Re: [PATCH v4 1/2] arcmsr: simplify ioctl data read/write

[Tomas Henzl <thenzl@redhat.com>]

On 09/12/2014 09:29 AM, Ching Huang wrote:
> From: Ching Huang <ching2048@areca.com.tw>
>
> This patch is to modify previous patch 13/17 and it is relative to
> http://git.infradead.org/users/hch/scsi-queue.git/tree/arcmsr-for-3.18:/drivers/scsi/arcmsr
>
> change in v4:
> 1. for readability, rename firstindex to getIndex, rename lastindex to putIndex
For some reason, the names head+tail areusual for a circular buffer.
But let us ignore the names, I don't care.
> 2. define ARCMSR_API_DATA_BUFLEN as 1032
> 3. simplify ioctl data read by macro CIRC_CNT_TO_END and CIRC_CNT
It's definitely better when you post renames and other non-functional changes in separate
patches, it's easier for the reviewer. 
>
> Signed-off-by: Ching Huang <ching2048@areca.com.tw>
> ---
>
> diff -uprN a/drivers/scsi/arcmsr/arcmsr_attr.c b/drivers/scsi/arcmsr/arcmsr_attr.c
> --- a/drivers/scsi/arcmsr/arcmsr_attr.c	2014-08-21 12:14:27.000000000 +0800
> +++ b/drivers/scsi/arcmsr/arcmsr_attr.c	2014-09-12 15:18:46.659125000 +0800
> @@ -50,6 +50,7 @@
>  #include <linux/errno.h>
>  #include <linux/delay.h>
>  #include <linux/pci.h>
> +#include <linux/circ_buf.h>
>  
>  #include <scsi/scsi_cmnd.h>
>  #include <scsi/scsi_device.h>
> @@ -68,7 +69,7 @@ static ssize_t arcmsr_sysfs_iop_message_
>  	struct device *dev = container_of(kobj,struct device,kobj);
>  	struct Scsi_Host *host = class_to_shost(dev);
>  	struct AdapterControlBlock *acb = (struct AdapterControlBlock *) host->hostdata;
> -	uint8_t *pQbuffer,*ptmpQbuffer;
> +	uint8_t *ptmpQbuffer;
>  	int32_t allxfer_len = 0;
>  	unsigned long flags;
>  
> @@ -78,57 +79,22 @@ static ssize_t arcmsr_sysfs_iop_message_
>  	/* do message unit read. */
>  	ptmpQbuffer = (uint8_t *)buf;
>  	spin_lock_irqsave(&acb->rqbuffer_lock, flags);
> -	if (acb->rqbuf_firstindex != acb->rqbuf_lastindex) {
> -		pQbuffer = &acb->rqbuffer[acb->rqbuf_firstindex];
> -		if (acb->rqbuf_firstindex > acb->rqbuf_lastindex) {
> -			if ((ARCMSR_MAX_QBUFFER - acb->rqbuf_firstindex) >= 1032) {
> -				memcpy(ptmpQbuffer, pQbuffer, 1032);
> -				acb->rqbuf_firstindex += 1032;
> -				acb->rqbuf_firstindex %= ARCMSR_MAX_QBUFFER;
> -				allxfer_len = 1032;
> -			} else {
> -				if (((ARCMSR_MAX_QBUFFER - acb->rqbuf_firstindex)
> -					+ acb->rqbuf_lastindex) > 1032) {
> -					memcpy(ptmpQbuffer, pQbuffer,
> -						ARCMSR_MAX_QBUFFER
> -						- acb->rqbuf_firstindex);
> -					ptmpQbuffer += ARCMSR_MAX_QBUFFER
> -						- acb->rqbuf_firstindex;
> -					memcpy(ptmpQbuffer, acb->rqbuffer, 1032
> -						- (ARCMSR_MAX_QBUFFER -
> -						acb->rqbuf_firstindex));
> -					acb->rqbuf_firstindex = 1032 -
> -						(ARCMSR_MAX_QBUFFER -
> -						acb->rqbuf_firstindex);
> -					allxfer_len = 1032;
> -				} else {
> -					memcpy(ptmpQbuffer, pQbuffer,
> -						ARCMSR_MAX_QBUFFER -
> -						acb->rqbuf_firstindex);
> -					ptmpQbuffer += ARCMSR_MAX_QBUFFER -
> -						acb->rqbuf_firstindex;
> -					memcpy(ptmpQbuffer, acb->rqbuffer,
> -						acb->rqbuf_lastindex);
> -					allxfer_len = ARCMSR_MAX_QBUFFER -
> -						acb->rqbuf_firstindex +
> -						acb->rqbuf_lastindex;
> -					acb->rqbuf_firstindex =
> -						acb->rqbuf_lastindex;
> -				}
> -			}
> -		} else {
> -			if ((acb->rqbuf_lastindex - acb->rqbuf_firstindex) > 1032) {
> -				memcpy(ptmpQbuffer, pQbuffer, 1032);
> -				acb->rqbuf_firstindex += 1032;
> -				allxfer_len = 1032;
> -			} else {
> -				memcpy(ptmpQbuffer, pQbuffer, acb->rqbuf_lastindex
> -					- acb->rqbuf_firstindex);
> -				allxfer_len = acb->rqbuf_lastindex -
> -					acb->rqbuf_firstindex;
> -				acb->rqbuf_firstindex = acb->rqbuf_lastindex;
> -			}
> +	if (acb->rqbuf_getIndex != acb->rqbuf_putIndex) {
> +		unsigned int tail = acb->rqbuf_getIndex;
> +		unsigned int head = acb->rqbuf_putIndex;
> +		unsigned int cnt_to_end = CIRC_CNT_TO_END(head, tail, ARCMSR_MAX_QBUFFER);
> +
> +		allxfer_len = CIRC_CNT(head, tail, ARCMSR_MAX_QBUFFER);
> +		if (allxfer_len > ARCMSR_API_DATA_BUFLEN)
> +			allxfer_len = ARCMSR_API_DATA_BUFLEN;
> +
> +		if (allxfer_len <= cnt_to_end)
> +			memcpy(ptmpQbuffer, acb->rqbuffer + tail, allxfer_len);
> +		else {
> +			memcpy(ptmpQbuffer, acb->rqbuffer + tail, cnt_to_end);
> +			memcpy(ptmpQbuffer + cnt_to_end, acb->rqbuffer, allxfer_len - cnt_to_end);
>  		}
> +		acb->rqbuf_getIndex = (acb->rqbuf_getIndex + allxfer_len) % ARCMSR_MAX_QBUFFER;
>  	}
>  	if (acb->acb_flags & ACB_F_IOPDATA_OVERFLOW) {
>  		struct QBUFFER __iomem *prbuffer;
> @@ -150,33 +116,32 @@ static ssize_t arcmsr_sysfs_iop_message_
>  	struct device *dev = container_of(kobj,struct device,kobj);
>  	struct Scsi_Host *host = class_to_shost(dev);
>  	struct AdapterControlBlock *acb = (struct AdapterControlBlock *) host->hostdata;
> -	int32_t my_empty_len, user_len, wqbuf_firstindex, wqbuf_lastindex;
> +	int32_t my_empty_len, user_len, wqbuf_getIndex, wqbuf_putIndex;
>  	uint8_t *pQbuffer, *ptmpuserbuffer;
>  	unsigned long flags;
>  
>  	if (!capable(CAP_SYS_ADMIN))
>  		return -EACCES;
> -	if (count > 1032)
> +	if (count > ARCMSR_API_DATA_BUFLEN)
>  		return -EINVAL;
>  	/* do message unit write. */
>  	ptmpuserbuffer = (uint8_t *)buf;
>  	user_len = (int32_t)count;
>  	spin_lock_irqsave(&acb->wqbuffer_lock, flags);
> -	wqbuf_lastindex = acb->wqbuf_lastindex;
> -	wqbuf_firstindex = acb->wqbuf_firstindex;
> -	if (wqbuf_lastindex != wqbuf_firstindex) {
> +	wqbuf_putIndex = acb->wqbuf_putIndex;
> +	wqbuf_getIndex = acb->wqbuf_getIndex;
> +	if (wqbuf_putIndex != wqbuf_getIndex) {
>  		arcmsr_write_ioctldata2iop(acb);
>  		spin_unlock_irqrestore(&acb->wqbuffer_lock, flags);
>  		return 0;	/*need retry*/
>  	} else {
> -		my_empty_len = (wqbuf_firstindex-wqbuf_lastindex - 1)
> -			&(ARCMSR_MAX_QBUFFER - 1);
> +		my_empty_len = ARCMSR_MAX_QBUFFER - 1;
This^ doesn't look like like an rename can you explain?

Let us stop here, or we end in an endless loop of corrections. The original 13/17
you are trying to improve here was at least without bugs (better said I haven't noticed) so
improving it now only complicates the process.
My suggestion is -let us skip this patch and focus only on fixing the spinlock problem found in 16/17.
OKay?

Cheers,
Tomas