Message ID | 1220901653-12812-1-git-send-email-becky.bruce@freescale.com (mailing list archive) |
---|---|
State | Rejected, archived |
Delegated to: | Kumar Gala |
Headers | show |
Becky Bruce wrote: > We're currently passing NULL, and really shouldn't be. This patch doesn't change that for cpm2, since port.dev is never initialized. :-P > Signed-off-by: Becky Bruce <becky.bruce@freescale.com> > --- > drivers/serial/cpm_uart/cpm_uart_cpm2.c | 6 +++--- > drivers/serial/ucc_uart.c | 4 ++-- > 2 files changed, 5 insertions(+), 5 deletions(-) cpm_uart_cpm1.c has the same problem. -Scott
On Sep 8, 2008, at 2:56 PM, Scott Wood wrote: > Becky Bruce wrote: >> We're currently passing NULL, and really shouldn't be. > > This patch doesn't change that for cpm2, since port.dev is never > initialized. :-P > >> Signed-off-by: Becky Bruce <becky.bruce@freescale.com> >> --- >> drivers/serial/cpm_uart/cpm_uart_cpm2.c | 6 +++--- >> drivers/serial/ucc_uart.c | 4 ++-- >> 2 files changed, 5 insertions(+), 5 deletions(-) > > cpm_uart_cpm1.c has the same problem. > > -Scott Good catch, thanks! Will respin. Cheers, B
On Mon, Sep 8, 2008 at 2:56 PM, Scott Wood <scottwood@freescale.com> wrote: > This patch doesn't change that for cpm2, since port.dev is never > initialized. :-P I don't initialize port->dev in the UCC UART driver, but it's still non-zero, so why isn't it initialized in the CPM UART driver? (P.S. Becky - please CC: me on any UCC UART or QE LIB patches)
Timur Tabi wrote: > On Mon, Sep 8, 2008 at 2:56 PM, Scott Wood <scottwood@freescale.com> wrote: > >> This patch doesn't change that for cpm2, since port.dev is never >> initialized. :-P > > I don't initialize port->dev in the UCC UART driver, Yes you do -- check line 1403. -Scott
diff --git a/drivers/serial/cpm_uart/cpm_uart_cpm2.c b/drivers/serial/cpm_uart/cpm_uart_cpm2.c index b8db4d3..141c0a3 100644 --- a/drivers/serial/cpm_uart/cpm_uart_cpm2.c +++ b/drivers/serial/cpm_uart/cpm_uart_cpm2.c @@ -136,7 +136,7 @@ int cpm_uart_allocbuf(struct uart_cpm_port *pinfo, unsigned int is_con) dma_addr = virt_to_bus(mem_addr); } else - mem_addr = dma_alloc_coherent(NULL, memsz, &dma_addr, + mem_addr = dma_alloc_coherent(pinfo->port.dev, memsz, &dma_addr, GFP_KERNEL); if (mem_addr == NULL) { @@ -163,8 +163,8 @@ int cpm_uart_allocbuf(struct uart_cpm_port *pinfo, unsigned int is_con) void cpm_uart_freebuf(struct uart_cpm_port *pinfo) { - dma_free_coherent(NULL, L1_CACHE_ALIGN(pinfo->rx_nrfifos * - pinfo->rx_fifosize) + + dma_free_coherent(pinfo->port.dev, L1_CACHE_ALIGN(pinfo->rx_nrfifos * + pinfo->rx_fifosize) + L1_CACHE_ALIGN(pinfo->tx_nrfifos * pinfo->tx_fifosize), (void __force *)pinfo->mem_addr, pinfo->dma_addr); diff --git a/drivers/serial/ucc_uart.c b/drivers/serial/ucc_uart.c index 5c5d18d..539c933 100644 --- a/drivers/serial/ucc_uart.c +++ b/drivers/serial/ucc_uart.c @@ -1009,7 +1009,7 @@ static int qe_uart_request_port(struct uart_port *port) rx_size = L1_CACHE_ALIGN(qe_port->rx_nrfifos * qe_port->rx_fifosize); tx_size = L1_CACHE_ALIGN(qe_port->tx_nrfifos * qe_port->tx_fifosize); - bd_virt = dma_alloc_coherent(NULL, rx_size + tx_size, &bd_dma_addr, + bd_virt = dma_alloc_coherent(port->dev, rx_size + tx_size, &bd_dma_addr, GFP_KERNEL); if (!bd_virt) { dev_err(port->dev, "could not allocate buffer descriptors\n"); @@ -1051,7 +1051,7 @@ static void qe_uart_release_port(struct uart_port *port) container_of(port, struct uart_qe_port, port); struct ucc_slow_private *uccs = qe_port->us_private; - dma_free_coherent(NULL, qe_port->bd_size, qe_port->bd_virt, + dma_free_coherent(port->dev, qe_port->bd_size, qe_port->bd_virt, qe_port->bd_dma_addr); ucc_slow_free(uccs);
We're currently passing NULL, and really shouldn't be. Signed-off-by: Becky Bruce <becky.bruce@freescale.com>