* [PATCH v2] memstick: rtsx: fix ms card data transfer bug
@ 2013-11-20 9:54 micky_ching
2014-01-24 11:59 ` Maxim Levitsky
0 siblings, 1 reply; 2+ messages in thread
From: micky_ching @ 2013-11-20 9:54 UTC (permalink / raw)
To: devel, linux-kernel
Cc: sameo, maximlevitsky, gregkh, oakad, rogerable, wei_wang, Micky Ching
From: Micky Ching <micky_ching@realsil.com.cn>
This patch is used to add support for ms card. The main difference
between ms card and mspro card is long data transfer mode. mspro card
can use auto mode DMA for long data transfer, but ms can not use this
mode, it should use normal mode DMA.
The memstick core add support for ms card, but the original driver will
make ms card fail at initialize, because it use auto mode DMA. This
patch can make ms card work properly.
Signed-off-by: Micky Ching <micky_ching@realsil.com.cn>
---
drivers/memstick/host/rtsx_pci_ms.c | 30 ++++++++++++++++++++----------
1 file changed, 20 insertions(+), 10 deletions(-)
diff --git a/drivers/memstick/host/rtsx_pci_ms.c b/drivers/memstick/host/rtsx_pci_ms.c
index 25f8f93..2a635b6 100644
--- a/drivers/memstick/host/rtsx_pci_ms.c
+++ b/drivers/memstick/host/rtsx_pci_ms.c
@@ -145,6 +145,8 @@ static int ms_transfer_data(struct realtek_pci_ms *host, unsigned char data_dir,
unsigned int length = sg->length;
u16 sec_cnt = (u16)(length / 512);
u8 val, trans_mode, dma_dir;
+ struct memstick_dev *card = host->msh->card;
+ bool pro_card = card->id.type == MEMSTICK_TYPE_PRO;
dev_dbg(ms_dev(host), "%s: tpc = 0x%02x, data_dir = %s, length = %d\n",
__func__, tpc, (data_dir == READ) ? "READ" : "WRITE",
@@ -152,19 +154,21 @@ static int ms_transfer_data(struct realtek_pci_ms *host, unsigned char data_dir,
if (data_dir == READ) {
dma_dir = DMA_DIR_FROM_CARD;
- trans_mode = MS_TM_AUTO_READ;
+ trans_mode = pro_card ? MS_TM_AUTO_READ : MS_TM_NORMAL_READ;
} else {
dma_dir = DMA_DIR_TO_CARD;
- trans_mode = MS_TM_AUTO_WRITE;
+ trans_mode = pro_card ? MS_TM_AUTO_WRITE : MS_TM_NORMAL_WRITE;
}
rtsx_pci_init_cmd(pcr);
rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, MS_TPC, 0xFF, tpc);
- rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, MS_SECTOR_CNT_H,
- 0xFF, (u8)(sec_cnt >> 8));
- rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, MS_SECTOR_CNT_L,
- 0xFF, (u8)sec_cnt);
+ if (pro_card) {
+ rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, MS_SECTOR_CNT_H,
+ 0xFF, (u8)(sec_cnt >> 8));
+ rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, MS_SECTOR_CNT_L,
+ 0xFF, (u8)sec_cnt);
+ }
rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, MS_TRANS_CFG, 0xFF, cfg);
rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, IRQSTAT0,
@@ -192,8 +196,14 @@ static int ms_transfer_data(struct realtek_pci_ms *host, unsigned char data_dir,
}
rtsx_pci_read_register(pcr, MS_TRANS_CFG, &val);
- if (val & (MS_INT_CMDNK | MS_INT_ERR | MS_CRC16_ERR | MS_RDY_TIMEOUT))
- return -EIO;
+ if (pro_card) {
+ if (val & (MS_INT_CMDNK | MS_INT_ERR |
+ MS_CRC16_ERR | MS_RDY_TIMEOUT))
+ return -EIO;
+ } else {
+ if (val & (MS_CRC16_ERR | MS_RDY_TIMEOUT))
+ return -EIO;
+ }
return 0;
}
@@ -462,8 +472,8 @@ static int rtsx_pci_ms_set_param(struct memstick_host *msh,
clock = 19000000;
ssc_depth = RTSX_SSC_DEPTH_500K;
- err = rtsx_pci_write_register(pcr, MS_CFG,
- 0x18, MS_BUS_WIDTH_1);
+ err = rtsx_pci_write_register(pcr, MS_CFG, 0x58,
+ MS_BUS_WIDTH_1 | PUSH_TIME_DEFAULT);
if (err < 0)
return err;
} else if (value == MEMSTICK_PAR4) {
--
1.7.9.5
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH v2] memstick: rtsx: fix ms card data transfer bug
2013-11-20 9:54 [PATCH v2] memstick: rtsx: fix ms card data transfer bug micky_ching
@ 2014-01-24 11:59 ` Maxim Levitsky
0 siblings, 0 replies; 2+ messages in thread
From: Maxim Levitsky @ 2014-01-24 11:59 UTC (permalink / raw)
To: micky_ching
Cc: devel, linux-kernel, sameo, gregkh, oakad, rogerable, wei_wang
On Wed, 2013-11-20 at 17:54 +0800, micky_ching@realsil.com.cn wrote:
> From: Micky Ching <micky_ching@realsil.com.cn>
>
> This patch is used to add support for ms card. The main difference
> between ms card and mspro card is long data transfer mode. mspro card
> can use auto mode DMA for long data transfer, but ms can not use this
> mode, it should use normal mode DMA.
>
> The memstick core add support for ms card, but the original driver will
> make ms card fail at initialize, because it use auto mode DMA. This
> patch can make ms card work properly.
Due to unexpected turn of events I bought an laptop with this card
reader.
So as soon as I have some time, I will too test this code
Really big thanks for writing this driver!
Best regards,
Maxim Levitsky
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2014-01-24 11:59 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-11-20 9:54 [PATCH v2] memstick: rtsx: fix ms card data transfer bug micky_ching
2014-01-24 11:59 ` Maxim Levitsky
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®