memstick: fix unreachable state in h_msb_read_page() in ms_block.c
PATCH 1/1 memstick: fix unreachable state in h_msb_read_page() in ms_block.c
From: Roger Tseng <[email protected]>
In h_msb_read_page() in ms_block.c, flow never reaches case MSB_RP_RECIVE_STATUS_REG. This causes error when MEMSTICK_INT_ERR is encountered and status error bits are going to be examined, but the status will never be copied back.
Fix it by transiting to MSB_RP_RECIVE_STATUS_REG right after MSB_RP_SEND_READ_STATUS_REG.
--- drivers/memstick/core/ms_block.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/memstick/core/ms_block.c +1 -1
@@ -401,7 +401,7 @@ again: sizeof(struct ms_status_register)))return 0 ; - msb->state =MSB_RP_RECEIVE_OOB_READ ; + msb->state =MSB_RP_RECIVE_STATUS_REG ;return 0 ;case MSB_RP_RECIVE_STATUS_REG :
Signature · 1 line
1.8.2
Maxim Levitsky
maintainerQuoted 20 lines · On Wed, 2013-10-30 at 15:54 +0800, [email protected] wrote:
From: Roger Tseng <[email protected]>
In
h_msb_read_page()inms_block.c, flow never reaches caseMSB_RP_RECIVE_STATUS_REG. This causes error whenMEMSTICK_INT_ERRis encountered and status error bits are going to be examined, but the status will never be copied back.Fix it by transiting to
MSB_RP_RECIVE_STATUS_REGright afterMSB_RP_SEND_READ_STATUS_REG.Signed-off-by Roger Tseng <[email protected]>--- drivers/memstick/core/ms_block.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/memstick/core/ms_block.c +0
-0
drivers/memstick/core/ms_block.c again: · line 407
Quoted 8 lines
drivers/memstick/core/ms_block.c +1
-1@@ -401,7 +401,7 @@ again: sizeof(struct ms_status_register)))return 0 ; - msb->state =MSB_RP_RECEIVE_OOB_READ ; + msb->state =MSB_RP_RECIVE_STATUS_REG ;return 0 ;case MSB_RP_RECIVE_STATUS_REG :
This is indeed a bug here. Probably added when I converted to use #defines for states.
So,
Btw, there is another minor or major (pun inteded) bug out there in this driver, which I still have no time to send patch for yet :-(
I accidentally didn't set the major and minor for the disk. It happens to work as usually there is just one card reader out there, and major is allocated dynamically.
The two lines in msb_init_disk,
msb->disk->major = major; msb->disk->first_minor = disk_id << MS_BLOCK_PART_SHIFT;
Somehow fell through the cracks. I had this in older revision.
Feel free to add it, or maybe I find time to compile and test this and send a patch.
You also welcome to write a driver for xD portion of the card reader too.
I back then implemented all the generic support, and you only need to expose an NAND driver (look at r852.c for my code).
Signature · 2 lines
Best regards,
Maxim Levitsky