>From: Nishanth Menon on Wednesday, January 20, 2010 6:02 AM
>
>Ameya Palande said the following on 01/20/2010 04:01 AM:
>> Hi Nishanth,
>>
>> On Wed, 2010-01-20 at 02:44 +0100, ext Nishanth Menon wrote:
>>
>>> If DSP baseimage requests for invalid OPPs, do not pass them
>>> to the rest of the linux kernel subsystem. with a test baseimage
>>> it was found to send
>>> IO_DispatchPM: WMDIOCTL_CONSTRAINT_REQUEST with parameter 0.
>>> this is an invalid OPP ID and should not be passed anywhere.
>>> we add a param check for this prior to passing it along to the
>>> workqueue.
>>>
>>> Cc: Ameya Palande <ameya.palande@xxxxxxxxx>
>>> Cc: Deepak Chitriki <deepak.chitriki@xxxxxx>
>>> Cc: Felipe Contreras <felipe.contreras@xxxxxxxxx>
>>> Cc: Hiroshi Doyu <hiroshi.doyu@xxxxxxxxx>
>>> Cc: Omar Ramirez Luna <omar.ramirez@xxxxxx>
>>>
>>> Signed-off-by: Nishanth Menon <nm@xxxxxx>
>>> ---
>>> drivers/dsp/bridge/wmd/io_sm.c | 9 ++++++---
>>> 1 files changed, 6 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/dsp/bridge/wmd/io_sm.c b/drivers/dsp/bridge/wmd/io_sm.c
>>> index 6c140c5..618bedb 100644
>>> --- a/drivers/dsp/bridge/wmd/io_sm.c
>>> +++ b/drivers/dsp/bridge/wmd/io_sm.c
>>> @@ -960,9 +960,12 @@ static void IO_DispatchPM(struct work_struct *work)
>>> pArg[1] = pIOMgr->pSharedMem->oppRequest.rqstOppPt;
>>> DBG_Trace(DBG_LEVEL7, "IO_DispatchPM : Value of OPP "
>>> "value =0x%x \n", pArg[1]);
>>> - status = pIOMgr->pIntfFxns->pfnDevCntrl(pIOMgr->
>>> - hWmdContext, WMDIOCTL_CONSTRAINT_REQUEST,
>>> - pArg);
>>> + if (!pArg[1])
>>> + status = DSP_EINVALIDARG;
>>> + else
>>> + status = pIOMgr->pIntfFxns->pfnDevCntrl(pIOMgr->
>>> + hWmdContext,
>>> + WMDIOCTL_CONSTRAINT_REQUEST, pArg);
>>> if (DSP_FAILED(status)) {
>>> DBG_Trace(DBG_LEVEL7, "IO_DispatchPM : Failed "
>>> "to set constraint = 0x%x \n",
>>>
>>
>> Instead of checking just for 0, isn't it better to check for any value
>> outside of valid OPPs? (Especially if we need robust implementation
>> which shouldn't break because of DSP BIOS blackbox!).
Dsp will ask only for the opps programmed to the shared memory... so no need, as long as the shared memory is filled with the right values.
>>
>>
>yeah, i agree. should check for the max level also. dropping this and
>will resend.
Dsp doesn't request opp 0 to kernel.
- omar
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
