From: "nobu (Nobuyoshi Nakada)" Date: 2021-10-19T11:03:57+00:00 Subject: [ruby-core:105678] [Ruby master Bug#18255] ioctl zeroes the last buffer byte Issue #18255 has been updated by nobu (Nobuyoshi Nakada). Status changed from Open to Feedback vihai (Daniele Orlandi) wrote: > The ioctl sanity check is triggered as if the buffer was too small however the size of the buffer passed to ioctl is correct. > > ``` > io.rb:116:in `ioctl': return value overflowed string (ArgumentError) > ``` > > If I append at least one byte to the buffer the ioctl does not raise an exception. Do you mean that `fd.ioctl(GPIO_GET_LINEHANDLE_IOCTL, "X"*364)` raises the exception? > It seems that the last byte of the buffer is zeroed: Yes, it's the sentinel byte added internally. ---------------------------------------- Bug #18255: ioctl zeroes the last buffer byte https://bugs.ruby-lang.org/issues/18255#change-94178 * Author: vihai (Daniele Orlandi) * Status: Feedback * Priority: Normal * Backport: 2.6: UNKNOWN, 2.7: UNKNOWN, 3.0: UNKNOWN ---------------------------------------- Hello, I'm running ruby 2.7.4p191 on an armv7 linux and experimenting with GPIO_GET_LINEHANDLE_IOCTL ioctl. The ioctl sanity check is triggered as if the buffer was too small however the size of the buffer passed to ioctl is correct. ``` io.rb:116:in `ioctl': return value overflowed string (ArgumentError) ``` If I append at least one byte to the buffer the ioctl does not raise an exception. It seems that the last byte of the buffer is zeroed: ``` puts "SIZE=#{req.bytesize}" req = req + "XXXXXXXXXX".b puts req.unpack("H*") fd.ioctl(GPIO_GET_LINEHANDLE_IOCTL, req) puts req.unpack("H*") ``` ``` SIZE=364 [...]0000000000000058585858585858585858 [...]0000000600000058585858585858585800 ``` I checked with a C program and the ioctl does not actually touch the buffer beyond the expected 364 bytes. The ioctl number does encode 364 as size: ``` #include #include void main() { printf("SIZE=%d", _IOC_SIZE(GPIO_GET_LINEHANDLE_IOCTL)); } ``` ``` SIZE=364 ``` -- https://bugs.ruby-lang.org/ Unsubscribe: