- 21 Oct, 2009 19 commits
-
-
Omar Ramirez Luna authored
Switching find_task_by_vpid call to pid_task, according to clean ups in pid module. Reference: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=17f98dcf6010a1cfd25d179fd0ce77d3dc2685c3Signed-off-by: Omar Ramirez Luna <omar.ramirez@ti.com>
-
Omar Ramirez Luna authored
For newer kernels the ioctl numbers have changed, passing ioctl number 1 or 2, will call kernel functions inside the ioctl syscall switch. This patch changes the base number for bridge ioctls to start on 101. For this to work, the same change has to be propagated to bridge API. This patch should be temporary until the _IOx kernel calls are used to define ioctl numbers withing bridge driver. Signed-off-by: Omar Ramirez Luna <omar.ramirez@ti.com>
-
Ramesh Gupta G authored
This patch provides OMAP4 Syslink release notes for L24.0.9 release. Signed-off-by: Ramesh Gupta G <grgupta@ti.com> Signed-off-by: Suman Anna <s-anna@ti.com>
-
Ramesh Gupta authored
This patch adds notify dependancies for Tesla bridgedriver Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
This patch provides OMAP4 Syslink release notes for L24.0.8 release. Signed-off-by: Ramesh Gupta G <grgupta@ti.com> Signed-off-by: Suman Anna <s-anna@ti.com>
-
Ramesh Gupta authored
This patch fixes coding style in tesla driver sources. Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
This patch fixes minor checkpatch warnings. Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
This patch fixes coding style and compilation warnings. Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
Use kernel bit operation functions instead of custom defined. Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
This patch cleans up tesla header files Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
Fixed typo in Makefile to build dspbridge Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Ramesh Gupta authored
Intial Bridge changes for new tree These are the bridge changes for OMAP4 and Notify Support. Require cleanup. Signed-off-by: Ramesh Gupta G <grgupta@ti.com>
-
Hari Kanigeri authored
Provide build support for tesla in omap4 setup Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
Hari Kanigeri authored
Port 3430 bridge files for tesla support. Also disables PM.
-
Hari Kanigeri authored
This patch adds tesla driver to teslabridge tree Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
C A Subramaniam authored
Fixed the bug where the L4 peripheral address table was getting corrupted after calling mmu module's function to update TLB entry. this resulted in the consecutive runs failure as wrong addresses gets programmed due to L4 peripheral table corruption. Signed-off-by: C A Subramaniam <subramaniam.ca@ti.com> Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
- 20 Oct, 2009 2 commits
-
-
Simon Que authored
In sysmgr_ioctl, the execution of CMD_SYSMGR_STOPCALLBACK does not set the return value in cargs.api_status. This caused an error sometimes because the status was uninitialized. Now it sets the status to 0 just like for the cases of CMD_SYSMGR_LOADCALLBACK and CMD_SYSMGR_STARTCALLBACK. Note that since the platform callbacks return void, we must set the api_status manually. Signed-off-by: Simon Que <sque@ti.com>
-
Simon Que authored
In proc4430, ducati_destroy() is called from both proc4430_destroy() and proc4430_stop(). Since ducati_setup() is called from proc4430_start(), the call to ducati_destroy() in proc4430_destroy() is invalid. It generates a crash. Signed-off-by: Simon Que <sque@ti.com>
-
- 16 Oct, 2009 19 commits
-
-
Hari Kanigeri authored
This fixes the bug in hw mmu module where the table entries that has memory segment information was getting corrupted after first run. This bug causes the consecutive runs to fail. This patch also fixes the memory leak associated with not calling iounmap and fixes a compile warning as one of the function is unused. Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
Suman Anna authored
Failed statements are printed in the IPC module xxx_destroy() functions only after checking if the status value indicates an error condition. This is to avoid confusion by the existing trace statement word 'failed'. This statement was being printed if the xxx_destroy jumps to exit label because the module's reference count is not yet zero. Signed-off-by: Suman Anna <s-anna@ti.com>
-
C A Subramaniam authored
This patch modifies the memory addresses in sysmgr to be in line with the new memory map changes required to use the TLB entries for Ducati. Signed-off-by: C A Subramaniam <subramaniam.ca@ti.com> Signed-off-by: Suman Anna <s-anna@ti.com>
-
Ramesh Gupta authored
This fixes the 16MB TLB issue. Signed-off-by: Ramesh Gupta G<grgupta@ti.com> Signed-off-by: C A Subramaniam <subramaniam.ca@ti.com>
-
Suman Anna authored
This patch programs all the required memory regions for Ducati as TLBs to provide a sw solution for the hw bug in SDC. A total of 49MB of physical memory is mapped, and the physical address is assumed to begin at 0x88000000. Signed-off-by: Suman Anna <s-anna@ti.com>
-
Hari Kanigeri authored
This patch de-couples SysManager calls from Proc module. Added new SysManger calls that get invoked from user space. Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
C A Subramaniam authored
This patch for the notify module fixes the following issues: - Random signal of exit event given to the user thread in the read ioctl call. - Memory leak because of the absence of iounmap() call. Signed-off-by: C A Subramaniam <subramaniam.ca@ti.com>
-
C A Subramaniam authored
Added the following changes - Removed array that was used to track ducati driver objects. - Removed array introduced in syslink 2-0-0-12 migration to track notify driver. - Introduced a list implementation for tracking notify objects. - Removed the index variable introduced in syslink 2-0-0-12 migration. - Re-organized notify_ducatidrv_create() function call. - Added support for multiple creates to be called with same driver name. Signed-off-by: C A Subramaniam <subramaniam.ca@ti.com>
-
Hari Kanigeri authored
This patch adds the changes to Proc start and Proc stop based on the validation on Zebu. Proc start function is extended to enable M3 clock and program Ducati MMU. Proc Stop is extended to put Cores in reset and disable M3 clock, and the option to disable individual Core and MMU. Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
Hari Kanigeri authored
-
Simon Que authored
platform_stop_callback now reads the config info for each module, just like platform_start_callback does. The problem was that it didn't read it before, nor did it save the config info from platform_start_callback for each Ducati core separately. Thus only the AppM3 config was saved, and then used for SysM3 cleanup. Signed-off-by: Simon Que <sque@ti.com>
-
Hari Kanigeri authored
This patch fixes the bug in notify drv read where the read thread wouldn't get blocked if the event list is empty. This bug resulted in board freeze in Zebu enviornment because the read thread was blockign all other threads from getting scheduled. Signed-off-by: Hari Kanigeri<h-kanigeri2@ti.com>
-
Hari Kanigeri authored
Fix the casting compile warning in platform destroy. Signed-off-by: Hari Kanigeri <h-kanigeri2@ti.com>
-
Imtiaz Ahmed authored
Added probe and remove functions, added platform_device_add, platform_device_alloc calls and also added few dev_dbgs to demo the usage of dev_dbg. Signed-off-by: Imtiaz Ahmed <x0116819@ti.com>
-
Jayan John authored
This patch fixes the issues found in.. <1> notify_ducati.c: we were not unmapping "driver_obj->ctrl_ptr". This is done using unmap_ducati_virt_mem() implemented now in ducatienabler. <2> ducatienabler.c: implemented unmap_ducati_virt_mem() to iounmap() the "shm_virt_addr" we get calling ioremap() in get_ducati_virt_mem(). <3> ducatienabler.h: added unmap_ducati_virt_mem(). <4> notify_ducatidriver.h: added unmap_ducati_virt_mem(). Signed-off-by: Jayan John <x00jayan@ti.com>
-
Jayan John authored
This patch fixes an issue found in notify_dispatcher.c where we were not calling iounmap() on "mailbx_hw_config.mbox_linear_addr" which we had obtained using ioremap() earlier in ntfy_disp_init(). Signed-off-by: Jayan John <x00jayan@ti.com>
-
Jayan John authored
This patch fixes an issue found in platform.c where we were not unlocking "platform_mem_state.gate" in case of a ioremap() failure. Signed-off-by: Jayan John <x00jayan@ti.com>
-
Jayan John authored
This patch fixes the issues found in.. <1> platform.c: - calling platform_mem_unmap() on all addreses on which we have called platform_mem_map(). - calling proc_mgr_detach for SysM3 and AppM3. <2> -proc4430.c: - proc4430_detach() fixed to unmap all reqd. addresses. - cleanup Signed-off-by: Jayan John <x00jayan@ti.com>
-
Simon Que authored
When a heapbuf object is created, there are two memory buffers allocated -- one for the handle and one for the object itself. That is in _heapbuf_create. In heapbuf_delete, both are freed. But in heapbuf_close, only the handle is freed. The object pointer is not freed and caused a memory leak. This patch adds a call to free the object. Signed-off-by: Simon Que <sque@ti.com>
-