crypto: marvell/cesa - irq balance

Balance the irqs of the marvell cesa driver over all
available cpus.
Currently all interrupts are handled by the first CPU.

From my testing with IPSec AES 256 SHA256
on my clearfog base with 2 Cores I get a 2x speed increase:

Before the patch: 26.74 Kpps
With the patch: 56.11 Kpps

Signed-off-by: Sven Auhagen <sven.auhagen@voleatech.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This commit is contained in:
Sven Auhagen 2020-07-21 06:40:27 +02:00 committed by Herbert Xu
parent c672041590
commit 28ee8b0912
2 changed files with 11 additions and 1 deletions

View File

@ -438,7 +438,7 @@ static int mv_cesa_probe(struct platform_device *pdev)
struct mv_cesa_dev *cesa; struct mv_cesa_dev *cesa;
struct mv_cesa_engine *engines; struct mv_cesa_engine *engines;
struct resource *res; struct resource *res;
int irq, ret, i; int irq, ret, i, cpu;
u32 sram_size; u32 sram_size;
if (cesa_dev) { if (cesa_dev) {
@ -505,6 +505,8 @@ static int mv_cesa_probe(struct platform_device *pdev)
goto err_cleanup; goto err_cleanup;
} }
engine->irq = irq;
/* /*
* Not all platforms can gate the CESA clocks: do not complain * Not all platforms can gate the CESA clocks: do not complain
* if the clock does not exist. * if the clock does not exist.
@ -548,6 +550,10 @@ static int mv_cesa_probe(struct platform_device *pdev)
if (ret) if (ret)
goto err_cleanup; goto err_cleanup;
/* Set affinity */
cpu = cpumask_local_spread(engine->id, NUMA_NO_NODE);
irq_set_affinity_hint(irq, get_cpu_mask(cpu));
crypto_init_queue(&engine->queue, CESA_CRYPTO_DEFAULT_MAX_QLEN); crypto_init_queue(&engine->queue, CESA_CRYPTO_DEFAULT_MAX_QLEN);
atomic_set(&engine->load, 0); atomic_set(&engine->load, 0);
INIT_LIST_HEAD(&engine->complete_queue); INIT_LIST_HEAD(&engine->complete_queue);
@ -570,6 +576,8 @@ static int mv_cesa_probe(struct platform_device *pdev)
clk_disable_unprepare(cesa->engines[i].zclk); clk_disable_unprepare(cesa->engines[i].zclk);
clk_disable_unprepare(cesa->engines[i].clk); clk_disable_unprepare(cesa->engines[i].clk);
mv_cesa_put_sram(pdev, i); mv_cesa_put_sram(pdev, i);
if (cesa->engines[i].irq > 0)
irq_set_affinity_hint(cesa->engines[i].irq, NULL);
} }
return ret; return ret;
@ -586,6 +594,7 @@ static int mv_cesa_remove(struct platform_device *pdev)
clk_disable_unprepare(cesa->engines[i].zclk); clk_disable_unprepare(cesa->engines[i].zclk);
clk_disable_unprepare(cesa->engines[i].clk); clk_disable_unprepare(cesa->engines[i].clk);
mv_cesa_put_sram(pdev, i); mv_cesa_put_sram(pdev, i);
irq_set_affinity_hint(cesa->engines[i].irq, NULL);
} }
return 0; return 0;

View File

@ -457,6 +457,7 @@ struct mv_cesa_engine {
atomic_t load; atomic_t load;
struct mv_cesa_tdma_chain chain; struct mv_cesa_tdma_chain chain;
struct list_head complete_queue; struct list_head complete_queue;
int irq;
}; };
/** /**