ntb_perf: Adjustments for perf_copy_chunk()
Subsystems (2)
cover ntb_perf: Adjustments for perf_copy_chunk()
From: Markus Elfring <[email protected]>
Date: Mon, 23 Sep 2024 11:00:22 +0200
A few update suggestions were taken into account
from static source code analysis.
Markus Elfring (2):
Delete duplicate dmaengine_unmap_put() call
Use common error handling code
drivers/ntb/test/ntb_perf.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-)
--
2.46.1
PATCH 1/2 ntb_perf: Delete duplicate dmaengine_unmap_put() call in perf_copy_chunk()
From: Markus Elfring <[email protected]>
Date: Mon, 23 Sep 2024 10:38:11 +0200
The function call “dmaengine_unmap_put(unmap)” was used in an if branch. The same call was immediately triggered by a subsequent goto statement. Thus avoid such a call repetition.
This issue was detected by using the Coccinelle software.
---
I imagine that a single function call “dmaengine_unmap_put(unmap)” can be sufficient for this use case.
https://elixir.bootlin.com/linux/v6.11/source/drivers/dma/dmaengine.c#L1377
How do you think about to improve resource management any further?
drivers/ntb/test/ntb_perf.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/ntb/test/ntb_perf.c +1 -3
@@ -839,10 +839,8 @@ static int perf_copy_chunk(struct perf_thread *pthr,dma_set_unmap (tx, unmap); ret =dma_submit_error (dmaengine_submit (tx)); -if (ret) { -dmaengine_unmap_put (unmap); +if (ret) goto err_free_resource; - }dmaengine_unmap_put (unmap);
--
2.46.1
PATCH 2/2 ntb_perf: Use common error handling code in perf_copy_chunk()
In reply to Markus ElfringFrom: Markus Elfring <[email protected]>
Date: Mon, 23 Sep 2024 10:48:17 +0200
Add a jump target so that a bit of exception handling can be better reused at the end of this function implementation.
--- drivers/ntb/test/ntb_perf.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-)
drivers/ntb/test/ntb_perf.c +7 -8
@@ -816,10 +816,9 @@ static int perf_copy_chunk(struct perf_thread *pthr, unmap->len = len; unmap->addr[0 ] =dma_map_page (dma_dev,virt_to_page (src),offset_in_page (src), len,DMA_TO_DEVICE ); -if (dma_mapping_error (dma_dev, unmap->addr[0 ])) { - ret = -EIO ; - goto err_free_resource; - } +if (dma_mapping_error (dma_dev, unmap->addr[0 ])) + goto e_io; + unmap->to_cnt =1 ; do {
@@ -829,10 +828,8 @@ static int perf_copy_chunk(struct perf_thread *pthr,msleep (DMA_MDELAY ); }while (!tx && (try++ <DMA_TRIES )); -if (!tx) { - ret = -EIO ; - goto err_free_resource; - } +if (!tx) + goto e_io; tx->callback = perf_dma_copy_callback; tx->callback_param = pthr;
@@ -850,6 +847,8 @@ static int perf_copy_chunk(struct perf_thread *pthr, ret_check_tsync:return likely (atomic_read (&pthr->perf->tsync) >0 ) ?0 : -EINTR ; +e_io: + ret = -EIO ; err_free_resource:dmaengine_unmap_put (unmap);
--
2.46.1