d before and after Link startup is carried out
> + * to allow variant specific Uni-Pro initialization.
> + */
> +struct ufs_hba_variant_ops {
> + const char *name;
> + int (*init)(struct ufs_hba *);
> + void(*exit)(struct ufs_hba *);
> +
On Wed, July 24, 2013, Sujit Reddy Thumma wrote:
> On 7/23/2013 2:04 PM, Seungwon Jeon wrote:
> > On Sat, July 20, 2013, Sujit Reddy Thumma wrote:
> >> On 7/19/2013 7:28 PM, Seungwon Jeon wrote:
> >>> On Tue, July 09, 2013, Sujit Reddy Thumma wrote:
> >>&g
On Wed, July 24, 2013, Sujit Reddy Thumma wrote:
> On 7/23/2013 1:57 PM, Seungwon Jeon wrote:
> > On Sat, July 20, 2013, Sujit Reddy Thumma wrote:
> >> On 7/19/2013 7:27 PM, Seungwon Jeon wrote:
> >>> On Tue, July 09, 2013, Sujit Reddy Thumma wrote:
> >>>
On Sat, July 20, 2013, Sujit Reddy Thumma wrote:
> On 7/19/2013 7:28 PM, Seungwon Jeon wrote:
> > On Tue, July 09, 2013, Sujit Reddy Thumma wrote:
> >> Error handling in UFS driver is broken and resets the host controller
> >> for fatal errors without re-initializati
On Sat, July 20, 2013, Sujit Reddy Thumma wrote:
> On 7/19/2013 7:27 PM, Seungwon Jeon wrote:
> > On Tue, July 09, 2013, Sujit Reddy Thumma wrote:
> >> As of now SCSI initiated error handling is broken because,
> >> the reset APIs don't try to bring back the devic
On Sat, July 20, 2013, Sujit Reddy Thumma wrote:
> On 7/19/2013 7:26 PM, Seungwon Jeon wrote:
> > On Tue, July 09, 2013 Sujit Reddy Thumma wrote:
> >> Currently, sending Task Management (TM) command to the card might
> >> be broken in some scenarios as listed below:
>
t fatal UIC error flags.
> + * @hba: per-adapter instance
> */
> -static void ufshcd_err_handler(struct ufs_hba *hba)
> +static void ufshcd_update_uic_error(struct ufs_hba *hba)
> {
> u32 reg;
>
> + /* PA_INIT_ERROR is fatal and needs UIC reset */
&
On Tue, July 09, 2013, Sujit Reddy Thumma wrote:
> There is a possible race condition in the hardware when the abort
> command is issued to terminate the ongoing SCSI command as described
> below:
>
> - A bit in the door-bell register is set in the controller for a
> new SCSI command.
> - In som
ED;
> + hba->ufshcd_state = UFSHCD_STATE_ERROR;
> + }
> + spin_unlock_irqrestore(hba->host->host_lock, flags);
> +
> + return err;
> +}
> +
> +/**
> + * ufshcd_eh_host_reset_handler - host reset handler registered to scsi layer
> + * @
__func__, tm_function);
> + if (ufshcd_clear_tm_cmd(hba, free_slot))
> + dev_WARN(hba->dev, "%s: unable clear tm cmd (slot %d)
> after timeout\n",
> + __func__, free_slot);
> + err = -ET
are raised only when the device is in critical need for
> bkops. Also the spec. recommends, host should ensure that the device
> doesn't go into such states.
>
> With your suggestion, when we disable bkops, the exception is raised and
> we enable bkops after which there is
tine.
> > Could you explain the reason dynamic allocation is needed?
> I don't like the idea of allocating it on stack, I suggest to statically
> allocate it on the query struct. What do you think?
Yes.
> >
> >> >> +
> >> >> + switch (opcode
REG_UTP_TRANSFER_REQ_DOOR_BELL,
> >>>> +mask, 0, 1000, 1000);
> >>> 4th argument should be (~mask) instead of '0', right?
> >>
> >> True, but not really for this implementation. It breaks the check in
> >> i
On Thu, July 11, 2013, Sujit Reddy Thumma wrote:
> On 7/10/2013 7:01 PM, Seungwon Jeon wrote:
> > I'm not sure that BKOPS with runtime-pm associates.
> > Do you think it's helpful for power management?
> > How about hibernation scheme for runtime-pm?
> >
> >> + /* Send query request */
> >> + err = ufshcd_send_query_request(hba, query, NULL, response);
> >> +
> >> + if (err) {
> >> + dev_err(hba->dev,
> >> + "%s: Sending flag query for idn %d failed, err = %d\
On Thu, July 11, 2013, Sujit Reddy Thumma wrote:
> On 7/10/2013 6:58 PM, Seungwon Jeon wrote:
> > On Tue, July 09, 2013, Sujit Reddy Thumma wrote:
> >> As part of device initialization sequence, sending NOP OUT UPIU and
> >> waiting for NOP IN UPIU response is mandat
runtime_put_sync(hba->dev);
> }
>
> /**
> @@ -2562,6 +2566,7 @@ static void ufshcd_async_scan(void *data,
> async_cookie_t cookie)
> hba->auto_bkops_enabled = false;
In the end of 'ufshcd_init', pm_runtime_get_sync is called, hence
auto_bkops_enabled may h
fshcd_query_attr(struct ufs_hba *hba, enum query_opcode opcode,
> + enum attr_idn idn, u8 index, u8 selector, u32 *attr_val)
> +{
> + struct ufs_query_req *query;
> + struct ufs_query_res *response;
> + int err = -ENOMEM;
> +
> + if (!attr
function is for initialization completion.
How about ufshcd_complete_dev_init for function name?
> +{
> + int i, retries, err = 0;
> + bool flag_res = 0;
> +
> + for (retries = QUERY_REQ_RETRIES; retries > 0; retries--) {
> + /* Set the fDeviceIntit
);
> -
> - /*
> - * get the result based on SCSI status response
> - * to notify the SCSI midlayer of the command status
> - */
> - scsi_status = result & MASK_SCSI_STATUS;
> - result = ufshcd_scsi_c
'From: Author name ' is expected in the head of e-mail.
I feel that maintainer has this role, though.
And there is need to check the base of tree for ufshcd.
' scsi: ufs: add support for query requests' is not merged finally.
I think we can talk with Santosh for those.
Thanks,
S
On Sunday, May 12, 2013, Dolev Raviv wrote:
Author should be identified here. (From: author...)
Except last patch of series(8/8), all of patches are not from you.
> Simplify operations with hiding mmio_base.
>
> Signed-off-by: Seungwon Jeon
> Signed-off-by: Dolev Raviv
I'm n
rr = ufshcd_wait_for_nop_cmd(hba, lrbp);
> +
> + if (err == -ETIMEDOUT) {
> + u32 reg;
> +
> + /* clear outstanding transaction before retry */
> + spin_lock_irqsave(hba->host->host_lock, flags);
> + ufshcd_utrl_clear(hba, INTERNAL_CMD_TAG);
c_suspend(struct mmc_host *host)
> >> else if (!mmc_host_is_spi(host))
> >> err = mmc_deselect_cards(host);
> >> host->card->state &= ~(MMC_STATE_HIGHSPEED |
> > MMC_STATE_HIGHSPEED_200);
> >> - mmc_release_host(host);
>
.
Depending on user scenario, write can be considered to be important.
It needs to check the starvation of write.
Thanks,
Seungwon Jeon
> Hi Chung
>
>
> > The below is my environment:
> > Kernel Version: linux-3.4
> > Card: eMMC4.5 (50MHz DDR mode, 8-bit buswidth) Ho
e
> requests between read requests in parallel read and write operations
> doesn't exceed 17 requests. Therefore, we can assume that fetching more
> than 17 write requests without hitting a read request can indicate that
> there is no read activity.
We can apply this experiment reg
rite requests are over
num_wr_reqs_to_start_packing?
That means individual request(1...17) will be issued with non-packing.
Could you explain your policy more?
> +
> +}
> +
> static u8 mmc_blk_prep_packed_list(struct mmc_queue *mq, struct request *req)
> {
> struct request_q
27 matches
Mail list logo
