pci: fix unneeded includes in public header file
PATCH 1/1 pci: fix unneeded includes in public header file
The rte_pci.h file includes more header files than are actually needed, which means that all users of it also include those headers. This patch removes the unneeded headers - adding them elsewhere where other components were requiring them but not including them directly.
--- drivers/bus/ifpga/rte_bus_ifpga.h | 1 + drivers/common/mlx5/mlx5_common.h | 1 + lib/librte_pci/rte_pci.h | 6 ------ 3 files changed, 2 insertions(+), 6 deletions(-)
drivers/bus/ifpga/rte_bus_ifpga.h +1 -0
@@ -17,6 +17,7 @@ extern "C" {#include <rte_bus.h> #include <rte_pci.h> +#include <rte_interrupts.h> #include <rte_spinlock.h> /** Name of Intel FPGA Bus */
drivers/common/mlx5/mlx5_common.h +1 -0
@@ -8,6 +8,7 @@#include <stdio.h> #include <rte_pci.h> +#include <rte_debug.h> #include <rte_atomic.h> #include <rte_log.h> #include <rte_kvargs.h>
lib/librte_pci/rte_pci.h +0 -6
@@ -16,17 +16,11 @@extern "C" {#endif -#include <stdio.h> #include <stdlib.h> #include <limits.h> -#include <errno.h> #include <sys/queue.h> -#include <stdint.h> #include <inttypes.h> -#include <rte_debug.h> -#include <rte_interrupts.h> -/** Formatting string for PCI device identifier: Ex: 0000:00:01.0 */ #define PCI_PRI_FMT "%.4" PRIx16 ":%.2" PRIx8 ":%.2" PRIx8 ".%" PRIx8#define PCI_PRI_STR_SIZE sizeof("XXXXXXXX:XX:XX.X")
Signature · 1 line
2.20.1
16/03/2020 12:29, Bruce Richardson:
Quoted 10 lines
The
rte_pci.hfile includes more header files than are actually needed, which means that all users of it also include those headers. This patch removes the unneeded headers - adding them elsewhere where other components were requiring them but not including them directly.Signed-off-by Bruce Richardson <[email protected]>
[...]
lib/librte_pci/rte_pci.h line 28
Quoted 12 lines
lib/librte_pci/rte_pci.h +0
-5-#include <stdio.h> #include <stdlib.h> #include <limits.h> -#include <errno.h> #include <sys/queue.h> -#include <stdint.h> #include <inttypes.h> -#include <rte_debug.h> -#include <rte_interrupts.h>
Removing useless includes (or lines in general) is always good.
Applied, thanks
Quoted 54 lines
The
rte_pci.hfile includes more header files than are actually needed, which means that all users of it also include those headers. This patch removes the unneeded headers - adding them elsewhere where other components were requiring them but not including them directly.Signed-off-by Bruce Richardson <[email protected]>--- drivers/bus/ifpga/rte_bus_ifpga.h | 1 + drivers/common/mlx5/mlx5_common.h | 1 + lib/librte_pci/rte_pci.h | 6 ------ 3 files changed, 2 insertions(+), 6 deletions(-)drivers/bus/ifpga/rte_bus_ifpga.h +1
-0@@ -17,6 +17,7 @@ extern "C" {#include <rte_bus.h> #include <rte_pci.h> +#include <rte_interrupts.h> #include <rte_spinlock.h> /** Name of Intel FPGA Bus */ drivers/common/mlx5/mlx5_common.h +1
-0@@ -8,6 +8,7 @@#include <stdio.h> #include <rte_pci.h> +#include <rte_debug.h> #include <rte_atomic.h> #include <rte_log.h> #include <rte_kvargs.h> lib/librte_pci/rte_pci.h +0
-2@@ -16,17 +16,11 @@extern "C" {#endif -#include <stdio.h> #include <stdlib.h> #include <limits.h> -#include <errno.h>
lib/librte_pci/rte_pci.h line 26
Quoted 10 lines
lib/librte_pci/rte_pci.h +0
-4#include <sys/queue.h> -#include <stdint.h> #include <inttypes.h> -#include <rte_debug.h> -#include <rte_interrupts.h> -/** Formatting string for PCI device identifier: Ex: 0000:00:01.0 */ #define PCI_PRI_FMT "%.4" PRIx16 ":%.2" PRIx8 ":%.2" PRIx8 ".%" PRIx8#define PCI_PRI_STR_SIZE sizeof("XXXXXXXX:XX:XX.X")
Patch is generating build errors with meson on my POWER system, though I don't see how they're POWER specific:
Patch is generating build errors on my POWER system:
[17/964] Compiling C object 'lib/76b5a35@@rte_pci@sta/librte_pci_rte_pci.c.o'.../lib/librte_pci/rte_pci.c: In function ‘rte_pci_device_name’:../lib/librte_pci/rte_pci.c:94:2: warning: implicit declaration of function ‘RTE_VERIFY’; did you mean ‘RTE_PRIO’? [-Wimplicit-function-declaration]
RTE_VERIFY(size >= PCI_PRI_STR_SIZE);
^~~~~~~~~~
RTE_PRIO../lib/librte_pci/rte_pci.c:94:2: warning: nested extern declaration of ‘RTE_VERIFY’ [-Wnested-externs]
[18/964] Compiling C object 'drivers/a715181@@tmp_rte_bus_ifpga@sta/bus_ifpga_ifpga_bus.c.o'.../drivers/bus/ifpga/ifpga_bus.c: In function ‘rte_ifpga_driver_register’:../drivers/bus/ifpga/ifpga_bus.c:50:2: warning: implicit declaration of function ‘RTE_VERIFY’; did you mean ‘RTE_PRIO’? [-Wimplicit-function-declaration]
RTE_VERIFY(driver);
^~~~~~~~~~
RTE_PRIO../drivers/bus/ifpga/ifpga_bus.c:50:2: warning: nested extern declaration of ‘RTE_VERIFY’ [-Wnested-externs]
[332/964] Linking target lib/librte_pci.so.20.0.2.
FAILED: lib/librte_pci.so.20.0.2
cc -o lib/librte_pci.so.20.0.2 'lib/76b5a35@@rte_pci@sta/librte_pci_rte_pci.c.o' -Wl,--no-undefined -Wl,--as-needed -Wl,-O1 -shared -fPIC -Wl,--start-group -Wl,-soname,librte_pci.so.20.0 -Wl,--no-as-needed -pthread -lm -ldl -lnuma lib/librte_eal.so.20.0.2 lib/librte_kvargs.so.20.0.2 -Wl,--end-group -Wl,--version-script=/home/drc/src/dpdk/lib/librte_pci/rte_pci_version.map '-Wl,-rpath,$ORIGIN/' -Wl,-rpath-link,/home/drc/src/dpdk/build/lib lib/76b5a35@@rte_pci@sta/librte_pci_rte_pci.c.o: In function `rte_pci_device_name':
rte_pci.c:(.text+0x38): undefined reference to `RTE_VERIFY'
rte_pci.c:(.text+0x6c): undefined reference to `RTE_VERIFY'
collect2: error: ld returned 1 exit status
[461/964] Compiling C object 'drivers/a715181@@tmp_rte_pmd_mlx5@sta/net_mlx5_mlx5_rxtx.c.o'.
ninja: build stopped: subcommand failed.
Added <rte_debug.h> to the affected files resolves the problem.
Dave
Quoted 6 lines · David Christensen
Quoted 6 lines
The
rte_pci.hfile includes more header files than are actually needed, which means that all users of it also include those headers. This patch removes the unneeded headers - adding them elsewhere where other components were requiring them but not including them directly.
Quoted 7 lines
Quoted 5 lines
Signed-off-by Bruce Richardson <[email protected]>Patch is generating build errors with meson on my POWER system, though I don't see how they're POWER specific:
There's a pending patch to enable Travis for POWER systems:
http://patches.dpdk.org/patch/65986/
Any reason it hasn't been accepted yet? Looks like it would have caught this problem.
Dave
Quoted 17 lines · David Christensen
Quoted 7 lines · David Christensen
Quoted 7 lines
The
rte_pci.hfile includes more header files than are actually needed, which means that all users of it also include those headers. This patch removes the unneeded headers - adding them elsewhere where other components
were requiring them but not including them directly.Quoted 8 lines
Quoted 5 lines
Signed-off-by Bruce Richardson <[email protected]>Patch is generating build errors with meson on my POWER system, though I don't see how they're POWER specific:
Quoted 6 lines
There's a pending patch to enable Travis for POWER systems:
http://patches.dpdk.org/patch/65986/
Any reason it hasn't been accepted yet? Looks like it would have caught this problem.
The reason is that this series has a change in the IOVA detection code. I have a lot of things to look at, so any help in reviews is appreciated.
Signature · 1 line
David Marchand
Quoted 22 lines · David Marchand
Quoted 19 lines · David Christensen
Quoted 7 lines · David Christensen
Quoted 7 lines
The
rte_pci.hfile includes more header files than are actually needed, which means that all users of it also include those headers. This patch removes the unneeded headers - adding them elsewhere where other components
were requiring them but not including them directly.Quoted 8 lines
Quoted 5 lines
Signed-off-by Bruce Richardson <[email protected]>Patch is generating build errors with meson on my POWER system, though I don't see how they're POWER specific:
There's a pending patch to enable Travis for POWER systems:
Quoted 7 lines
Quoted 4 lines
http://patches.dpdk.org/patch/65986/
Any reason it hasn't been accepted yet? Looks like it would have caught this problem.
The reason is that this series has a change in the IOVA detection code. I have a lot of things to look at, so any help in reviews is appreciated.
Reviewed...
Quoted 5 lines
--
David Marchand