as DWL-120 that I asked you about turns out
to use at76c50x-usb as a driver. It wasn't a total loss, as that driver crashes
my kernels, thus it needs some work. I hope to get to it someday.
Thanks for your efforts on merging the code.
|
Thanks for helping keep SourceForge clean.
X